fix(models): route auto-discovered providers by wire protocol (#41)

- replace npmToLLMProvider map with npmToWireProtocol (openai/anthropic/google)
- add createAutoRoutedGoogleProvider so @ai-sdk/google proxies work
  (fixes opencode/gemini-* failing with "no LLM provider mapping")
- strip the genai-injected v1beta segment for proxies whose base URL
  already carries a version (e.g. opencode's /zen/v1)
- preserve openai-compat fallback and clearer error for unroutable providers
- document auto-routing in README and providers docs; update CreateProvider godoc
- add regression tests for wire routing and version-path rewriting

Fixes #41

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,79 @@
+name: Bug Report
+description: Report a bug or issue with Kit
+title: "fix: "
+labels: ["bug"]
+body:
+  - type: textarea
+    id: description
+    attributes:
+      label: Bug Description
+      description: What happened? What did you expect to happen?
+      placeholder: |
+        The BorderColor field in ToolRenderConfig is documented but never applied 
+        during tool rendering. I expected the tool block to render with my custom 
+        color, but it uses the default styling instead.
+    validations:
+      required: true
+
+  - type: textarea
+    id: reproduction
+    attributes:
+      label: Steps to Reproduce
+      description: Provide clear steps to reproduce the issue
+      placeholder: |
+        1. Create an extension with `api.RegisterToolRenderer(ext.ToolRenderConfig{...})`
+        2. Set `BorderColor: "#89b4fa"` in the config
+        3. Run a tool that uses this renderer
+        4. Observe the border color is not applied
+      render: markdown
+    validations:
+      required: true
+
+  - type: textarea
+    id: code
+    attributes:
+      label: Relevant Code / Configuration
+      description: Paste any code, configuration, or error messages
+      placeholder: |
+        ```go
+        api.RegisterToolRenderer(ext.ToolRenderConfig{
+            ToolName:    "bash",
+            DisplayName: "Shell",
+            BorderColor: "#a6e3a1",  // This is ignored!
+            Background:  "#1e1e2e",  // This is ignored!
+        })
+        ```
+      render: go
+
+  - type: input
+    id: component
+    attributes:
+      label: Affected Component
+      description: Which part of Kit is affected?
+      placeholder: e.g., extensions, ui, tool rendering, session management
+
+  - type: input
+    id: version
+    attributes:
+      label: Kit Version
+      description: What version of Kit are you running?
+      placeholder: e.g., v0.1.0, commit hash, or "main"
+
+  - type: textarea
+    id: context
+    attributes:
+      label: Additional Context
+      description: Any other context, proposed fixes, or related issues
+      placeholder: |
+        The issue appears to be in `internal/ui/messages.go:RenderToolMessage()` 
+        which ignores the BorderColor and Background fields from ToolRendererData.
+
+  - type: checkboxes
+    id: terms
+    attributes:
+      label: Checklist
+      options:
+        - label: I've searched existing issues and this hasn't been reported yet
+          required: true
+        - label: I've tested with the latest version of Kit
+          required: false

.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1,11 @@
+blank_issues_enabled: false
+contact_links:
+  - name: Kit Documentation
+    url: https://github.com/mark3labs/kit/tree/main/www/pages
+    about: Check the documentation before filing an issue
+  - name: Extension Examples
+    url: https://github.com/mark3labs/kit/tree/main/examples/extensions
+    about: See working extension examples for reference
+  - name: Discussions
+    url: https://github.com/mark3labs/kit/discussions
+    about: For questions, ideas, or general discussion

.github/ISSUE_TEMPLATE/documentation.yml

@@ -0,0 +1,40 @@
+name: Documentation Issue
+description: Report missing, incorrect, or unclear documentation
+title: "docs: "
+labels: ["documentation"]
+body:
+  - type: textarea
+    id: description
+    attributes:
+      label: Documentation Issue
+      description: What's wrong or missing in the documentation?
+      placeholder: |
+        The ToolRenderConfig documentation mentions BorderColor and Background fields,
+        but the code doesn't actually use them. The docs should either be updated 
+        to reflect reality, or the bug should be fixed.
+    validations:
+      required: true
+
+  - type: input
+    id: location
+    attributes:
+      label: Documentation Location
+      description: Where is the affected documentation?
+      placeholder: e.g., README.md, examples/extensions/tool-renderer-demo.go, pkg/kit docs
+
+  - type: textarea
+    id: suggestion
+    attributes:
+      label: Suggested Improvement
+      description: How should the documentation be improved?
+      placeholder: |
+        Add a note that BorderColor and Background are not yet implemented, 
+        or fix the bug and document the correct behavior.
+
+  - type: checkboxes
+    id: terms
+    attributes:
+      label: Checklist
+      options:
+        - label: I've checked that this documentation issue still exists in the latest version
+          required: true

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,64 @@
+name: Feature Request
+description: Suggest a new feature or enhancement for Kit
+title: "feat: "
+labels: ["enhancement"]
+body:
+  - type: textarea
+    id: description
+    attributes:
+      label: Feature Description
+      description: What would you like to see added or changed?
+      placeholder: |
+        I'd like to be able to customize the border color of tool result blocks 
+        dynamically based on the tool type or result status.
+    validations:
+      required: true
+
+  - type: textarea
+    id: motivation
+    attributes:
+      label: Motivation / Use Case
+      description: Why is this feature needed? What problem does it solve?
+      placeholder: |
+        When running multiple tools in sequence, it's hard to visually distinguish 
+        between file reads (blue), shell commands (green), and errors (red) 
+        without custom border colors.
+    validations:
+      required: true
+
+  - type: textarea
+    id: proposed
+    attributes:
+      label: Proposed Implementation
+      description: How do you think this should work? (optional)
+      placeholder: |
+        Extend `ToolRenderConfig` to accept a function that receives the tool 
+        result and returns a color based on the content:
+        
+        ```go
+        BorderColorFunc: func(result string, isError bool) string {
+            if isError {
+                return "#f38ba8"
+            }
+            return "#89b4fa"
+        }
+        ```
+      render: go
+
+  - type: checkboxes
+    id: alternatives
+    attributes:
+      label: Alternatives Considered
+      options:
+        - label: I've considered workarounds or alternative approaches
+          required: false
+
+  - type: checkboxes
+    id: terms
+    attributes:
+      label: Checklist
+      options:
+        - label: I've searched existing issues and this hasn't been requested yet
+          required: true
+        - label: This feature aligns with Kit's design philosophy (TUI-first, extension-based)
+          required: false

.gitignore

@@ -3,6 +3,7 @@
 .env
 .kit/*
 !.kit/extensions/
+!.kit/prompts/
 aidocs/
 *.log
 /kit

.kit/extensions/go-edit-lint.go

@@ -1,228 +0,0 @@
-//go:build ignore
-
-package main
-
-import (
-	"context"
-	"encoding/json"
-	"fmt"
-	"os/exec"
-	"path/filepath"
-	"strings"
-	"time"
-
-	"kit/ext"
-)
-
-const (
-	diagnosticsTimeout = 20 * time.Second
-	maxOutputBytes     = 12_000
-)
-
-type toolPathInput struct {
-	Path string `json:"path"`
-}
-
-type lintResult struct {
-	Output string
-	Err    error
-}
-
-func Init(api ext.API) {
-	api.OnSessionStart(func(_ ext.SessionStartEvent, ctx ext.Context) {
-		ctx.Print("go-edit-lint extension loaded - will run gopls and golangci-lint on Go file edits")
-	})
-
-	api.OnToolResult(func(e ext.ToolResultEvent, ctx ext.Context) *ext.ToolResultResult {
-		if e.IsError || !isEditOrWrite(e.ToolName) {
-			return nil
-		}
-
-		absPath, ok := resolveGoFilePath(e.Input, ctx.CWD)
-		if !ok {
-			return nil
-		}
-
-		report := runGoDiagnostics(ctx.CWD, absPath)
-		
-		// Check if there are issues and add explicit prompt for the LLM to react
-		goplsIssues, lintIssues := countIssues(report)
-		hasIssues := goplsIssues > 0 || lintIssues > 0
-		
-		var enhanced string
-		if hasIssues {
-			enhanced = e.Content + "\n\n" + report + "\n\n⚠️ DIAGNOSTICS FOUND: Please review the issues above and fix them before proceeding."
-		} else {
-			enhanced = e.Content + "\n\n" + report
-		}
-
-		// Show TUI message block for diagnostics visibility (only if there are issues)
-		if hasIssues {
-			var msgLines []string
-			msgLines = append(msgLines, fmt.Sprintf("File: %s", filepath.Base(absPath)))
-			if goplsIssues > 0 {
-				msgLines = append(msgLines, fmt.Sprintf("gopls: %d issue(s)", goplsIssues))
-			}
-			if lintIssues > 0 {
-				msgLines = append(msgLines, fmt.Sprintf("golangci-lint: %d issue(s)", lintIssues))
-			}
-			msgLines = append(msgLines, "", "⚠️ Please fix these issues before proceeding.")
-
-			borderColor := "#f9e2af" // yellow
-			if goplsIssues > 0 && lintIssues > 0 {
-				borderColor = "#f38ba8" // red
-			}
-
-			ctx.PrintBlock(ext.PrintBlockOpts{
-				Text:        strings.Join(msgLines, "\n"),
-				BorderColor: borderColor,
-				Subtitle:    "go-edit-lint",
-			})
-		}
-
-		return &ext.ToolResultResult{Content: &enhanced}
-	})
-}
-
-func isEditOrWrite(toolName string) bool {
-	return strings.EqualFold(toolName, "edit") || strings.EqualFold(toolName, "write")
-}
-
-func resolveGoFilePath(inputJSON, cwd string) (string, bool) {
-	var args toolPathInput
-	if err := json.Unmarshal([]byte(inputJSON), &args); err != nil || args.Path == "" {
-		return "", false
-	}
-
-	absPath := args.Path
-	if !filepath.IsAbs(absPath) {
-		absPath = filepath.Join(cwd, absPath)
-	}
-
-	if strings.ToLower(filepath.Ext(absPath)) != ".go" {
-		return "", false
-	}
-
-	return absPath, true
-}
-
-func runGoDiagnostics(cwd, absPath string) string {
-	gopls := runGopls(cwd, absPath)
-	lint := runGolangCILint(cwd, "./...")
-
-	return fmt.Sprintf(
-		"<go_diagnostics file=%q>\n[gopls]\n%s\n\n[golangci-lint]\n%s\n</go_diagnostics>",
-		filepath.Base(absPath),
-		formatToolResult(gopls, "No diagnostics."),
-		formatToolResult(lint, "No lint issues."),
-	)
-}
-
-func runGopls(cwd, absPath string) lintResult {
-	ctx, cancel := context.WithTimeout(context.Background(), diagnosticsTimeout)
-	defer cancel()
-
-	cmd := exec.CommandContext(ctx, "gopls", "check", absPath)
-	cmd.Dir = cwd
-	out, err := cmd.CombinedOutput()
-
-	if ctx.Err() == context.DeadlineExceeded {
-		return lintResult{Err: fmt.Errorf("timed out after %s", diagnosticsTimeout)}
-	}
-
-	if err != nil {
-		return lintResult{Output: truncate(string(out), maxOutputBytes), Err: fmt.Errorf("failed to run gopls check: %w", err)}
-	}
-
-	return lintResult{Output: truncate(string(out), maxOutputBytes)}
-}
-
-func runGolangCILint(cwd, target string) lintResult {
-	ctx, cancel := context.WithTimeout(context.Background(), diagnosticsTimeout)
-	defer cancel()
-
-	args := []string{
-		"run",
-		target,
-		"--show-stats=false",
-		"--output.text.path", "stdout",
-		"--output.text.colors=false",
-		"--output.text.print-issued-lines=false",
-	}
-	cmd := exec.CommandContext(ctx, "golangci-lint", args...)
-	cmd.Dir = cwd
-	out, err := cmd.CombinedOutput()
-
-	if ctx.Err() == context.DeadlineExceeded {
-		return lintResult{Err: fmt.Errorf("timed out after %s", diagnosticsTimeout)}
-	}
-
-	trimmed := truncate(string(out), maxOutputBytes)
-	if err == nil {
-		return lintResult{Output: trimmed}
-	}
-
-	exitErr, ok := err.(*exec.ExitError)
-	if ok && exitErr.ExitCode() == 1 {
-		return lintResult{Output: trimmed}
-	}
-
-	return lintResult{Output: trimmed, Err: fmt.Errorf("failed to run golangci-lint: %w", err)}
-}
-
-func formatToolResult(res lintResult, emptyFallback string) string {
-	var lines []string
-	if res.Err != nil {
-		lines = append(lines, "ERROR: "+res.Err.Error())
-	}
-	out := strings.TrimSpace(res.Output)
-	if out == "" {
-		if res.Err == nil {
-			lines = append(lines, emptyFallback)
-		}
-	} else {
-		lines = append(lines, out)
-	}
-	if len(lines) == 0 {
-		return emptyFallback
-	}
-	return strings.Join(lines, "\n")
-}
-
-func truncate(s string, max int) string {
-	if len(s) <= max {
-		return s
-	}
-	return s[:max] + "\n... output truncated ..."
-}
-
-func countIssues(report string) (goplsCount, lintCount int) {
-	// Extract gopls section
-	goplsStart := strings.Index(report, "[gopls]")
-	lintStart := strings.Index(report, "[golangci-lint]")
-	endTag := strings.Index(report, "</go_diagnostics>")
-
-	if goplsStart != -1 && lintStart != -1 {
-		goplsSection := report[goplsStart:lintStart]
-		// Count non-empty lines excluding the header and "No diagnostics." message
-		for _, line := range strings.Split(goplsSection, "\n") {
-			line = strings.TrimSpace(line)
-			if line != "" && line != "[gopls]" && line != "No diagnostics." {
-				goplsCount++
-			}
-		}
-	}
-
-	if lintStart != -1 && endTag != -1 {
-		lintSection := report[lintStart:endTag]
-		// Count non-empty lines excluding the header and "No lint issues." message
-		for _, line := range strings.Split(lintSection, "\n") {
-			line = strings.TrimSpace(line)
-			if line != "" && line != "[golangci-lint]" && line != "No lint issues." {
-				lintCount++
-			}
-		}
-	}
-
-	return goplsCount, lintCount
-}

.kit/prompts/acp-smoke-test.md

@@ -0,0 +1,37 @@
+---
+description: Run ACP smoke test against opencode/kimi-k2.5 to verify JSON-RPC stdio works
+---
+
+Run the ACP smoke test to verify the Kit ACP server works correctly over JSON-RPC stdio with streaming responses.
+
+## Steps
+
+1. Build the kit binary:
+   ```bash
+   go build -o output/kit ./cmd/kit
+   ```
+
+2. Run the smoke test Python script against opencode/kimi-k2.5:
+   ```bash
+   python3 scripts/acp_smoke_test.py
+   ```
+
+3. Verify the output shows:
+   - `session/new` returns a valid `sessionId`
+   - `session/prompt` streams `agent_thought_chunk` notifications (reasoning)
+   - `session/prompt` streams `agent_message_chunk` notifications (response)
+   - Final result has `stopReason: "end_turn"`
+   - `✓ SMOKE TEST PASSED` at the end
+
+4. If the test fails, check:
+   - `output/kit` binary exists and is executable
+   - `OPENCODE_API_KEY` or `OPENCODE_ZEN_API_KEY` environment variable is set
+   - `scripts/acp_smoke_test.py` exists
+   - The model `opencode/kimi-k2.5` is available (`kit models opencode | grep kimi-k2.5`)
+
+5. For testing with a different model, edit the script or set the `MODEL` variable:
+   ```bash
+   MODEL=anthropic/claude-sonnet-4-5 python3 scripts/acp_smoke_test.py
+   ```
+
+The smoke test exercises the full ACP protocol: session lifecycle, streaming notifications, and tool-free prompt completion.

.kit/prompts/code-audit.md

@@ -0,0 +1,146 @@
+---
+description: Read-only audit for dead code, duplication, boundary violations, and refactor opportunities
+---
+
+Perform a comprehensive **read-only** audit of this repository and report
+findings. **Do not edit, rename, or delete any files.** Optional focus / scope
+hints from the user: $@
+
+## Scope
+
+If the user supplied focus hints above (a package path, a subsystem name, a
+concern like "TUI" or "extensions"), scope the audit accordingly. Otherwise
+audit the whole repo, prioritising the highest-traffic packages first
+(`cmd/`, `internal/`, `pkg/kit/` for this repo).
+
+## Steps
+
+1. **Map the repo first**:
+   - `ls` / `find` the top-level layout and list every Go package
+   - Read `AGENTS.md`, `README.md`, and any `pkg/*/doc.go` to understand the
+     intended architectural boundaries (SDK vs internal vs TUI vs cmd vs
+     extension surface)
+   - Note the public SDK surface (`pkg/kit/`) and any documented invariants
+     (e.g. "no dependency name leakage", "UI never imports extensions
+     directly") — these define what counts as a violation
+
+2. **Hunt for dead code**:
+   - Run `go vet ./...` and capture warnings
+   - Use `grep` to find exported symbols (`^func [A-Z]`, `^type [A-Z]`,
+     `^var [A-Z]`, `^const [A-Z]`) and cross-reference call sites. Symbols
+     with zero non-test references inside the module are suspects
+   - Check for unreferenced files, `// TODO: remove` markers, commented-out
+     blocks, and `_ = x` discard patterns
+   - If `staticcheck`, `deadcode`, or `unused` are available on PATH, run
+     them and include their output verbatim
+   - **Do not delete anything** — list candidates with file:line and a
+     confidence level (high / medium / low)
+
+3. **Find unnecessary duplication**:
+   - Look for near-identical function bodies, struct shapes, or switch
+     statements across packages — `grep` for repeated function signatures
+     and copy-pasted string literals / error messages is a fast first pass
+   - Distinguish *coincidental* duplication (two things that happen to look
+     alike but evolve independently) from *unnecessary* duplication (same
+     intent, drifting in lockstep) — only flag the latter
+   - For each cluster, propose where the extracted helper should live
+     (which package, which file) and whether it crosses a boundary
+
+4. **Check concerns / boundary violations**:
+   - **SDK leakage**: grep `pkg/kit/` for imports of `internal/...` types
+     in exported signatures, and for dependency-name leakage in exported
+     names / godoc (e.g. library jargon appearing in `LLM*` types)
+   - **UI ↔ extensions**: grep `internal/ui/` for any import of
+     `internal/extensions/` — per AGENTS.md the UI must not import
+     extensions directly; converters in `cmd/root.go` should bridge them
+   - **cmd vs internal**: business logic living in `cmd/` that should be
+     in `internal/` (and vice versa)
+   - **Cyclic risk**: packages that import each other transitively or that
+     reach across sibling boundaries unexpectedly
+   - For each violation, cite the offending import / signature with
+     file:line
+
+5. **Spot refactor opportunities**:
+   - Long functions (>80 lines) doing multiple unrelated things
+   - Deeply nested conditionals that flatten well with early returns
+   - Repeated `if err != nil { return fmt.Errorf("...: %w", err) }` chains
+     that could become helpers — but only where the wrapping context is
+     genuinely uniform
+   - Structs with too many fields that hint at split responsibilities
+   - Exported APIs that would be cleaner with options structs / functional
+     options
+   - Tests that share setup boilerplate ripe for a helper
+   - Flag each with: location, current shape (1-2 lines), proposed shape
+     (1-2 lines), and estimated risk (low / medium / high)
+
+6. **Cross-check against project rules**:
+   - Re-read `AGENTS.md` "Key Patterns" section and verify nothing in your
+     findings contradicts the documented gotchas (Yaegi interface ban,
+     `prog.Send()` from `Update()`, function-field bug, etc.) — if a
+     "refactor" would reintroduce a known pitfall, drop it from the report
+     and note why
+
+7. **Write the report** as your final message (do not write it to disk)
+   structured as:
+
+   ```
+   # Code Audit Report
+
+   ## Summary
+   - N dead-code candidates
+   - N duplication clusters
+   - N boundary violations
+   - N refactor opportunities
+
+   ## Dead Code
+   ### High confidence
+   - path/to/file.go:LINE — symbol — reason
+
+   ### Medium confidence
+   ...
+
+   ## Duplication
+   ### Cluster: <short name>
+   - Sites: file:line, file:line, …
+   - Suggested home: package/path
+   - Notes: …
+
+   ## Boundary Violations
+   - Rule: <which rule from AGENTS.md / project convention>
+   - Offender: file:line
+   - Fix sketch: …
+
+   ## Refactor Opportunities
+   - Location: file:line
+   - Current: …
+   - Proposed: …
+   - Risk: low/medium/high
+   - Why it's worth it: …
+
+   ## Suggested Next Steps
+   1. …
+   2. …
+   ```
+
+8. **End the report with an explicit reminder** that no files were modified,
+   and recommend the user pick the highest-leverage items to act on
+   manually (or via a follow-up `/fix-issue` style prompt) rather than
+   running a sweeping refactor.
+
+## Guidelines
+
+- **Read-only, always**: no `edit`, no `write`, no `git commit`, no `go mod
+  tidy`. Use only `read`, `grep`, `find`, `ls`, and read-only `bash`
+  commands (`go vet`, `go build -o /tmp/...`, `staticcheck`, etc.)
+- **Cite every finding** with `path/to/file.go:LINE` so the user can jump
+  straight to it
+- **Be honest about confidence**: false positives in a code audit are
+  expensive — prefer "medium confidence, worth a look" over confidently
+  wrong claims
+- **Quantity isn't quality**: 10 sharp findings beat 100 nitpicks. Cut
+  anything that's purely stylistic unless it directly causes one of the
+  four issue categories above
+- **Skip generated code** (`*.pb.go`, `*_gen.go`, anything under
+  `vendor/`) and obvious third-party copies
+- **Don't propose architectural rewrites** — stay within the existing
+  shape of the repo and recommend incremental, reviewable changes

.kit/prompts/commit-push.md

@@ -0,0 +1,30 @@
+---
+description: Stage, commit, and push changes with an auto-generated conventional commit message
+---
+
+Review the current git status and diff, then stage all changes, write a concise conventional commit message, commit, and push to the current branch.
+
+## Steps
+
+1. **Check status**: `git status` — understand what has changed
+2. **Review the diff**: `git diff` (and `git diff --cached` if anything is already staged) — read the actual changes
+3. **Stage everything**: `git add -A`
+4. **Craft the commit message** following Conventional Commits:
+   - Format: `<type>(<scope>): <short summary>`
+   - Types: `feat`, `fix`, `refactor`, `chore`, `docs`, `test`, `perf`, `build`
+   - Scope: optional, the subsystem affected (e.g. `ui`, `cmd`, `config`)
+   - Summary: imperative mood, lowercase, no trailing period, ≤72 chars
+   - Body: add a blank line then bullet points for non-trivial changes
+   - Do **not** include "Generated by" or similar noise
+5. **Commit**: `git commit -m "<message>"`
+6. **Push**: `git push`
+
+## Guidelines
+
+- Read the actual diff — do not guess from filenames alone
+- Prefer one well-scoped commit; do not split unless the changes are clearly unrelated
+- Keep the subject line under 72 characters
+- Use the body to explain *what* and *why*, not *how*
+- If there is nothing to commit, say so and stop
+
+$@

.kit/prompts/create-pr.md

@@ -0,0 +1,47 @@
+---
+description: Open a GitHub PR for the current branch using the repo's PR template
+---
+
+Open a GitHub pull request for the current branch, filling out the repository's PR template with a description grounded in the actual commits and diff.
+
+## Steps
+
+1. **Verify the branch is pushed**:
+   - `git status -sb` and `git log @{u}..HEAD --oneline 2>/dev/null` — if there is no upstream or unpushed commits, run `git push -u origin "$(git branch --show-current)"` first
+   - If the working tree is dirty, stop and tell the user to commit first (suggest `/commit-push`)
+2. **Gather context**:
+   - `git log origin/main..HEAD --oneline` — list of commits going into the PR
+   - `git diff origin/main...HEAD --stat` then `git diff origin/main...HEAD` — read the actual changes
+   - Identify the linked issue (from commit messages, branch name, or extra user input: $@) — capture as `Fixes #N` if applicable
+3. **Locate the PR template**:
+   - Check `.github/pull_request_template.md`, `.github/PULL_REQUEST_TEMPLATE.md`, or `docs/pull_request_template.md`
+   - If none exists, use a minimal `## Description` / `## Type of Change` / `## Checklist` structure
+4. **Draft the PR body** by filling out the template:
+   - **Description**: 1–3 short paragraphs explaining *what* changed and *why*, grounded in the diff. Include a brief before/after example for new APIs when useful.
+   - **Fixes #N**: only if there is a real linked issue
+   - **Type of Change**: tick the single most accurate box with `[x]` (leave others as `[ ]`)
+   - **Checklist**: tick items that are genuinely true (style, self-review, tests added, docs updated)
+   - **Additional Information**: bullet list of added / modified files and any backward-compatibility notes
+   - Remove template sections explicitly marked "remove if not applicable" (e.g. MCP Spec Compliance) when they don't apply
+5. **Write the body to a temp file**: `/tmp/pr-body-<branch-or-issue>.md` — never inline a long body via `--body`, always use `--body-file`
+6. **Choose the title**: prefer the subject of the primary commit if it already follows Conventional Commits; otherwise craft one in the same style (`<type>(<scope>): <imperative summary>`, ≤72 chars)
+7. **Create the PR**:
+   ```
+   gh pr create \
+     --title "<title>" \
+     --body-file /tmp/pr-body-<...>.md \
+     --base main \
+     --head "$(git branch --show-current)"
+   ```
+   Use the repo's actual default branch if it isn't `main` (`gh repo view --json defaultBranchRef -q .defaultBranchRef.name`)
+8. **Report the PR URL** returned by `gh` and stop
+
+## Guidelines
+
+- Read the diff and commit messages — do **not** invent features that aren't in the code
+- One PR per logical change; if the branch contains unrelated commits, surface that and ask before continuing
+- Keep the description focused on reviewer-relevant information (what / why), not a replay of the diff
+- Only check checklist boxes that are actually satisfied; leave the rest unchecked rather than lying
+- If `gh` is not authenticated (`gh auth status` fails), stop and tell the user
+
+$@

.kit/prompts/feature-request.md

@@ -0,0 +1,86 @@
+---
+description: Create a feature request using the GitHub template
+---
+
+Create a feature request for the Kit repository. The user wants to request: $@
+
+## Feature Request Template
+
+This prompt uses the `feature_request` GitHub template which requires:
+
+| Field | Required | Purpose |
+|-------|----------|---------|
+| **Feature Description** | Yes | What should be added or changed |
+| **Motivation / Use Case** | Yes | Why is this needed? What problem does it solve? |
+| **Proposed Implementation** | No | How do you think this should work? |
+
+## Steps
+
+1. **Understand the request** from the user input: $@
+   - What capability is missing?
+   - What would the ideal behavior look like?
+
+2. **Ask clarifying questions** if needed:
+   - "What problem does this solve for you?"
+   - "How would you expect this to work?"
+   - "Are there similar features in other tools you use?"
+
+3. **Craft the title** using conventional format:
+   - `feat: <short description>`
+   - Lowercase, imperative mood, ≤72 chars
+   - Good examples:
+     - `feat: add keyboard shortcut for clearing input`
+     - `feat: support custom themes per extension`
+     - `feat: add fuzzy matching to model selector`
+   - Bad examples:
+     - `Feature request: can we have...` (too vague)
+     - `It would be nice if...` (not imperative)
+
+4. **Build the body** with the template fields:
+
+   **Feature Description:**
+   - Clear statement of what to add/change
+   - Be specific about the behavior
+   - Include UI/UX details if relevant
+
+   **Motivation / Use Case:**
+   - What problem does this solve?
+   - Current workaround (if any) and why it's insufficient
+   - Who benefits from this feature?
+
+   **Proposed Implementation** (optional but helpful):
+   - High-level approach
+   - API changes if applicable
+   - Example usage code
+
+5. **Create the issue**:
+   ```bash
+   gh issue create --template feature_request --title "feat: ..." --body "..."
+   ```
+
+6. **Confirm success**:
+   - Show the issue URL and number
+   - Mention it was created with the feature_request template
+
+## Guidelines
+
+- Focus on the *problem* first, then the solution
+- Include concrete examples of how the feature would be used
+- Consider edge cases and mention them
+- If proposing API changes, show before/after code
+- Check if similar features exist in related tools (mention them for reference)
+- Align with Kit's philosophy: TUI-first, extension-based, keyboard-driven
+
+## Example
+
+User: `/feature-request I want to be able to customize tool border colors dynamically`
+
+You:
+1. Title: `feat: dynamic border colors for tool results based on status`
+2. Body:
+   - **Feature Description**: Allow `ToolRenderConfig` to accept a function that determines border color based on tool result content or status, enabling dynamic visual feedback.
+   - **Motivation**: When running multiple tools, it's hard to distinguish file reads (blue), shell commands (green), and errors (red) without custom colors per result.
+   - **Proposed Implementation**: Add `BorderColorFunc` callback that receives `(result string, isError bool)` and returns a color string.
+
+3. Execute: `gh issue create --template feature_request --title "feat: ..." --body "..."`
+4. Confirm: Created issue #43 using feature_request template

.kit/prompts/file-issue.md

@@ -0,0 +1,100 @@
+---
+description: File a GitHub issue using the appropriate template
+---
+
+File a GitHub issue for the Kit repository. The user wants to create an issue about: $@
+
+## Issue Templates Available
+
+This repository has structured issue templates. You MUST use the appropriate template:
+
+| Type | Template | Use For |
+|------|----------|---------|
+| `bug` | `bug_report` | Something is broken, not working as expected |
+| `feat` | `feature_request` | New feature, enhancement, improvement |
+| `docs` | `documentation` | Missing, incorrect, or unclear documentation |
+
+## Steps
+
+1. **Determine the issue type** from the user input: $@
+   - Bug → use `--template bug_report`
+   - Feature → use `--template feature_request`  
+   - Documentation → use `--template documentation`
+
+2. **Ask clarifying questions** if critical info is missing:
+   - For bugs: "What were you doing when this happened?" (reproduction steps)
+   - For features: "What problem does this solve?" (motivation)
+   - For docs: "Where did you look for this information?" (location)
+
+3. **Craft the title** using conventional format:
+   - `<type>: <short description>`
+   - Lowercase, imperative mood, ≤72 chars
+   - Examples:
+     - `fix: ToolRenderConfig BorderColor ignored during rendering`
+     - `feat: add keyboard shortcut for clearing input`
+     - `docs: clarify extension widget lifecycle`
+
+4. **File the issue** using the template:
+   ```bash
+   # For bugs
+   gh issue create --template bug_report --title "fix: ..." --body "..."
+   
+   # For features
+   gh issue create --template feature_request --title "feat: ..." --body "..."
+   
+   # For documentation
+   gh issue create --template documentation --title "docs: ..." --body "..."
+   ```
+
+   The template will guide the user through the required fields. You need to provide:
+   - **Bug reports**: Description, reproduction steps, expected vs actual behavior
+   - **Feature requests**: Description, motivation/use case, optional proposed implementation
+   - **Documentation**: Description, location of docs, suggested improvement
+
+5. **Confirm success** by showing:
+   - The issue URL
+   - The issue number
+   - Which template was used
+
+## Template Field Guide
+
+### Bug Report (`bug_report`)
+Required fields in the body:
+- **Bug Description** - what happened vs expected
+- **Steps to Reproduce** - numbered list to recreate the bug
+- **Relevant Code** - code snippets, configuration, error messages
+- **Component** - which part of Kit (ui, extensions, session, etc.)
+- **Version** - Kit version or commit hash
+
+### Feature Request (`feature_request`)
+Required fields in the body:
+- **Feature Description** - what to add/change
+- **Motivation / Use Case** - why this is needed
+- **Proposed Implementation** - how it could work (optional)
+
+### Documentation (`documentation`)
+Required fields in the body:
+- **Documentation Issue** - what's wrong or missing
+- **Documentation Location** - file or URL where docs exist
+- **Suggested Improvement** - how to fix the docs
+
+## Guidelines
+
+- ALWAYS use `--template <name>` instead of bare `gh issue create`
+- Include file paths and line numbers when you know them
+- Use triple backticks for code blocks
+- Keep the body factual - avoid speculation unless in "Proposed Fix" section
+- If you're unsure about technical details, say so in the issue
+- For UI bugs, describe what you see vs what you expect
+- For API bugs, include the relevant struct/function names
+
+## Example Usage
+
+User: `/file-issue The ToolRenderConfig BorderColor field is documented but never used in rendering`
+
+You: 
+1. Determine this is a **bug** (documented field doesn't work)
+2. Use `--template bug_report`
+3. Gather: reproduction steps (register renderer with BorderColor), expected (custom color), actual (default color)
+4. Create issue with title `fix: ToolRenderConfig BorderColor and Background fields are ignored`
+5. Confirm: Created issue #42 using bug_report template

.kit/prompts/fix-issue.md

@@ -0,0 +1,61 @@
+---
+description: Implement the fix/feature/docs change requested by a GitHub issue
+---
+
+Resolve GitHub issue #$1 by reading it, classifying it, and producing the appropriate code or doc change. **Stop once the working tree contains the change** — committing, pushing, and opening a PR are handled by `/commit-push` and `/create-pr`.
+
+## Steps
+
+1. **Fetch the issue**:
+   - Run: gh issue view $1 --json number,title,body,labels,state,author,comments
+   - If the issue is closed, stop and ask the user whether to proceed
+   - Read the **entire** thread including comments — the latest comment often refines the ask
+
+2. **Classify the issue** from labels, title prefix, and body content:
+   - `bug` / `fix:` → reproduce, then fix
+   - `enhancement` / `feature` / `feat:` → design, then implement
+   - `documentation` / `docs:` → locate and update docs
+   - `question` / `discussion` → answer in a comment, do **not** write code
+   - Anything else → ask the user how to proceed
+
+3. **Create a working branch** off the default branch:
+   - `git checkout main && git pull --ff-only`
+   - Branch name: <type>/$1-<slug> (e.g. `fix/42-borderColor-ignored`, `feat/57-keyboard-clear`, `docs/63-widget-lifecycle`)
+
+4. **Do the work** based on type:
+
+   ### Bug (`bug` label / `fix:` title)
+   - Reproduce the failure first (write a failing test if feasible) — if you cannot reproduce, comment on the issue asking for clarification and stop
+   - Locate the root cause; do not patch symptoms
+   - Add or extend a regression test that fails before and passes after the fix
+   - Run `go test ./... -race` and `golangci-lint run`
+
+   ### Feature (`enhancement` / `feature` label / `feat:` title)
+   - Re-read the motivation and proposed implementation in the issue body
+   - For large, ambiguous, or breaking changes, sketch the design in a comment on the issue and wait for sign-off before writing code
+   - Implement behind sensible defaults; add godoc on every exported symbol
+   - Add unit tests covering the new behaviour and edge cases
+   - Update `README.md` / `docs/` if the public surface changed
+   - Run `go test ./... -race` and `golangci-lint run`
+
+   ### Documentation (`documentation` label / `docs:` title)
+   - Open the file/URL referenced in the issue's "Documentation Location"
+   - Apply the suggested improvement; verify code samples compile (`go build ./...`)
+   - No tests required, but run `golangci-lint run` if Go files were touched
+
+5. **Report**:
+   - Branch name (`git branch --show-current`)
+   - Summary of files changed (`git status -s`) and the diff highlights
+   - Test/lint results (pass/fail with key output)
+   - Suggest the next step explicitly:
+     - `/commit-push` to commit with a Conventional Commit subject (the message should reference `(#$1)` and include `Fixes #$1` so merge auto-closes)
+     - then `/create-pr $1` to open the pull request
+
+## Guidelines
+
+- This prompt **stops at a clean working tree with the change applied** — do not run `git commit`, `git push`, or `gh pr create`
+- If the issue is unclear, post a clarifying comment on the issue and stop; do not guess
+- Keep the change scoped to the issue; surface unrelated cleanups separately
+- For breaking changes or architecture shifts, propose the design on the issue first and wait for maintainer sign-off
+- If the issue is a duplicate or already fixed on `main`, comment with the reference and stop
+- Do not close the issue manually — the eventual PR's `Fixes #$1` handles that on merge

.kit/prompts/new-prompt.md

@@ -0,0 +1,84 @@
+---
+description: Scaffold a new prompt template in .kit/prompts/
+---
+
+Create a new kit prompt template. The user wants a prompt that does: $@
+
+## What a prompt template is
+
+A prompt template is a `.md` file in `.kit/prompts/` (project-local) or `~/.kit/prompts/` (global).
+It becomes a `/slug` slash command in the kit input box — typed as `/filename` with optional arguments.
+
+## File format
+
+```
+---
+description: One-line description shown in autocomplete
+---
+
+Body text of the prompt. Reference user-supplied arguments
+with positional placeholders (see "Argument placeholders" below).
+```
+
+- **Filename** → slug: `commit-push.md` becomes `/commit-push`
+- **Frontmatter**: only `description` is recognised; keep it under ~80 chars
+- **Body**: plain markdown; the full text is submitted as the user's message when the template fires
+- **Required args**: kit infers required positional args from the highest `$N` it finds *outside* backtick/tilde code fences — a stray `$2` in active prose means kit will refuse to run without 2 arguments
+
+## Argument placeholders
+
+kit performs shell-style substitution before sending the prompt to the model:
+
+- `$1`, `$2`, … — positional arguments (1-indexed)
+- `${1}`, `${2}`, … — same, brace form (use when followed by digits/letters: `${1}_suffix`)
+- `$@` — all arguments joined by spaces (zero or more, optional)
+- `$+` — all arguments, **at least one required**
+- `$ARGUMENTS` / `${ARGUMENTS}` — alias for `$@`
+- `${@:N}` — args from the Nth onwards (1-indexed, bash-style)
+- `${@:N:L}` — `L` args starting from the Nth
+
+### ⚠️ Critical: code fences and inline code preserve placeholders verbatim
+
+Anything inside triple-backtick fences, `~~~` fences, or single-backtick `inline` code spans is **left untouched** so example code samples don't get corrupted. That means:
+
+- An inline-coded `gh issue view $1` stays literal `$1` in the model's input ❌
+- The same command without backticks: gh issue view $1 → expands to `gh issue view 42` ✓
+
+**Rule of thumb:** if you want a placeholder to substitute, keep it outside backticks and fences. If you want a literal `$1` in the output (e.g. teaching the user shell syntax), put it inside backticks.
+
+### Workarounds for "I want it to look like code AND substitute"
+
+1. **Drop the backticks** around just the placeholder portion — the rest can still read as a command line in prose
+2. **Use a 4-space-indented code block** instead of a triple-backtick fence — kit only skips backtick/tilde fences, so indentation-style code blocks still get substitution:
+
+       git push -u origin "$(git branch --show-current)"
+       gh pr create --title "fix: ... (#$1)" --base main
+
+3. **Bind once, reference loosely**: put `Issue: $1` at the top in prose, then leave the backticked examples literal — the model will substitute mentally
+
+## Steps
+
+1. **Understand the workflow** the user described in $@ — ask a clarifying question if the intent is ambiguous
+2. **Choose a filename**: short, lowercase, hyphen-separated, descriptive (e.g. `code-review.md`)
+3. **Write the description**: one sentence, imperative, fits in autocomplete
+4. **Decide on arguments**:
+   - No args needed → omit placeholders entirely
+   - One required value (issue number, PR url, file path) → use `$1`
+   - Free-form trailing context → end with a single `$@` line
+   - Multiple distinct values → use `$1`, `$2`, … and document each at the top
+5. **Draft the body**:
+   - Open with a single sentence stating the goal, weaving in `$1`/`$@` where the value belongs
+   - Use `## Steps` for multi-step workflows; use plain prose for simple prompts
+   - Be specific: name commands, flags, and file paths where relevant
+   - **Audit every backtick and code fence**: any `$N` or `$@` inside them will not expand — was that intentional? If not, apply one of the workarounds above
+6. **Write the file** to `.kit/prompts/<slug>.md`
+7. **Verify substitution** by mentally (or actually) replacing `$1`/`$@` with a sample value and confirming every reference resolves — and that the prompt's *own* example snippets don't accidentally bump the required-arg count (wrap illustrative `$N` examples in triple-backtick fences, not 4-space indentation, so `RequiredArgs()` ignores them)
+8. **Confirm** by showing the final file content and the slash command that activates it (e.g. `/code-review 42`)
+
+## Guidelines
+
+- Keep prompts action-oriented — they should tell kit *what to do*, not just *what to think about*
+- Prefer concrete steps over vague instructions
+- A prompt that does one thing well beats one that tries to cover every edge case
+- If the workflow already exists as a prompt, suggest extending it instead of duplicating
+- When in doubt about substitution behaviour, write the file and run `/<slug> testvalue` once to confirm — wrong placement of backticks is the #1 failure mode

.kit/prompts/release-tagger.md

@@ -0,0 +1,70 @@
+---
+description: Semantic version tagging workflow - analyzes commits and tags releases
+---
+
+# Release Tagging Workflow
+
+Tag a new version of this Go project following semantic versioning.
+
+## Steps
+
+1. **Fetch remote tags**: `git fetch --tags origin`
+
+2. **Find latest version**: `git tag -l | sort -V | tail -5` to see recent tags
+
+3. **Analyze changes since last tag**:
+   - `git log <latest-tag>..HEAD --oneline` - list commits
+   - `git diff <latest-tag>..HEAD --stat` - see file stats
+   - `git diff <latest-tag>..HEAD --name-only` - see changed files
+
+4. **Determine version bump** (Semantic Versioning):
+   - **MAJOR (X.0.0)**: Breaking API changes, incompatible modifications
+   - **MINOR (0.X.0)**: New features, backward-compatible additions
+   - **PATCH (0.0.X)**: Bug fixes, backward-compatible fixes
+   
+   Look for indicators:
+   - `feat:` or `feature:` commits → MINOR
+   - `fix:` or `bugfix:` commits → PATCH
+   - `breaking:` or `BREAKING CHANGE:` → MAJOR
+   - Breaking API changes in `pkg/` or public interfaces → MAJOR
+   - New commands, flags, or features → MINOR
+   - Documentation-only changes → PATCH (or skip)
+
+5. **Calculate new version**: Increment appropriate segment, reset lower segments to 0
+
+6. **Draft tag message**:
+   - Summarize key changes from commits
+   - Group by type (Features, Fixes, Breaking Changes)
+   - Keep concise but informative
+
+7. **Create annotated tag**: `git tag -a vX.Y.Z -m "vX.Y.Z - <summary>\n\n<detailed list>"`
+
+8. **Push tag**: `git push origin vX.Y.Z`
+
+## Guidelines
+
+- Always fetch remote tags first to avoid conflicts
+- Use annotated tags (`-a`) with descriptive messages
+- Follow semver strictly - when in doubt, prefer conservative bump (patch over minor)
+- For Go projects, changes to `pkg/` or exported APIs warrant careful version consideration
+- If no changes since last tag, suggest skipping the release
+- Include commit summaries in the tag message body
+
+## Example Tag Message Format
+
+```
+v0.30.1 - Bug fixes for model handling and UI improvements
+
+Fixes:
+- Properly handle think tags from Qwen/DeepSeek models
+- Handle custom provider model persistence and bare model names
+
+Improvements:
+- UI style refactoring and cleanup
+```
+
+Wait for the user to confirm the version and message before executing tag commands.
+
+---
+
+$@

.kit/prompts/update-docs.md

@@ -0,0 +1,52 @@
+---
+description: Audit and update project documentation (README and docs site) for a recent change
+---
+
+Review recent code changes, identify all documentation surfaces that should
+mention them, and update each one — grounded in the actual diff, not guesses.
+
+## Steps
+
+1. **Identify the change**:
+   - If the user input ($@) names a commit / PR / branch / topic, use that as the focus
+   - Otherwise inspect `git log origin/main..HEAD --oneline` and `git diff origin/main...HEAD --stat` to discover what shipped on the current branch
+   - Read the actual diff (`git diff origin/main...HEAD`) — never document features that aren't in the code
+
+2. **Inventory the doc surfaces**:
+   - `README.md` at the repo root
+   - Any docs site (commonly `www/`, `docs/`, `site/`) — list its pages and identify the one(s) most thematically related to the change
+   - Inline godoc / API reference comments on the new exported symbols
+   - `CHANGELOG.md` if the project keeps one
+   - Any `examples/` directory entries that demonstrate the affected area
+
+3. **Audit each surface** with `grep`:
+   - Search for the names of related existing APIs (e.g. if you added `IterTools`, grep for `ListTools`) to find every page that already discusses the area
+   - Decide for each hit: does it need a cross-reference, a side-by-side comparison, or to stay untouched?
+
+4. **Decide where new content lives**:
+   - Prefer extending an existing page over creating a new one
+   - For a docs site, place new sections near related content (check the page's `## Heading` outline first)
+   - Skip surfaces that genuinely don't apply (e.g. a server-focused README for a client-only change) and say so explicitly
+
+5. **Draft the updates**:
+   - Lead with a one-sentence statement of what's new and why
+   - Show concrete code examples copied from real signatures — verify against the source files
+   - Include a comparison / "when to use which" table when adding an alternative to an existing API
+   - Note backwards-compatibility behaviour if relevant
+
+6. **Verify the docs build** before committing:
+   - For vocs / docusaurus / mkdocs sites, run the local build command (e.g. `npx vocs build`, `mkdocs build`) and fix any MDX/markdown errors
+   - For godoc, run `go vet ./...` and `go doc <pkg> <Symbol>` to sanity-check rendering
+
+7. **Report**:
+   - List every file changed and every file deliberately left alone (with a one-line reason)
+   - Suggest the next step (typically `/commit-push`) — do not auto-commit unless asked
+
+## Guidelines
+
+- Read the diff before writing anything — invented API names erode trust faster than missing docs
+- One change per doc commit; keep doc updates separate from code changes when possible
+- Match the existing voice and formatting of each surface (headings, code-fence languages, table styles)
+- Prefer linking between pages over duplicating content
+
+$@

.opencode/opencode.jsonc

@@ -1,8 +0,0 @@
-{
-  "$schema": "https://opencode.ai/config.json",
-  "permission": {
-    "external_directory": {
-      "~/go/**": "deny"
-    }
-  }
-}

AGENTS.md

@@ -100,3 +100,21 @@ Positional args are the prompt. `@file` args attach file content. Key flags: `--
 - Never guess or manually search the filesystem for external projects
 - Example: `btca ask -r https://github.com/user/repo -q "How does X work?"`
 - See `.agents/skills/btca-cli/SKILL.md` for full btca usage
+
+## BTCA Configured Resources
+The following external repositories are configured in `btca.config.jsonc` for research:
+
+- bubbletea
+- lipgloss
+- bubbles
+- glamour
+- fantasy
+- catwalk
+- crush
+- pi
+- iteratr
+- yaegi
+- acp-go-sdk
+- opencode
+- herald
+- herald-md

README.md

@@ -18,7 +18,8 @@ A powerful, extensible AI coding agent CLI with multi-provider support, built-in
 ## Features
 
 - **Multi-Provider LLM Support**: Anthropic, OpenAI, Google Gemini, Ollama, Azure OpenAI, AWS Bedrock, OpenRouter, and more
-- **Built-in Core Tools**: bash, read, write, edit, grep, find, ls, subagent - no MCP overhead
+- **Built-in Core Tools**: bash (with interactive sudo password prompt), read, write, edit, grep, find, ls, subagent - no MCP overhead
+- **Smart @ Attachments**: Binary files auto-detected via MIME type, MCP resources via `@mcp:server:uri`
 - **MCP Integration**: Connect external MCP servers for expanded capabilities
 - **Extension System**: Write custom tools, commands, widgets, and UI modifications in Go
 - **Theming**: 22 built-in color themes (KITT, Catppuccin, Dracula, Nord, etc.) with runtime switching, persistence, and custom theme files
@@ -28,7 +29,7 @@ A powerful, extensible AI coding agent CLI with multi-provider support, built-in
 - **Session Management**: Tree-based conversation history with branching support
 - **Non-Interactive Mode**: Script-friendly positional args with JSON output
 - **ACP Server**: Run Kit as an [Agent Client Protocol](https://agentclientprotocol.com) agent over stdio
-- **Go SDK**: Embed Kit in your own applications
+- **Go SDK**: Embed Kit in your own applications with full agent lifecycle events (30+ event types) and behavior-modifying hooks
 
 ## Installation
 
@@ -125,8 +126,14 @@ model: anthropic/claude-sonnet-latest
 max-tokens: 4096
 temperature: 0.7
 stream: true
+thinking-level: off       # off, none, minimal, low, medium, high
+no-core-tools: false      # set to true to disable all built-in core tools
 ```
 
+All of the above keys can also be set programmatically via the SDK
+(`kit.Options.MaxTokens`, `Options.Temperature`, `Options.ThinkingLevel`, etc.)
+without touching config files — see [SDK options](#with-options).
+
 ### Environment Variables
 
 ```bash
@@ -151,6 +158,16 @@ mcpServers:
   search:
     type: remote
     url: "https://mcp.example.com/search"
+
+  pubmed:
+    type: remote
+    url: "https://pubmed.mcp.example.com"
+    noOAuth: true  # skip OAuth for public servers that don't require auth
+
+  builds:
+    type: remote
+    url: "https://builds.mcp.example.com"
+    tasksMode: always  # async task execution — see MCP Tasks below
 ```
 
 ## CLI Reference
@@ -179,19 +196,22 @@ mcpServers:
 --compact                Enable compact output mode
 --auto-compact           Auto-compact conversation near context limit
 
-# Extensions
+# Extensions and tools
 --extension, -e          Load additional extension file(s) (repeatable)
 --no-extensions          Disable all extensions
+--no-core-tools          Disable all built-in core tools (bash, read, write, edit, grep, find, ls, subagent)
 --prompt-template        Load a specific prompt template by name
 --no-prompt-templates    Disable prompt template loading
 
 # Generation parameters
---max-tokens             Maximum tokens in response (default: 4096)
+--max-tokens             Maximum tokens in response (default: 8192, auto-raised up to 32768 for models with larger known output limits)
 --temperature            Randomness 0.0-1.0 (default: 0.7)
 --top-p                  Nucleus sampling 0.0-1.0 (default: 0.95)
 --top-k                  Limit top K tokens (default: 40)
 --stop-sequences         Custom stop sequences (comma-separated)
---thinking-level         Extended thinking level: off, minimal, low, medium, high (default: off)
+--frequency-penalty      Penalize frequent tokens 0.0-2.0 (default: 0.0)
+--presence-penalty       Penalize present tokens 0.0-2.0 (default: 0.0)
+--thinking-level         Extended thinking level: off, none, minimal, low, medium, high (default: off)
 
 # System
 --config                 Config file path (default: ~/.kit.yml)
@@ -203,9 +223,10 @@ mcpServers:
 
 ```bash
 # Authentication (for OAuth-enabled providers)
-kit auth login [provider]    # Start OAuth flow (e.g., anthropic)
-kit auth logout [provider]   # Remove credentials for provider
-kit auth status              # Check authentication status
+kit auth login [provider]          # Start OAuth flow (e.g., anthropic)
+kit auth login [provider] --set-default  # Set provider's default model as system default
+kit auth logout [provider]         # Remove credentials for provider
+kit auth status                    # Check authentication status
 
 # Model database
 kit models [provider]        # List available models (optionally filter by provider)
@@ -287,7 +308,7 @@ kit -e examples/extensions/minimal.go
 
 ### Extension Capabilities
 
-**Lifecycle Events**: OnSessionStart, OnSessionShutdown, OnBeforeAgentStart, OnAgentStart, OnAgentEnd, OnToolCall, OnToolExecutionStart, OnToolOutput, OnToolExecutionEnd, OnToolResult, OnInput, OnMessageStart, OnMessageUpdate, OnMessageEnd, OnModelChange, OnContextPrepare, OnBeforeFork, OnBeforeSessionSwitch, OnBeforeCompact, OnCustomEvent, OnSubagentStart, OnSubagentChunk, OnSubagentEnd
+**Lifecycle Events**: OnSessionStart, OnSessionShutdown, OnBeforeAgentStart, OnAgentStart, OnAgentEnd, OnToolCall, OnToolCallInputStart, OnToolCallInputDelta, OnToolCallInputEnd, OnToolExecutionStart, OnToolOutput, OnToolExecutionEnd, OnToolResult, OnInput, OnMessageStart, OnMessageUpdate, OnMessageEnd, OnModelChange, OnContextPrepare, OnBeforeFork, OnBeforeSessionSwitch, OnBeforeCompact, OnCustomEvent, OnSubagentStart, OnSubagentChunk, OnSubagentEnd
 
 **Custom Components**:
 - **Tools**: Add new tools the LLM can invoke
@@ -317,39 +338,40 @@ kit -e examples/extensions/minimal.go
 
 See the `examples/extensions/` directory:
 
-- `minimal.go` - Clean UI with custom footer
-- `auto-commit.go` - Auto-commit on shutdown
-- `bookmark.go` - Bookmark conversations
-- `branded-output.go` - Branded output rendering
-- `compact-notify.go` - Notification on compaction
-- `confirm-destructive.go` - Confirm destructive operations
-- `context-inject.go` - Inject context into conversations
-- `conversation-manager.go` - **NEW** Tree navigation, branch summarization, and fresh context loops
-- `custom-editor-demo.go` - Vim-like modal editor
-- `dev-reload.go` - Development live-reload
-- `header-footer-demo.go` - Custom headers and footers
-- `inline-bash.go` - Inline bash execution
-- `interactive-shell.go` - Interactive shell integration
-- `kit-kit.go` - Kit-in-Kit (sub-agent spawning)
-- `lsp-diagnostics.go` - LSP diagnostic integration
-- `notify.go` - Desktop notifications
-- `overlay-demo.go` - Modal dialogs
-- `permission-gate.go` - Permission gating for tools
-- `pirate.go` - Pirate-themed personality
-- `plan-mode.go` - Read-only planning mode
-- `project-rules.go` - Project-specific rules
-- `prompt-demo.go` - Interactive prompts (select/confirm/input)
-- `prompt-templates.go` - **NEW** Frontmatter-driven templates with model switching and skill injection
-- `protected-paths.go` - Path protection for sensitive files
-- `subagent-widget.go` - Multi-agent orchestration with status widget
-- `subagent-test.go` - Subagent testing utilities
-- `summarize.go` - Conversation summarization
-- `tool-logger.go` - Log all tool calls
-- `neon-theme.go` - Custom theme registration and switching
-- `tool-renderer-demo.go` - Custom tool call rendering
-- `widget-status.go` - Persistent status widgets
+- [`minimal.go`](examples/extensions/minimal.go) - Clean UI with custom footer
+- [`auto-commit.go`](examples/extensions/auto-commit.go) - Auto-commit on shutdown
+- [`bookmark.go`](examples/extensions/bookmark.go) - Bookmark conversations
+- [`branded-output.go`](examples/extensions/branded-output.go) - Branded output rendering
+- [`bridge-demo.go`](examples/extensions/bridge_demo.go) - Bridged SDK API demo (tree navigation, skills, templates, model resolution)
+- [`compact-notify.go`](examples/extensions/compact-notify.go) - Notification on compaction
+- [`confirm-destructive.go`](examples/extensions/confirm-destructive.go) - Confirm destructive operations
+- [`context-inject.go`](examples/extensions/context-inject.go) - Inject context into conversations
+- [`conversation-manager.go`](examples/extensions/conversation-manager.go) - **NEW** Tree navigation, branch summarization, and fresh context loops
+- [`custom-editor-demo.go`](examples/extensions/custom-editor-demo.go) - Vim-like modal editor
+- [`dev-reload.go`](examples/extensions/dev-reload.go) - Development live-reload
+- [`header-footer-demo.go`](examples/extensions/header-footer-demo.go) - Custom headers and footers
+- [`inline-bash.go`](examples/extensions/inline-bash.go) - Inline bash execution
+- [`interactive-shell.go`](examples/extensions/interactive-shell.go) - Interactive shell integration
+- [`kit-kit.go`](examples/extensions/kit-kit.go) - Kit-in-Kit (sub-agent spawning)
+- [`lsp-diagnostics.go`](examples/extensions/lsp-diagnostics.go) - LSP diagnostic integration
+- [`notify.go`](examples/extensions/notify.go) - Desktop notifications
+- [`overlay-demo.go`](examples/extensions/overlay-demo.go) - Modal dialogs
+- [`permission-gate.go`](examples/extensions/permission-gate.go) - Permission gating for tools
+- [`pirate.go`](examples/extensions/pirate.go) - Pirate-themed personality
+- [`plan-mode.go`](examples/extensions/plan-mode.go) - Read-only planning mode
+- [`project-rules.go`](examples/extensions/project-rules.go) - Project-specific rules
+- [`prompt-demo.go`](examples/extensions/prompt-demo.go) - Interactive prompts (select/confirm/input)
+- [`prompt-templates.go`](examples/extensions/prompt-templates.go) - **NEW** Frontmatter-driven templates with model switching and skill injection
+- [`protected-paths.go`](examples/extensions/protected-paths.go) - Path protection for sensitive files
+- [`subagent-widget.go`](examples/extensions/subagent-widget.go) - Multi-agent orchestration with status widget
+- [`subagent-test.go`](examples/extensions/subagent-test.go) - Subagent testing utilities
+- [`summarize.go`](examples/extensions/summarize.go) - Conversation summarization
+- [`tool-logger.go`](examples/extensions/tool-logger.go) - Log all tool calls
+- [`neon-theme.go`](examples/extensions/neon-theme.go) - Custom theme registration and switching
+- [`tool-renderer-demo.go`](examples/extensions/tool-renderer-demo.go) - Custom tool call rendering
+- [`widget-status.go`](examples/extensions/widget-status.go) - Persistent status widgets
 
-Also see `.kit/extensions/go-edit-lint.go` (in this repo) for a project-local extension example that runs gopls and golangci-lint on Go file edits.
+Also see [`.kit/extensions/go-edit-lint.go`](.kit/extensions/go-edit-lint.go) (in this repo) for a project-local extension example that runs gopls and golangci-lint on Go file edits.
 
 ### Loading Extensions
 
@@ -406,7 +428,7 @@ func TestMyExtension(t *testing.T) {
 - `AssertPrinted()`, `AssertPrintedContains()` — Verify output
 - `AssertToolRegistered()`, `AssertCommandRegistered()` — Verify registration
 
-See `examples/extensions/tool-logger_test.go` for a complete example with 14 test cases covering tool calls, input handling, and session lifecycle.
+See [`examples/extensions/tool-logger_test.go`](examples/extensions/tool-logger_test.go) for a complete example with 14 test cases covering tool calls, input handling, and session lifecycle.
 
 ### Prompt Templates
 
@@ -428,10 +450,13 @@ Focus on $1 specifically.
 
 **Argument placeholders:**
 - `$1`, `$2`, etc. — Individual arguments
-- `$@` or `$ARGUMENTS` — All arguments
+- `$@` or `$ARGUMENTS` — All arguments (zero or more)
+- `$+` — All arguments (one or more required; error if none given)
 - `${@:2}` — Arguments from position 2 onwards
 - `${@:1:3}` — 3 arguments starting at position 1
 
+Placeholders inside fenced code blocks (```) and inline code spans are ignored.
+
 Disable templates with `--no-prompt-templates` or load a specific template with `--prompt-template <name>`.
 
 ## Session Management
@@ -477,9 +502,18 @@ During an interactive session, use these slash commands:
 | `/import <path>` | Import and switch to a session from a JSONL file |
 | `/share` | Upload session to GitHub Gist and get a shareable viewer URL |
 | `/tree` | Navigate the session tree |
-| `/fork` | Branch from an earlier message |
+| `/fork` | Fork to new session from an earlier message |
 | `/new` | Start a fresh session |
 
+### Keyboard Shortcuts
+
+| Shortcut | Description |
+|----------|-------------|
+| `Ctrl+X e` | Open `$VISUAL`/`$EDITOR` to compose or edit your prompt |
+| `Ctrl+X s` | Steer — inject a system-level instruction mid-turn |
+| `ESC ESC` | Cancel the current operation (tool call or streaming) |
+| `↑` / `↓` | Navigate prompt history |
+
 ## Go SDK
 
 Embed Kit in your Go applications:
@@ -522,16 +556,37 @@ host, err := kit.New(ctx, &kit.Options{
     SystemPrompt: "You are a helpful bot",
     ConfigFile:   "/path/to/config.yml",
     MaxSteps:     10,
-    Streaming:    true,
+    Streaming:    ptr(true), // *bool: nil = unset (default true), &false = off
     Quiet:        true,
 
+    // Generation parameters (override env/config/per-model defaults)
+    MaxTokens:        16384,             // 0 = auto-resolve (env → config → per-model → 8192 floor)
+    ThinkingLevel:    "medium",          // "off", "none", "minimal", "low", "medium", "high"
+    Temperature:      ptr(float32(0.2)), // pointer so 0.0 != unset; nil = provider default
+    TopP:             nil,                // nil = leave provider/per-model default
+    TopK:             nil,
+    FrequencyPenalty: nil,
+    PresencePenalty:  nil,
+
+    // Provider configuration (override env/config without reaching into viper)
+    ProviderAPIKey: "sk-...",                      // "" = use config / provider env var
+    ProviderURL:    "https://proxy.internal/v1",   // "" = provider default
+    TLSSkipVerify:  false,                         // only takes effect when true
+
     // Session options
     SessionPath:  "./session.jsonl",  // Open specific session
     Continue:     true,                // Resume most recent session
     NoSession:    true,                // Ephemeral mode
 
     // Tool options
-    ExtraTools:   []kit.Tool{...},     // Additional tools alongside defaults
+    Tools:            []kit.Tool{...},     // Replace default tool set entirely
+    ExtraTools:       []kit.Tool{...},     // Add tools alongside defaults
+    DisableCoreTools: true,                // Disable all built-in core tools; also controllable via
+                                           // --no-core-tools flag, KIT_NO_CORE_TOOLS env var,
+                                           // or no-core-tools: true in .kit.yml
+
+    // Configuration
+    SkipConfig:   true,                   // Skip .kit.yml files (viper defaults + env vars still apply)
 
     // Compaction
     AutoCompact:  true,                // Auto-compact near context limit
@@ -540,6 +595,151 @@ host, err := kit.New(ctx, &kit.Options{
 })
 ```
 
+**Generation & provider fields** (added in v0.55+) let SDK consumers configure
+Kit entirely in-code without `viper.Set()` workarounds or shipping a `.kit.yml`.
+Precedence is `Options` > `KIT_*` env vars > `.kit.yml` > per-model defaults
+(`modelSettings` / `customModels`) > provider-level defaults. Sampling params
+are pointer types so explicit `0.0` is distinguishable from "leave alone"; a
+non-zero `MaxTokens` suppresses automatic right-sizing the same way `--max-tokens`
+does on the CLI.
+
+### Functional options (`NewAgent`)
+
+For simple programmatic setups, `kit.NewAgent` offers an ergonomic
+functional-options front door over `kit.New`. Streaming is **enabled by
+default**; pass `kit.WithStreaming(false)` to opt out.
+
+```go
+host, err := kit.NewAgent(ctx,
+    kit.WithModel("anthropic/claude-sonnet-4-5-20250929"),
+    kit.WithSystemPrompt("You are a helpful assistant."),
+    kit.WithMaxTokens(8192),
+    kit.WithThinkingLevel("medium"),
+    kit.Ephemeral(), // in-memory session, no persistence
+)
+```
+
+Available options: `WithModel`, `WithSystemPrompt`, `WithStreaming`,
+`WithMaxTokens`, `WithThinkingLevel`, `WithTools`, `WithExtraTools`,
+`WithProviderAPIKey`, `WithProviderURL`, `WithConfigFile`, `WithDebug`, and
+`Ephemeral`. For advanced configuration not covered by the helpers (custom MCP
+config, in-process MCP servers, session backends, MCP task tuning) construct an
+`Options` value explicitly and call `kit.New`.
+
+### Per-instance config isolation
+
+Each `kit.New` / `kit.NewAgent` call owns an **isolated configuration store**,
+so constructing multiple Kit instances in the same process is safe: setting the
+model, thinking level, or generation parameters on one never affects another,
+and runtime mutators (`SetModel`, `SetThinkingLevel`) only touch the owning
+instance. This makes subagent spawning and multi-Kit embedding race-free with
+no external synchronization required.
+
+### MCP OAuth (remote MCP servers)
+
+When a remote MCP server returns 401, Kit runs the full OAuth flow (dynamic
+client registration → PKCE → token exchange → persistence) but delegates the
+user-facing step — showing the authorization URL and receiving the callback —
+to an `MCPAuthHandler` that you pass explicitly via `Options.MCPAuthHandler`.
+If nil, OAuth is disabled and the authorization-required error surfaces to the
+caller; the SDK never auto-opens a browser or binds a localhost port.
+
+```go
+// CLI/TUI apps: opens the system browser + prints status to stderr.
+authHandler, _ := kit.NewCLIMCPAuthHandler()
+defer authHandler.Close()
+
+host, _ := kit.New(ctx, &kit.Options{
+    MCPAuthHandler: authHandler,
+})
+
+// Custom UX: reuse the SDK's port + callback server, supply your own
+// presentation via OnAuthURL (TUI modal, QR code, web redirect, etc.).
+//   h, _ := kit.NewDefaultMCPAuthHandler()
+//   h.OnAuthURL = func(server, authURL string) { myUI.Show(server, authURL) }
+//
+// Full control (web apps, daemons): implement kit.MCPAuthHandler yourself —
+// no localhost binding, no side effects.
+```
+
+Tokens are persisted to `$XDG_CONFIG_HOME/.kit/mcp_tokens.json` by default; swap
+in a custom `MCPTokenStoreFactory` for encrypted, DB-backed, or in-memory
+storage. See the [SDK options docs](/sdk/options#mcp-oauth-authorization) for
+the full matrix.
+
+### MCP Tasks (long-running tools)
+
+Kit advertises [MCP task support](https://modelcontextprotocol.io/specification/2025-11-25/basic/utilities/tasks)
+during `initialize`, so cooperating MCP servers can respond to `tools/call`
+with a `taskId` instead of blocking the connection. Kit then polls
+`tasks/get` / `tasks/result` until the task reaches a terminal state, and
+best-effort `tasks/cancel`s on context cancellation.
+
+Defaults are safe — a server that doesn't advertise task capability runs
+synchronously, exactly as before. Opt in per server via `tasksMode` in
+`.kit.yml` (`auto` | `never` | `always`) or programmatically through the SDK:
+
+```go
+host, _ := kit.New(ctx, &kit.Options{
+    MCPTaskMode: map[string]kit.MCPTaskMode{
+        "build-server": kit.MCPTaskModeAlways,
+    },
+    MCPTaskTimeout:  15 * time.Minute,
+    MCPTaskProgress: func(p kit.MCPTaskProgress) {
+        log.Printf("%s: %s", p.TaskID, p.Status)
+    },
+})
+
+tasks, _ := host.ListMCPTasks(ctx, "build-server")
+_, _    = host.CancelMCPTask(ctx, "build-server", tasks[0].TaskID)
+```
+
+See the [configuration docs](/configuration#mcp-tasks-long-running-tools) and
+[SDK options → MCP Tasks](/sdk/options#mcp-tasks) for the full surface.
+
+### Custom Tools
+
+Create custom tools with automatic schema generation — no external dependencies needed:
+
+```go
+type SearchInput struct {
+    Query string `json:"query" description:"Search query"`
+}
+
+searchTool := kit.NewTool("search", "Search the codebase",
+    func(ctx context.Context, input SearchInput) (kit.ToolOutput, error) {
+        return kit.TextResult("Found: ..."), nil
+    },
+)
+
+host, _ := kit.New(ctx, &kit.Options{
+    ExtraTools: []kit.Tool{searchTool}, // adds alongside built-in tools
+})
+```
+
+Use `kit.NewParallelTool` for tools safe to run concurrently. Binary data (images, audio, etc.) in `ToolOutput.Data` is automatically forwarded to the LLM when `MediaType` is set. See the [SDK docs](/sdk/overview) for full details on struct tags, `ToolOutput` fields, and `ToolCallIDFromContext`.
+
+#### Return Helpers
+
+| Helper | Description |
+| --- | --- |
+| `kit.TextResult(content)` | Successful text result |
+| `kit.ErrorResult(content)` | Error result (LLM sees it as a tool error) |
+| `kit.ImageResult(content, data, mediaType)` | Image result with binary data (e.g. `"image/png"`) |
+| `kit.MediaResult(content, data, mediaType)` | Non-image media result (e.g. `"audio/mpeg"`) |
+
+#### ToolOutput Fields
+
+```go
+kit.ToolOutput{
+    Content:   "result text",     // text returned to the LLM
+    IsError:   false,             // true = LLM sees this as an error
+    Data:      pngBytes,          // optional binary data (images, audio)
+    MediaType: "image/png",       // MIME type for binary Data
+    Metadata:  map[string]any{},  // opaque metadata for hooks/UI (not sent to LLM)
+}
+```
+
 ### With Callbacks
 
 ```go
@@ -555,7 +755,7 @@ unsub2 := host.OnToolResult(func(e kit.ToolResultEvent) {
 })
 defer unsub2()
 
-unsub3 := host.OnStreaming(func(e kit.MessageUpdateEvent) {
+unsub3 := host.OnMessageUpdate(func(e kit.MessageUpdateEvent) {
     print(e.Chunk)
 })
 defer unsub3()
@@ -592,6 +792,45 @@ host, _ := kit.New(ctx, &kit.Options{
 })
 ```
 
+### Runtime Skills & Context Files
+
+For multi-tenant hosts (chatbots, per-user agents, web services), the SDK
+lets you swap skills and `AGENTS.md`-style context files **after** Kit
+construction. Every mutation recomposes the system prompt and applies it to
+the agent so the next turn picks up the new instructions — no restart needed.
+
+```go
+// Programmatic skill (no file on disk required).
+host.AddSkill(&kit.Skill{
+    Name:        "polite-french",
+    Description: "Respond in French and always greet the user.",
+    Content:     "Always reply in French. Open every response with 'Bonjour'.",
+})
+
+// Or load one from disk.
+host.LoadAndAddSkill("/var/skills/refund-policy.md")
+
+// Per-user AGENTS.md content pulled from a database.
+host.AddContextFileContent(
+    fmt.Sprintf("session://%s/AGENTS.md", userID),
+    rulesFromDB,
+)
+
+// Tear down session-specific state on logout.
+host.RemoveSkill("polite-french")
+host.RemoveContextFile(fmt.Sprintf("session://%s/AGENTS.md", userID))
+
+// Or replace the whole set atomically.
+host.SetSkills(activeSkillsForUser)
+host.SetContextFiles(activeContextForUser)
+```
+
+Skills dedupe by `Name`, context files dedupe by `Path` (which can be any
+opaque identifier — it doesn't have to be a real filesystem path). All
+mutators and readers (`GetSkills`, `GetContextFiles`) are safe to call
+concurrently from multiple goroutines. See the [SDK overview docs](/sdk/overview#runtime-skills-and-context-files)
+for the full reference.
+
 ## Advanced Usage
 
 ### Subagent Pattern
@@ -733,6 +972,31 @@ This automatically defaults to `custom/custom` without needing to specify a mode
 - Reasoning and temperature support
 - Optional `CUSTOM_API_KEY` environment variable or `--provider-api-key` flag
 
+### Auto-routed Providers
+
+Any provider in the [models.dev](https://models.dev) database can be used as
+`provider/model` without a dedicated native integration. Kit auto-routes the
+request through the matching **wire protocol** based on the provider's npm package
+(or per-model override), using its `api` URL as the base:
+
+| npm package | Wire protocol |
+|-------------|---------------|
+| `@ai-sdk/openai` | OpenAI (Responses API) |
+| `@ai-sdk/openai-compatible` | OpenAI (chat completions) |
+| `@ai-sdk/anthropic` | Anthropic |
+| `@ai-sdk/google` | Google Gemini |
+
+Providers with an `api` URL but an unrecognized npm package fall back to the
+OpenAI-compatible wire. Because routing follows the wire protocol, aggregator/proxy
+providers work across all of their models — including Claude, GPT, *and* Gemini
+routes:
+
+```bash
+kit --model opencode/claude-haiku-4-5 "Hello"     # → Anthropic wire
+kit --model opencode/gpt-5 "Hello"                # → OpenAI wire
+kit --model opencode/gemini-3.5-flash "Hello"     # → Google wire
+```
+
 ### Model String Format
 
 ```bash

cmd/acp.go

@@ -11,6 +11,7 @@ import (
 	"os/signal"
 	"syscall"
 
+	"github.com/charmbracelet/log"
 	acp "github.com/coder/acp-go-sdk"
 
 	"github.com/mark3labs/kit/internal/acpserver"
@@ -54,6 +55,8 @@ func runACP(cmd *cobra.Command, _ []string) error {
 		conn.SetLogger(slog.New(slog.NewTextHandler(os.Stderr, &slog.HandlerOptions{
 			Level: slog.LevelDebug,
 		})))
+		// Also set charmbracelet/log level for acpserver package logging
+		log.SetLevel(log.DebugLevel)
 	}
 
 	// Wait for either the client to disconnect or a signal.

cmd/auth.go

@@ -11,6 +11,7 @@ import (
 
 	"charm.land/huh/v2"
 	"github.com/mark3labs/kit/internal/auth"
+	"github.com/mark3labs/kit/internal/ui"
 	kit "github.com/mark3labs/kit/pkg/kit"
 	"github.com/spf13/cobra"
 )
@@ -54,9 +55,13 @@ Available providers:
   - anthropic: Anthropic Claude API (OAuth)
   - openai:    OpenAI ChatGPT Plus/Pro (Codex OAuth)
 
-Example:
+Flags:
+  --set-default   Set this provider's default model as the system default
+
+Examples:
   kit auth login anthropic
-  kit auth login openai`,
+  kit auth login openai
+  kit auth login openai --set-default`,
 	Args: cobra.ExactArgs(1),
 	RunE: runAuthLogin,
 }
@@ -99,10 +104,43 @@ Example:
 	RunE: runAuthStatus,
 }
 
+var (
+	loginSetDefault bool
+)
+
+// defaultModels maps providers to their recommended default models.
+// These are used when --set-default flag is passed to auth login.
+var defaultModels = map[string]string{
+	"anthropic": "anthropic/claude-sonnet-4-5-20250929",
+	"openai":    "openai/gpt-5.4",
+}
+
+// setDefaultModelIfRequested sets the default model for the given provider
+// if the --set-default flag was provided.
+func setDefaultModelIfRequested(provider string) error {
+	if !loginSetDefault {
+		return nil
+	}
+
+	model, ok := defaultModels[provider]
+	if !ok {
+		return fmt.Errorf("no default model configured for provider: %s", provider)
+	}
+
+	if err := ui.SaveModelPreference(model); err != nil {
+		return fmt.Errorf("failed to save model preference: %w", err)
+	}
+
+	fmt.Printf("\n✓ Set default model to: %s\n", model)
+	return nil
+}
+
 func init() {
 	authCmd.AddCommand(authLoginCmd)
 	authCmd.AddCommand(authLogoutCmd)
 	authCmd.AddCommand(authStatusCmd)
+
+	authLoginCmd.Flags().BoolVar(&loginSetDefault, "set-default", false, "Set this provider's default model as the system default after login")
 }
 
 func runAuthLogin(cmd *cobra.Command, args []string) error {
@@ -288,6 +326,17 @@ func loginAnthropic() error {
 	fmt.Println("\n🎉 Your OAuth credentials will now be used for Anthropic API calls.")
 	fmt.Println("💡 You can check your authentication status with: kit auth status")
 
+	// Set default model if requested
+	if err := setDefaultModelIfRequested("anthropic"); err != nil {
+		return err
+	}
+
+	// Remind users how to set this as default if they didn't use --set-default
+	if !loginSetDefault {
+		fmt.Println("\n💡 To set Anthropic as your default model, run:")
+		fmt.Println("   kit auth login anthropic --set-default")
+	}
+
 	return nil
 }
 
@@ -454,6 +503,17 @@ func loginOpenAI() error {
 	fmt.Println("\n🎉 Your OAuth credentials will now be used for OpenAI API calls.")
 	fmt.Println("💡 You can check your authentication status with: kit auth status")
 
+	// Set default model if requested
+	if err := setDefaultModelIfRequested("openai"); err != nil {
+		return err
+	}
+
+	// Remind users how to set this as default if they didn't use --set-default
+	if !loginSetDefault {
+		fmt.Println("\n💡 To set OpenAI as your default model, run:")
+		fmt.Println("   kit auth login openai --set-default")
+	}
+
 	return nil
 }
 
@@ -504,13 +564,13 @@ func startOpenAICallbackServer(expectedState string) (*callbackServer, error) {
 		}
 
 		// Return success page
-		w.Header().Set("Content-Type", "text/html")
+		w.Header().Set("Content-Type", "text/html; charset=utf-8")
 		w.WriteHeader(http.StatusOK)
 		_, _ = fmt.Fprintf(w, `<!DOCTYPE html>
 <html>
 <head><title>Authentication Successful</title></head>
 <body style="font-family: sans-serif; text-align: center; padding: 50px;">
-<h1>✓ Authentication Successful</h1>
+<h1>&#10003; Authentication Successful</h1>
 <p>You can close this window and return to the terminal.</p>
 </body>
 </html>`)

cmd/extension_context.go

@@ -0,0 +1,473 @@
+package cmd
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"strings"
+
+	"github.com/spf13/viper"
+	"golang.org/x/term"
+
+	"github.com/mark3labs/kit/internal/app"
+	"github.com/mark3labs/kit/internal/auth"
+	"github.com/mark3labs/kit/internal/extbridge"
+	"github.com/mark3labs/kit/internal/extensions"
+	"github.com/mark3labs/kit/internal/models"
+	"github.com/mark3labs/kit/internal/ui"
+	kit "github.com/mark3labs/kit/pkg/kit"
+)
+
+// extensionContextDeps groups the runtime dependencies needed to wire up
+// an extensions.Context for the interactive TUI mode.
+type extensionContextDeps struct {
+	ctx          context.Context
+	cwd          string
+	modelName    string
+	interactive  bool
+	kitInstance  *kit.Kit
+	appInstance  *app.App
+	usageTracker *ui.UsageTracker
+}
+
+// buildInteractiveExtensionContext returns an extensions.Context with every
+// field except Print / PrintInfo / PrintError populated. Callers must set
+// the three print routes appropriately for their phase (startup buffering
+// vs. live runtime routing).
+//
+// This consolidates two near-identical 400-line literal expressions that
+// previously appeared inline in runNormalMode.
+func buildInteractiveExtensionContext(deps extensionContextDeps) extensions.Context {
+	kitInstance := deps.kitInstance
+	appInstance := deps.appInstance
+	usageTracker := deps.usageTracker
+	ctx := deps.ctx
+
+	return extensions.Context{
+		CWD:         deps.cwd,
+		Model:       deps.modelName,
+		Interactive: deps.interactive,
+		PrintBlock: func(opts extensions.PrintBlockOpts) {
+			appInstance.PrintBlockFromExtension(opts)
+		},
+		SendMessage:   func(text string) { appInstance.Run(text) },
+		CancelAndSend: func(text string) { appInstance.InterruptAndSend(text) },
+		Abort:         func() { appInstance.Abort() },
+		IsIdle:        func() bool { return !appInstance.IsBusy() },
+		Compact: func(cfg extensions.CompactConfig) error {
+			return appInstance.CompactAsync(cfg.CustomInstructions, cfg.OnComplete, cfg.OnError)
+		},
+		SendMultimodalMessage: func(text string, files []extensions.FilePart) {
+			parts := make([]kit.LLMFilePart, len(files))
+			for i, f := range files {
+				parts[i] = kit.LLMFilePart{
+					Filename:  f.Filename,
+					Data:      f.Data,
+					MediaType: f.MediaType,
+				}
+			}
+			appInstance.RunWithFiles(text, parts)
+		},
+		GetSessionUsage: func() extensions.SessionUsage {
+			if usageTracker == nil {
+				return extensions.SessionUsage{}
+			}
+			stats := usageTracker.GetSessionStats()
+			return extensions.SessionUsage{
+				TotalInputTokens:      stats.TotalInputTokens,
+				TotalOutputTokens:     stats.TotalOutputTokens,
+				TotalCacheReadTokens:  stats.TotalCacheReadTokens,
+				TotalCacheWriteTokens: stats.TotalCacheWriteTokens,
+				TotalCost:             stats.TotalCost,
+				RequestCount:          stats.RequestCount,
+			}
+		},
+		Exit: func() { appInstance.QuitFromExtension() },
+		SetWidget: func(config extensions.WidgetConfig) {
+			kitInstance.Extensions().SetWidget(config)
+			go appInstance.NotifyWidgetUpdate()
+		},
+		RemoveWidget: func(id string) {
+			kitInstance.Extensions().RemoveWidget(id)
+			go appInstance.NotifyWidgetUpdate()
+		},
+		SetHeader: func(config extensions.HeaderFooterConfig) {
+			kitInstance.Extensions().SetHeader(config)
+			go appInstance.NotifyWidgetUpdate()
+		},
+		RemoveHeader: func() {
+			kitInstance.Extensions().RemoveHeader()
+			go appInstance.NotifyWidgetUpdate()
+		},
+		SetFooter: func(config extensions.HeaderFooterConfig) {
+			kitInstance.Extensions().SetFooter(config)
+			go appInstance.NotifyWidgetUpdate()
+		},
+		RemoveFooter: func() {
+			kitInstance.Extensions().RemoveFooter()
+			go appInstance.NotifyWidgetUpdate()
+		},
+		PromptSelect: func(config extensions.PromptSelectConfig) extensions.PromptSelectResult {
+			ch := make(chan app.PromptResponse, 1)
+			appInstance.SendPromptRequest(app.PromptRequestEvent{
+				PromptType: "select",
+				Message:    config.Message,
+				Options:    config.Options,
+				ResponseCh: ch,
+			})
+			resp := <-ch
+			if resp.Cancelled {
+				return extensions.PromptSelectResult{Cancelled: true}
+			}
+			return extensions.PromptSelectResult{Value: resp.Value, Index: resp.Index}
+		},
+		PromptConfirm: func(config extensions.PromptConfirmConfig) extensions.PromptConfirmResult {
+			ch := make(chan app.PromptResponse, 1)
+			def := "false"
+			if config.DefaultValue {
+				def = "true"
+			}
+			appInstance.SendPromptRequest(app.PromptRequestEvent{
+				PromptType: "confirm",
+				Message:    config.Message,
+				Default:    def,
+				ResponseCh: ch,
+			})
+			resp := <-ch
+			if resp.Cancelled {
+				return extensions.PromptConfirmResult{Cancelled: true}
+			}
+			return extensions.PromptConfirmResult{Value: resp.Confirmed}
+		},
+		PromptInput: func(config extensions.PromptInputConfig) extensions.PromptInputResult {
+			ch := make(chan app.PromptResponse, 1)
+			appInstance.SendPromptRequest(app.PromptRequestEvent{
+				PromptType:  "input",
+				Message:     config.Message,
+				Placeholder: config.Placeholder,
+				Default:     config.Default,
+				ResponseCh:  ch,
+			})
+			resp := <-ch
+			if resp.Cancelled {
+				return extensions.PromptInputResult{Cancelled: true}
+			}
+			return extensions.PromptInputResult{Value: resp.Value}
+		},
+		SetUIVisibility: func(v extensions.UIVisibility) {
+			kitInstance.Extensions().SetUIVisibility(v)
+			go appInstance.NotifyWidgetUpdate()
+		},
+		GetContextStats: func() extensions.ContextStats {
+			s := kitInstance.GetContextStats()
+			return extensions.ContextStats{
+				EstimatedTokens: s.EstimatedTokens,
+				ContextLimit:    s.ContextLimit,
+				UsagePercent:    s.UsagePercent,
+				MessageCount:    s.MessageCount,
+			}
+		},
+		SetEditor: func(config extensions.EditorConfig) {
+			kitInstance.Extensions().SetEditor(config)
+			// Always use a goroutine for NotifyWidgetUpdate: prog.Send()
+			// deadlocks if called synchronously from inside BubbleTea's
+			// Update() handler. All call sites use go-routines uniformly.
+			go appInstance.NotifyWidgetUpdate()
+		},
+		ResetEditor: func() {
+			kitInstance.Extensions().ResetEditor()
+			go appInstance.NotifyWidgetUpdate()
+		},
+		GetMessages: func() []extensions.SessionMessage {
+			return kitInstance.Extensions().GetSessionMessages()
+		},
+		GetSessionPath: func() string {
+			return kitInstance.GetSessionPath()
+		},
+		AppendEntry: func(entryType string, data string) (string, error) {
+			return kitInstance.Extensions().AppendEntry(entryType, data)
+		},
+		GetEntries: func(entryType string) []extensions.ExtensionEntry {
+			return kitInstance.Extensions().GetEntries(entryType)
+		},
+		SetEditorText: func(text string) {
+			appInstance.SetEditorTextFromExtension(text)
+		},
+		SetStatus: func(key string, text string, priority int) {
+			kitInstance.Extensions().SetStatus(extensions.StatusBarEntry{
+				Key:      key,
+				Text:     text,
+				Priority: priority,
+			})
+			go appInstance.NotifyWidgetUpdate()
+		},
+		RemoveStatus: func(key string) {
+			kitInstance.Extensions().RemoveStatus(key)
+			go appInstance.NotifyWidgetUpdate()
+		},
+		GetOption: func(name string) string {
+			return kitInstance.Extensions().GetOption(name)
+		},
+		SetOption: func(name string, value string) {
+			kitInstance.Extensions().SetOption(name, value)
+		},
+		SetModel: func(modelString string) error {
+			// Capture previous model for the ModelChange event.
+			previousModel := kitInstance.Extensions().GetContext().Model
+			err := kitInstance.SetModel(context.Background(), modelString)
+			if err != nil {
+				return err
+			}
+			// Notify TUI so it updates model in status bar.
+			p, m, _ := models.ParseModelString(modelString)
+			appInstance.NotifyModelChanged(p, m)
+			// Update the context's Model field so handlers see it.
+			kitInstance.Extensions().UpdateContextModel(modelString)
+			// Fire OnModelChange event to extensions.
+			kitInstance.Extensions().EmitModelChange(modelString, previousModel, "extension")
+			// Update usage tracker with new model info for correct token counting.
+			if usageTracker != nil {
+				newProvider, newModel, _ := models.ParseModelString(modelString)
+				if newProvider != "unknown" && newModel != "unknown" && newProvider != "ollama" {
+					registry := models.GetGlobalRegistry()
+					if modelInfo := registry.LookupModel(newProvider, newModel); modelInfo != nil {
+						// Check OAuth status for Anthropic models
+						isOAuth := false
+						if newProvider == "anthropic" {
+							_, source, err := auth.GetAnthropicAPIKey(viper.GetString("provider-api-key"))
+							if err == nil && strings.HasPrefix(source, "stored OAuth") {
+								isOAuth = true
+							}
+						}
+						usageTracker.UpdateModelInfo(modelInfo, newProvider, isOAuth)
+					}
+				}
+			}
+			return nil
+		},
+		GetAvailableModels: func() []extensions.ModelInfoEntry {
+			return kitInstance.GetAvailableModels()
+		},
+		EmitCustomEvent: func(name string, data string) {
+			kitInstance.Extensions().EmitCustomEvent(name, data)
+		},
+		Complete: func(req extensions.CompleteRequest) (extensions.CompleteResponse, error) {
+			return kitInstance.ExecuteCompletion(context.Background(), req)
+		},
+		SuspendTUI: func(callback func()) error {
+			return appInstance.SuspendTUI(callback)
+		},
+		RenderMessage: func(rendererName, content string) {
+			renderer := kitInstance.Extensions().GetMessageRenderer(rendererName)
+			if renderer == nil || renderer.Render == nil {
+				appInstance.PrintFromExtension("", content)
+				return
+			}
+			w, _, _ := term.GetSize(int(os.Stdout.Fd()))
+			if w == 0 {
+				w = 80
+			}
+			rendered := renderer.Render(content, w)
+			appInstance.PrintFromExtension("", rendered)
+		},
+		ReloadExtensions: func() error {
+			err := kitInstance.Extensions().Reload()
+			if err != nil {
+				return err
+			}
+			// Notify TUI that widgets/status/commands may have changed.
+			go appInstance.NotifyWidgetUpdate()
+			return nil
+		},
+		GetAllTools: func() []extensions.ToolInfo {
+			return kitInstance.Extensions().GetToolInfos()
+		},
+		SetActiveTools: func(names []string) {
+			kitInstance.Extensions().SetActiveTools(names)
+		},
+		RegisterTheme: func(name string, config extensions.ThemeColorConfig) {
+			tc := func(c extensions.ThemeColor) [2]string { return [2]string{c.Light, c.Dark} }
+			ui.RegisterThemeFromConfig(name,
+				tc(config.Primary), tc(config.Secondary),
+				tc(config.Success), tc(config.Warning),
+				tc(config.Error), tc(config.Info),
+				tc(config.Text), tc(config.Muted),
+				tc(config.VeryMuted), tc(config.Background),
+				tc(config.Border), tc(config.MutedBorder),
+				tc(config.System), tc(config.Tool),
+				tc(config.Accent), tc(config.Highlight),
+				tc(config.MdHeading), tc(config.MdLink),
+				tc(config.MdKeyword), tc(config.MdString),
+				tc(config.MdNumber), tc(config.MdComment),
+			)
+		},
+		SetTheme: func(name string) error {
+			return ui.ApplyTheme(name)
+		},
+		ListThemes: func() []string {
+			return ui.ListThemes()
+		},
+		ShowOverlay: func(config extensions.OverlayConfig) extensions.OverlayResult {
+			ch := make(chan app.OverlayResponse, 1)
+			appInstance.SendOverlayRequest(app.OverlayRequestEvent{
+				Title:       config.Title,
+				Content:     config.Content.Text,
+				Markdown:    config.Content.Markdown,
+				BorderColor: config.Style.BorderColor,
+				Background:  config.Style.Background,
+				Width:       config.Width,
+				MaxHeight:   config.MaxHeight,
+				Anchor:      string(config.Anchor),
+				Actions:     config.Actions,
+				ResponseCh:  ch,
+			})
+			resp := <-ch
+			if resp.Cancelled {
+				return extensions.OverlayResult{Cancelled: true, Index: -1}
+			}
+			return extensions.OverlayResult{
+				Action: resp.Action,
+				Index:  resp.Index,
+			}
+		},
+		SpawnSubagent: func(config extensions.SubagentConfig) (*extensions.SubagentHandle, *extensions.SubagentResult, error) {
+			return extbridge.SpawnSubagent(ctx, kitInstance, config)
+		},
+		// -------------------------------------------------------------------
+		// Tree Navigation API
+		// -------------------------------------------------------------------
+		GetTreeNode: func(entryID string) *extensions.TreeNode {
+			node := kitInstance.GetTreeNode(entryID)
+			if node == nil {
+				return nil
+			}
+			return &extensions.TreeNode{
+				ID:        node.ID,
+				ParentID:  node.ParentID,
+				Type:      node.Type,
+				Role:      node.Role,
+				Content:   node.Content,
+				Model:     node.Model,
+				Provider:  node.Provider,
+				Timestamp: node.Timestamp,
+				Children:  node.Children,
+			}
+		},
+		GetCurrentBranch: func() []extensions.TreeNode {
+			nodes := kitInstance.GetCurrentBranch()
+			result := make([]extensions.TreeNode, len(nodes))
+			for i, n := range nodes {
+				result[i] = extensions.TreeNode{
+					ID:        n.ID,
+					ParentID:  n.ParentID,
+					Type:      n.Type,
+					Role:      n.Role,
+					Content:   n.Content,
+					Model:     n.Model,
+					Provider:  n.Provider,
+					Timestamp: n.Timestamp,
+					Children:  n.Children,
+				}
+			}
+			return result
+		},
+		GetChildren: func(parentID string) []string {
+			return kitInstance.GetChildren(parentID)
+		},
+		NavigateTo: func(entryID string) extensions.TreeNavigationResult {
+			err := kitInstance.NavigateTo(entryID)
+			if err != nil {
+				return extensions.TreeNavigationResult{Success: false, Error: err.Error()}
+			}
+			return extensions.TreeNavigationResult{Success: true}
+		},
+		SummarizeBranch: func(fromID, toID string) string {
+			summary, _ := kitInstance.SummarizeBranch(fromID, toID)
+			return summary
+		},
+		CollapseBranch: func(fromID, toID, summary string) extensions.TreeNavigationResult {
+			err := kitInstance.CollapseBranch(fromID, toID, summary)
+			if err != nil {
+				return extensions.TreeNavigationResult{Success: false, Error: err.Error()}
+			}
+			return extensions.TreeNavigationResult{Success: true}
+		},
+
+		// -------------------------------------------------------------------
+		// Skill Loading API
+		// -------------------------------------------------------------------
+		LoadSkill: func(path string) (*extensions.Skill, string) {
+			s, err := kitInstance.LoadSkillForExtension(path)
+			return s, err
+		},
+		LoadSkillsFromDir: func(dir string) extensions.SkillLoadResult {
+			return kitInstance.LoadSkillsFromDirForExtension(dir)
+		},
+		DiscoverSkills: func() extensions.SkillLoadResult {
+			skills := kitInstance.DiscoverSkillsForExtension()
+			return extensions.SkillLoadResult{Skills: skills}
+		},
+		InjectSkillAsContext: func(skillName string) string {
+			skills := kitInstance.DiscoverSkillsForExtension()
+			for _, s := range skills {
+				if s.Name == skillName {
+					appInstance.Run(fmt.Sprintf("<skill name=%q>\n%s\n</skill>", s.Name, s.Content))
+					return ""
+				}
+			}
+			return fmt.Sprintf("skill not found: %s", skillName)
+		},
+		InjectRawSkillAsContext: func(path string) string {
+			s, err := kitInstance.LoadSkillForExtension(path)
+			if err != "" {
+				return err
+			}
+			appInstance.Run(fmt.Sprintf("<skill name=%q>\n%s\n</skill>", s.Name, s.Content))
+			return ""
+		},
+		GetAvailableSkills: func() []extensions.Skill {
+			return kitInstance.DiscoverSkillsForExtension()
+		},
+
+		// -------------------------------------------------------------------
+		// Template Parsing API
+		// -------------------------------------------------------------------
+		ParseTemplate: func(name, content string) extensions.PromptTemplate {
+			return kit.ParseTemplate(name, content)
+		},
+		RenderTemplate: func(tpl extensions.PromptTemplate, vars map[string]string) string {
+			return kit.RenderTemplate(tpl, vars)
+		},
+		ParseArguments: func(input string, pattern extensions.ArgumentPattern) extensions.ParseResult {
+			return kit.ParseArguments(input, pattern)
+		},
+		SimpleParseArguments: func(input string, count int) []string {
+			return kit.SimpleParseArguments(input, count)
+		},
+		EvaluateModelConditional: func(condition string) bool {
+			return kit.EvaluateModelConditional(kitInstance.Extensions().GetContext().Model, condition)
+		},
+		RenderWithModelConditionals: func(content string) string {
+			return kit.RenderWithModelConditionals(content, kitInstance.Extensions().GetContext().Model)
+		},
+
+		// -------------------------------------------------------------------
+		// Model Resolution API
+		// -------------------------------------------------------------------
+		ResolveModelChain: func(preferences []string) extensions.ModelResolutionResult {
+			return kit.ResolveModelChain(preferences)
+		},
+		GetModelCapabilities: func(model string) (extensions.ModelCapabilities, string) {
+			return kit.GetModelCapabilities(model)
+		},
+		CheckModelAvailable: func(model string) bool {
+			return kit.CheckModelAvailable(model)
+		},
+		GetCurrentProvider: func() string {
+			return kit.GetCurrentProvider(kitInstance.Extensions().GetContext().Model)
+		},
+		GetCurrentModelID: func() string {
+			return kit.GetCurrentModelID(kitInstance.Extensions().GetContext().Model)
+		},
+	}
+}

cmd/setup.go

@@ -41,7 +41,6 @@ func BuildAppOptions(mcpConfig *config.Config, modelName string, serverNames, to
 		StreamingEnabled: viper.GetBool("stream"),
 		Quiet:            quietFlag,
 		Debug:            viper.GetBool("debug"),
-		CompactMode:      viper.GetBool("compact"),
 	}
 }
 
@@ -131,7 +130,6 @@ func SetupCLIForNonInteractive(k *kit.Kit) (*ui.CLI, error) {
 		Agent:          agentAdapter,
 		ModelString:    viper.GetString("model"),
 		Debug:          viper.GetBool("debug"),
-		Compact:        viper.GetBool("compact"),
 		Quiet:          quietFlag,
 		ShowDebug:      false,
 		ProviderAPIKey: viper.GetString("provider-api-key"),

examples/extensions/dev-reload.go

@@ -7,10 +7,12 @@
 // development: edit your extension source, then type /reload to pick up
 // changes immediately.
 //
-// Event handlers, slash commands, tool renderers, message renderers, and
-// keyboard shortcuts update immediately. Extension-defined tools are NOT
-// updated (they are baked into the agent at creation time and require a
-// restart).
+// Note: Extensions in autoloaded directories (~/.config/kit/extensions/
+// and .kit/extensions/) are automatically reloaded on save. The /reload
+// command is useful for extensions loaded via -e from other locations.
+//
+// Event handlers, slash commands, tool definitions, tool renderers,
+// message renderers, and keyboard shortcuts all update immediately.
 //
 // Commands:
 //   /reload   — hot-reload all extensions from disk

examples/extensions/inline-bash.go

@@ -10,13 +10,21 @@ import (
 	"kit/ext"
 )
 
+// re matches !{...} with non-greedy content.
+var re = regexp.MustCompile(`!\{([^}]+)\}`)
+
 // Init expands inline bash expressions in user prompts before they reach the
-// LLM. Text like !{git branch --show-current} is replaced with the command's
-// stdout.
+// LLM. Text like !{git rev-parse --abbrev-ref HEAD} is replaced with the
+// command's stdout.
+//
+// In interactive mode the expansion happens at submit time via an editor
+// interceptor, so the expanded text is also visible in the user message
+// block on screen. In non-interactive mode (CLI, script, queue) the
+// expansion happens via OnInput transform.
 //
 // Examples:
 //
-//	"Fix the tests on !{git branch --show-current}"
+//	"Fix the tests on !{git rev-parse --abbrev-ref HEAD}"
 //	  → "Fix the tests on main"
 //
 //	"The current directory is !{pwd}"
@@ -24,29 +32,59 @@ import (
 //
 // Usage: kit -e examples/extensions/inline-bash.go
 func Init(api ext.API) {
-	// Matches !{...} with non-greedy content.
-	re := regexp.MustCompile(`!\{([^}]+)\}`)
+	// ── Interactive mode: editor interceptor ──────────────────────────
+	// Intercept Enter / Ctrl+D so we can expand !{...} BEFORE the
+	// SubmitMsg is created. This ensures the expanded text appears in
+	// the user message block on screen as well as in the LLM prompt.
+	api.OnSessionStart(func(_ ext.SessionStartEvent, ctx ext.Context) {
+		if !ctx.Interactive {
+			return
+		}
+		ctx.SetEditor(ext.EditorConfig{
+			HandleKey: func(key string, currentText string) ext.EditorKeyAction {
+				if (key == "enter" || key == "ctrl+d") && re.MatchString(currentText) {
+					expanded := expand(currentText)
+					// Clear the textarea asynchronously — calling
+					// SetEditorText synchronously from inside Update()
+					// would deadlock the BubbleTea event loop.
+					go ctx.SetEditorText("")
+					return ext.EditorKeyAction{
+						Type:       ext.EditorKeySubmit,
+						SubmitText: expanded,
+					}
+				}
+				return ext.EditorKeyAction{Type: ext.EditorKeyPassthrough}
+			},
+		})
+	})
 
+	// ── Non-interactive fallback: OnInput transform ──────────────────
+	// For CLI, script, and queue sources the editor interceptor is not
+	// active, so we fall back to OnInput which still rewrites the
+	// prompt text sent to the LLM.
 	api.OnInput(func(ev ext.InputEvent, ctx ext.Context) *ext.InputResult {
-		if !re.MatchString(ev.Text) {
+		if ev.Source == "interactive" || !re.MatchString(ev.Text) {
 			return nil
 		}
 
-		expanded := re.ReplaceAllStringFunc(ev.Text, func(match string) string {
-			// Extract the command between !{ and }.
-			cmd := re.FindStringSubmatch(match)[1]
-			cmd = strings.TrimSpace(cmd)
-
-			out, err := exec.Command("bash", "-c", cmd).Output()
-			if err != nil {
-				return match // keep original on error
-			}
-			return strings.TrimSpace(string(out))
-		})
-
 		return &ext.InputResult{
 			Action: "transform",
-			Text:   expanded,
+			Text:   expand(ev.Text),
 		}
 	})
 }
+
+// expand replaces every !{cmd} in text with the command's stdout.
+// On error the original !{cmd} token is preserved.
+func expand(text string) string {
+	return re.ReplaceAllStringFunc(text, func(match string) string {
+		cmd := re.FindStringSubmatch(match)[1]
+		cmd = strings.TrimSpace(cmd)
+
+		out, err := exec.Command("bash", "-c", cmd).Output()
+		if err != nil {
+			return match // keep original on error
+		}
+		return strings.TrimSpace(string(out))
+	})
+}

examples/extensions/kit-telegram/main.go

@@ -168,6 +168,10 @@ var (
 	// Test
 	pendingTest *PendingTest
 
+	// Typing indicator
+	typingTicker *time.Ticker
+	typingStop   chan struct{}
+
 	// Latest context for background goroutines
 	latestCtx    ext.Context
 	latestCtxSet bool
@@ -203,8 +207,23 @@ func configDir() string {
 	return filepath.Join(home, ".config", "kit")
 }
 
+func globalConfigDir() string {
+	home, _ := os.UserHomeDir()
+	return filepath.Join(home, ".config", "kit")
+}
+
 func configPath() string {
-	return filepath.Join(configDir(), "kit-telegram.json")
+	// Prefer project-local config, fall back to global config.
+	local := filepath.Join(configDir(), "kit-telegram.json")
+	if _, err := os.Stat(local); err == nil {
+		return local
+	}
+	global := filepath.Join(globalConfigDir(), "kit-telegram.json")
+	if _, err := os.Stat(global); err == nil {
+		return global
+	}
+	// Neither exists — return local path (will be created on connect).
+	return local
 }
 
 func failureLogDir() string {
@@ -387,6 +406,14 @@ func tgEditMessageText(token string, chatID int64, messageID int, text string) (
 	return &msg, nil
 }
 
+func tgSendChatAction(token string, chatID int64, action string) error {
+	_, err := telegramRequest(token, "sendChatAction", map[string]any{
+		"chat_id": chatID,
+		"action":  action,
+	}, 15)
+	return err
+}
+
 // ──────────────────────────────────────────────
 // Error classification
 // ──────────────────────────────────────────────
@@ -637,6 +664,48 @@ func clearHealthTimer() {
 	}
 }
 
+// ──────────────────────────────────────────────
+// Typing indicator
+// ──────────────────────────────────────────────
+
+func startTypingLoop() {
+	mu.Lock()
+	defer mu.Unlock()
+	if typingTicker != nil {
+		return
+	}
+	cfg := config
+	if cfg == nil || !cfg.Enabled {
+		return
+	}
+	token := cfg.BotToken
+	chatID := cfg.ChatID
+	typingTicker = time.NewTicker(4 * time.Second)
+	typingStop = make(chan struct{})
+	// Send immediately, then every 4 seconds.
+	go func() {
+		tgSendChatAction(token, chatID, "typing")
+		for {
+			select {
+			case <-typingTicker.C:
+				tgSendChatAction(token, chatID, "typing")
+			case <-typingStop:
+				return
+			}
+		}
+	}()
+}
+
+func stopTypingLoop() {
+	mu.Lock()
+	defer mu.Unlock()
+	if typingTicker != nil {
+		typingTicker.Stop()
+		close(typingStop)
+		typingTicker = nil
+	}
+}
+
 // ──────────────────────────────────────────────
 // Polling lifecycle
 // ──────────────────────────────────────────────
@@ -2105,6 +2174,7 @@ func Init(api ext.API) {
 		mu.Unlock()
 
 		sendShutdownDisconnectedMessage()
+		stopTypingLoop()
 		stopPolling()
 		clearHealthTimer()
 		clearFooter()
@@ -2128,6 +2198,7 @@ func Init(api ext.API) {
 		mu.Unlock()
 
 		report("run.start", fmt.Sprintf("runId=%d", run.ID))
+		startTypingLoop()
 		ensureProgressMessage()
 		updateProgressMessage()
 	})
@@ -2140,6 +2211,8 @@ func Init(api ext.API) {
 		run := activeRun
 		mu.Unlock()
 
+		stopTypingLoop()
+
 		if run != nil {
 			// Capture final response from event
 			if e.Response != "" {

examples/extensions/subagent-monitor_test.go

@@ -13,7 +13,7 @@ import (
 // without panicking and properly guards nil ctx calls.
 func TestSubagentMonitor_SessionStart(t *testing.T) {
 	harness := test.New(t)
-	harness.LoadFile("../../.kit/extensions/subagent-monitor.go")
+	harness.LoadFile("./subagent-monitor.go")
 
 	// Emit SessionStart - should not panic even with nil ctx functions
 	_, err := harness.Emit(extensions.SessionStartEvent{SessionID: "test-session"})
@@ -26,7 +26,7 @@ func TestSubagentMonitor_SessionStart(t *testing.T) {
 // creates entries and emits widget updates.
 func TestSubagentMonitor_SubagentLifecycle(t *testing.T) {
 	harness := test.New(t)
-	harness.LoadFile("../../.kit/extensions/subagent-monitor.go")
+	harness.LoadFile("./subagent-monitor.go")
 
 	// Start session
 	_, err := harness.Emit(extensions.SessionStartEvent{SessionID: "test-session"})
@@ -84,7 +84,7 @@ func TestSubagentMonitor_SubagentLifecycle(t *testing.T) {
 // TestSubagentMonitor_MultipleSubagents verifies multiple parallel subagents.
 func TestSubagentMonitor_MultipleSubagents(t *testing.T) {
 	harness := test.New(t)
-	harness.LoadFile("../../.kit/extensions/subagent-monitor.go")
+	harness.LoadFile("./subagent-monitor.go")
 
 	_, err := harness.Emit(extensions.SessionStartEvent{SessionID: "test-session"})
 	if err != nil {
@@ -130,11 +130,63 @@ func TestSubagentMonitor_MultipleSubagents(t *testing.T) {
 	time.Sleep(100 * time.Millisecond)
 }
 
+// TestSubagentMonitor_ConcurrentSubagents verifies no panics when multiple
+// subagents emit events concurrently from different goroutines.
+func TestSubagentMonitor_ConcurrentSubagents(t *testing.T) {
+	harness := test.New(t)
+	harness.LoadFile("./subagent-monitor.go")
+
+	_, err := harness.Emit(extensions.SessionStartEvent{SessionID: "test-session"})
+	if err != nil {
+		t.Fatalf("SessionStart should not error: %v", err)
+	}
+
+	// Start 5 subagents concurrently
+	done := make(chan struct{}, 5)
+	for i := range 5 {
+		go func(idx int) {
+			defer func() { done <- struct{}{} }()
+
+			callID := fmt.Sprintf("concurrent-%d", idx)
+			task := fmt.Sprintf("concurrent task %d", idx)
+
+			_, _ = harness.Emit(extensions.SubagentStartEvent{
+				ToolCallID: callID,
+				Task:       task,
+			})
+
+			// Emit many chunks rapidly
+			for j := range 20 {
+				_, _ = harness.Emit(extensions.SubagentChunkEvent{
+					ToolCallID: callID,
+					Task:       task,
+					ChunkType:  "text",
+					Content:    fmt.Sprintf("agent %d chunk %d", idx, j),
+				})
+			}
+
+			_, _ = harness.Emit(extensions.SubagentEndEvent{
+				ToolCallID: callID,
+				Task:       task,
+				Response:   "done",
+			})
+		}(i)
+	}
+
+	// Wait for all goroutines
+	for range 5 {
+		<-done
+	}
+
+	// Allow any final processing
+	time.Sleep(200 * time.Millisecond)
+}
+
 // TestSubagentMonitor_SessionShutdown verifies shutdown doesn't panic
 // even with nil ctx functions.
 func TestSubagentMonitor_SessionShutdown(t *testing.T) {
 	harness := test.New(t)
-	harness.LoadFile("../../.kit/extensions/subagent-monitor.go")
+	harness.LoadFile("./subagent-monitor.go")
 
 	// Start then shutdown
 	_, err := harness.Emit(extensions.SessionStartEvent{SessionID: "test-session"})

examples/extensions/sudo-handler.go

@@ -0,0 +1,153 @@
+//go:build ignore
+
+// sudo-handler.go - Extension to handle sudo password prompts securely
+//
+// This extension intercepts bash commands containing "sudo" and:
+// 1. Checks if sudo credentials are already cached (via sudo -n)
+// 2. If not cached, prompts the user for their password (with masking)
+// 3. Temporarily sets SUDO_PASSWORD environment variable for execution
+// 4. The bash tool automatically uses sudo -S -p '' to pipe the password
+//
+// Usage: kit -e examples/extensions/sudo-handler.go
+//
+// Security notes:
+// - Password is only stored in memory for the duration of the session
+// - Password is never logged or displayed
+// - Each session requires re-authentication (sudo -k is used)
+// - The SUDO_PASSWORD env var is set only during tool execution
+
+package main
+
+import (
+	"encoding/json"
+	"os"
+	"strings"
+	"sync"
+
+	"kit/ext"
+)
+
+var (
+	// cachedPassword stores the sudo password for the session
+	cachedPassword string
+	// hasCachedPassword tracks if we have a valid cached password
+	hasCachedPassword bool
+	// mu protects cached password access
+	mu sync.RWMutex
+)
+
+// Init sets up the sudo handler extension
+func Init(api ext.API) {
+	api.OnToolCall(func(tc ext.ToolCallEvent, ctx ext.Context) *ext.ToolCallResult {
+		if tc.ToolName != "bash" {
+			return nil
+		}
+
+		// Parse the command from tool input
+		var input struct {
+			Command string `json:"command"`
+		}
+		if err := json.Unmarshal([]byte(tc.Input), &input); err != nil {
+			return nil
+		}
+
+		// Check if command contains sudo
+		if !containsSudo(input.Command) {
+			return nil
+		}
+
+		// Check if we already have cached credentials
+		mu.RLock()
+		password := cachedPassword
+		hasCached := hasCachedPassword
+		mu.RUnlock()
+
+		if hasCached {
+			// Use cached password
+			os.Setenv("SUDO_PASSWORD", password)
+			return nil
+		}
+
+		// No cached password - prompt user
+		result := ctx.PromptInput(ext.PromptInputConfig{
+			Message:     "🔐 Sudo password required for:\n  " + truncateCommand(input.Command, 60),
+			Placeholder: "Enter your password",
+		})
+
+		if result.Cancelled {
+			return &ext.ToolCallResult{
+				Block:  true,
+				Reason: "Sudo password prompt cancelled by user",
+			}
+		}
+
+		if result.Value == "" {
+			return &ext.ToolCallResult{
+				Block:  true,
+				Reason: "No password provided",
+			}
+		}
+
+		// Cache the password for this session
+		mu.Lock()
+		cachedPassword = result.Value
+		hasCachedPassword = true
+		mu.Unlock()
+
+		// Set environment variable for the bash tool to use
+		os.Setenv("SUDO_PASSWORD", result.Value)
+
+		// Show confirmation (without revealing password)
+		ctx.PrintInfo("Sudo password cached for this session")
+
+		return nil
+	})
+
+	// Clear cached password when session ends
+	api.OnSessionShutdown(func(event ext.SessionShutdownEvent, ctx ext.Context) {
+		mu.Lock()
+		cachedPassword = ""
+		hasCachedPassword = false
+		mu.Unlock()
+		os.Unsetenv("SUDO_PASSWORD")
+	})
+}
+
+// containsSudo checks if the command contains sudo as a command (not in a string)
+func containsSudo(command string) bool {
+	// Simple check for sudo as a word, not inside quotes or as part of another word
+	lower := strings.ToLower(command)
+
+	// Check for sudo at start or after separators
+	patterns := []string{
+		"sudo ",
+		"sudo\t",
+		";sudo ",
+		"&& sudo ",
+		"|| sudo ",
+		"| sudo ",
+		"$(sudo ",
+		"`sudo ",
+	}
+
+	for _, pattern := range patterns {
+		if strings.Contains(lower, pattern) {
+			return true
+		}
+	}
+
+	// Check if command starts with sudo
+	if strings.HasPrefix(lower, "sudo ") {
+		return true
+	}
+
+	return false
+}
+
+// truncateCommand truncates a long command for display
+func truncateCommand(cmd string, maxLen int) string {
+	if len(cmd) <= maxLen {
+		return cmd
+	}
+	return cmd[:maxLen-3] + "..."
+}

examples/sdk/README.md

@@ -42,4 +42,14 @@ defer host.Close()
 response, err := host.Prompt(ctx, "Hello!")
 ```
 
+Or use the functional-options constructor for quick setups (streaming defaults on):
+
+```go
+host, err := kit.NewAgent(ctx,
+    kit.WithModel("anthropic/claude-sonnet-4-5-20250929"),
+    kit.WithSystemPrompt("You are a helpful assistant."),
+    kit.Ephemeral(),
+)
+```
+
 See the [SDK README](../../pkg/kit/README.md) for the full API reference.

examples/sdk/basic/main.go

@@ -62,7 +62,7 @@ func main() {
 		}
 	})
 	// Subscribe to streaming chunks.
-	host3.OnStreaming(func(e kit.MessageUpdateEvent) {
+	host3.OnMessageUpdate(func(e kit.MessageUpdateEvent) {
 		fmt.Print(e.Chunk)
 	})
 

go.mod

@@ -1,102 +1,104 @@
 module github.com/mark3labs/kit
 
-go 1.26.1
+go 1.26.3
 
 require (
 	charm.land/bubbles/v2 v2.1.0
-	charm.land/bubbletea/v2 v2.0.2
-	charm.land/fantasy v0.17.1
+	charm.land/bubbletea/v2 v2.0.6
+	charm.land/fantasy v0.25.0
 	charm.land/huh/v2 v2.0.3
-	charm.land/lipgloss/v2 v2.0.2
-	github.com/alecthomas/chroma/v2 v2.23.1
+	charm.land/lipgloss/v2 v2.0.3
+	github.com/alecthomas/chroma/v2 v2.26.1
+	github.com/atotto/clipboard v0.1.4
 	github.com/aymanbagabas/go-udiff v0.4.1
 	github.com/charmbracelet/fang v1.0.0
 	github.com/charmbracelet/log v1.0.0
-	github.com/coder/acp-go-sdk v0.6.3
-	github.com/indaco/herald v0.10.0
-	github.com/indaco/herald-md v0.1.0
-	github.com/mark3labs/mcp-go v0.46.0
+	github.com/charmbracelet/openai-go v0.0.0-20260319145158-d0740cc34266
+	github.com/charmbracelet/ultraviolet v0.0.0-20260525132238-948f4557a654
+	github.com/charmbracelet/x/editor v0.2.0
+	github.com/clipperhouse/displaywidth v0.11.0
+	github.com/clipperhouse/uax29/v2 v2.7.0
+	github.com/coder/acp-go-sdk v0.13.0
+	github.com/fsnotify/fsnotify v1.10.1
+	github.com/indaco/herald v0.13.0
+	github.com/indaco/herald-md v0.3.0
+	github.com/mark3labs/mcp-go v0.54.1
 	github.com/spf13/cobra v1.10.2
 	github.com/spf13/viper v1.21.0
 	github.com/traefik/yaegi v0.16.1
-	golang.org/x/term v0.41.0
+	golang.org/x/term v0.43.0
 	gopkg.in/yaml.v3 v3.0.1
 )
 
 require (
 	cloud.google.com/go v0.123.0 // indirect
-	cloud.google.com/go/auth v0.19.0 // indirect
+	cloud.google.com/go/auth v0.20.0 // indirect
 	cloud.google.com/go/auth/oauth2adapt v0.2.8 // indirect
 	cloud.google.com/go/compute/metadata v0.9.0 // indirect
-	github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.0 // indirect
-	github.com/Azure/azure-sdk-for-go/sdk/internal v1.11.2 // indirect
-	github.com/atotto/clipboard v0.1.4 // indirect
-	github.com/aws/aws-sdk-go-v2 v1.41.5 // indirect
-	github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.8 // indirect
-	github.com/aws/aws-sdk-go-v2/config v1.32.13 // indirect
-	github.com/aws/aws-sdk-go-v2/credentials v1.19.13 // indirect
-	github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.18.21 // indirect
-	github.com/aws/aws-sdk-go-v2/internal/configsources v1.4.21 // indirect
-	github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.7.21 // indirect
-	github.com/aws/aws-sdk-go-v2/internal/ini v1.8.6 // indirect
-	github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.7 // indirect
-	github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.13.21 // indirect
-	github.com/aws/aws-sdk-go-v2/service/signin v1.0.9 // indirect
-	github.com/aws/aws-sdk-go-v2/service/sso v1.30.14 // indirect
-	github.com/aws/aws-sdk-go-v2/service/ssooidc v1.35.18 // indirect
-	github.com/aws/aws-sdk-go-v2/service/sts v1.41.10 // indirect
-	github.com/aws/smithy-go v1.24.2 // indirect
+	github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.1 // indirect
+	github.com/Azure/azure-sdk-for-go/sdk/internal v1.12.0 // indirect
+	github.com/aws/aws-sdk-go-v2 v1.41.8 // indirect
+	github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.10 // indirect
+	github.com/aws/aws-sdk-go-v2/config v1.32.19 // indirect
+	github.com/aws/aws-sdk-go-v2/credentials v1.19.18 // indirect
+	github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.18.24 // indirect
+	github.com/aws/aws-sdk-go-v2/internal/configsources v1.4.24 // indirect
+	github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.7.24 // indirect
+	github.com/aws/aws-sdk-go-v2/internal/v4a v1.4.25 // indirect
+	github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.9 // indirect
+	github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.13.24 // indirect
+	github.com/aws/aws-sdk-go-v2/service/signin v1.1.0 // indirect
+	github.com/aws/aws-sdk-go-v2/service/sso v1.30.18 // indirect
+	github.com/aws/aws-sdk-go-v2/service/ssooidc v1.36.1 // indirect
+	github.com/aws/aws-sdk-go-v2/service/sts v1.42.2 // indirect
+	github.com/aws/smithy-go v1.26.0 // indirect
 	github.com/catppuccin/go v0.3.0 // indirect
 	github.com/cespare/xxhash/v2 v2.3.0 // indirect
 	github.com/charmbracelet/anthropic-sdk-go v0.0.0-20260223140439-63879b0b8dab // indirect
 	github.com/charmbracelet/colorprofile v0.4.3 // indirect
 	github.com/charmbracelet/harmonica v0.2.0 // indirect
 	github.com/charmbracelet/lipgloss v1.1.1-0.20250404203927-76690c660834 // indirect
-	github.com/charmbracelet/openai-go v0.0.0-20260319145158-d0740cc34266 // indirect
-	github.com/charmbracelet/ultraviolet v0.0.0-20260330092749-0f94982c930b // indirect
 	github.com/charmbracelet/x/cellbuf v0.0.15 // indirect
-	github.com/charmbracelet/x/exp/charmtone v0.0.0-20260330094520-2dce04b6f8a4 // indirect
+	github.com/charmbracelet/x/exp/charmtone v0.0.0-20260527151214-009e6338d40d // indirect
 	github.com/charmbracelet/x/exp/ordered v0.1.0 // indirect
-	github.com/charmbracelet/x/exp/slice v0.0.0-20260330094520-2dce04b6f8a4 // indirect
+	github.com/charmbracelet/x/exp/slice v0.0.0-20260527151214-009e6338d40d // indirect
 	github.com/charmbracelet/x/exp/strings v0.1.0 // indirect
 	github.com/charmbracelet/x/json v0.2.0 // indirect
 	github.com/charmbracelet/x/termios v0.1.1 // indirect
 	github.com/charmbracelet/x/windows v0.2.2 // indirect
-	github.com/clipperhouse/displaywidth v0.11.0 // indirect
-	github.com/clipperhouse/uax29/v2 v2.7.0 // indirect
-	github.com/dlclark/regexp2 v1.11.5 // indirect
+	github.com/dlclark/regexp2 v1.12.0 // indirect
+	github.com/dlclark/regexp2/v2 v2.1.1 // indirect
 	github.com/dustin/go-humanize v1.0.1 // indirect
 	github.com/felixge/httpsnoop v1.0.4 // indirect
-	github.com/fsnotify/fsnotify v1.9.0 // indirect
-	github.com/go-json-experiment/json v0.0.0-20260214004413-d219187c3433 // indirect
+	github.com/go-json-experiment/json v0.0.0-20260520185125-572e7c383686 // indirect
 	github.com/go-logfmt/logfmt v0.6.1 // indirect
 	github.com/go-logr/logr v1.4.3 // indirect
 	github.com/go-logr/stdr v1.2.2 // indirect
 	github.com/go-viper/mapstructure/v2 v2.5.0 // indirect
 	github.com/goccy/go-yaml v1.19.2 // indirect
-	github.com/golang-jwt/jwt/v5 v5.3.0 // indirect
 	github.com/google/go-cmp v0.7.0 // indirect
-	github.com/google/jsonschema-go v0.4.2 // indirect
+	github.com/google/jsonschema-go v0.4.3 // indirect
 	github.com/google/s2a-go v0.1.9 // indirect
 	github.com/google/uuid v1.6.0 // indirect
-	github.com/googleapis/enterprise-certificate-proxy v0.3.14 // indirect
-	github.com/googleapis/gax-go/v2 v2.20.0 // indirect
+	github.com/googleapis/enterprise-certificate-proxy v0.3.16 // indirect
+	github.com/googleapis/gax-go/v2 v2.22.0 // indirect
 	github.com/gorilla/websocket v1.5.3 // indirect
-	github.com/kaptinlin/go-i18n v0.3.0 // indirect
-	github.com/kaptinlin/jsonpointer v0.4.17 // indirect
-	github.com/kaptinlin/jsonschema v0.7.7 // indirect
-	github.com/kaptinlin/messageformat-go v0.4.19 // indirect
+	github.com/kaptinlin/go-i18n v0.4.5 // indirect
+	github.com/kaptinlin/jsonpointer v0.4.25 // indirect
+	github.com/kaptinlin/jsonschema v0.7.13 // indirect
+	github.com/kaptinlin/messageformat-go v0.6.0 // indirect
 	github.com/mitchellh/hashstructure/v2 v2.0.2 // indirect
 	github.com/muesli/mango v0.2.0 // indirect
 	github.com/muesli/mango-cobra v1.3.0 // indirect
 	github.com/muesli/mango-pflag v0.2.0 // indirect
 	github.com/muesli/roff v0.1.0 // indirect
-	github.com/pelletier/go-toml/v2 v2.3.0 // indirect
+	github.com/pelletier/go-toml/v2 v2.3.1 // indirect
 	github.com/sagikazarmark/locafero v0.12.0 // indirect
+	github.com/santhosh-tekuri/jsonschema/v6 v6.0.2 // indirect
 	github.com/spf13/afero v1.15.0 // indirect
 	github.com/spf13/cast v1.10.0 // indirect
 	github.com/subosito/gotenv v1.6.0 // indirect
-	github.com/tidwall/gjson v1.18.0 // indirect
+	github.com/tidwall/gjson v1.19.0 // indirect
 	github.com/tidwall/match v1.2.0 // indirect
 	github.com/tidwall/pretty v1.2.1 // indirect
 	github.com/tidwall/sjson v1.2.5 // indirect
@@ -104,38 +106,38 @@ require (
 	github.com/yosida95/uritemplate/v3 v3.0.2 // indirect
 	github.com/yuin/goldmark v1.8.2 // indirect
 	go.opentelemetry.io/auto/sdk v1.2.1 // indirect
-	go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.67.0 // indirect
-	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.67.0 // indirect
-	go.opentelemetry.io/otel v1.42.0 // indirect
-	go.opentelemetry.io/otel/metric v1.42.0 // indirect
-	go.opentelemetry.io/otel/trace v1.42.0 // indirect
+	go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.69.0 // indirect
+	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.69.0 // indirect
+	go.opentelemetry.io/otel v1.44.0 // indirect
+	go.opentelemetry.io/otel/metric v1.44.0 // indirect
+	go.opentelemetry.io/otel/trace v1.44.0 // indirect
 	go.yaml.in/yaml/v3 v3.0.4 // indirect
-	golang.org/x/crypto v0.49.0 // indirect
-	golang.org/x/exp v0.0.0-20260312153236-7ab1446f8b90 // indirect
-	golang.org/x/net v0.52.0 // indirect
+	golang.org/x/crypto v0.52.0 // indirect
+	golang.org/x/exp v0.0.0-20260528193900-50dc527dd6c7 // indirect
+	golang.org/x/net v0.55.0 // indirect
 	golang.org/x/oauth2 v0.36.0 // indirect
 	golang.org/x/time v0.15.0 // indirect
-	google.golang.org/api v0.273.0 // indirect
-	google.golang.org/genai v1.52.0 // indirect
-	google.golang.org/genproto/googleapis/rpc v0.0.0-20260319201613-d00831a3d3e7 // indirect
-	google.golang.org/grpc v1.79.3 // indirect
+	google.golang.org/api v0.282.0 // indirect
+	google.golang.org/genai v1.58.0 // indirect
+	google.golang.org/genproto/googleapis/rpc v0.0.0-20260526163538-3dc84a4a5aaa // indirect
+	google.golang.org/grpc v1.81.1 // indirect
 	google.golang.org/protobuf v1.36.11 // indirect
 	gopkg.in/yaml.v2 v2.4.0 // indirect
 )
 
 require (
 	github.com/aymanbagabas/go-osc52/v2 v2.0.1 // indirect
-	github.com/charmbracelet/x/ansi v0.11.6
+	github.com/charmbracelet/x/ansi v0.11.7
 	github.com/charmbracelet/x/term v0.2.2 // indirect
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
 	github.com/lucasb-eyer/go-colorful v1.4.0 // indirect
-	github.com/mattn/go-isatty v0.0.20 // indirect
-	github.com/mattn/go-runewidth v0.0.21 // indirect
+	github.com/mattn/go-isatty v0.0.22 // indirect
+	github.com/mattn/go-runewidth v0.0.23 // indirect
 	github.com/muesli/cancelreader v0.2.2 // indirect
 	github.com/muesli/termenv v0.16.0 // indirect
 	github.com/rivo/uniseg v0.4.7 // indirect
-	github.com/spf13/pflag v1.0.10 // indirect
+	github.com/spf13/pflag v1.0.10
 	golang.org/x/sync v0.20.0 // indirect
-	golang.org/x/sys v0.42.0 // indirect
-	golang.org/x/text v0.35.0
+	golang.org/x/sys v0.45.0 // indirect
+	golang.org/x/text v0.37.0
 )

go.sum

@@ -1,69 +1,69 @@
 charm.land/bubbles/v2 v2.1.0 h1:YSnNh5cPYlYjPxRrzs5VEn3vwhtEn3jVGRBT3M7/I0g=
 charm.land/bubbles/v2 v2.1.0/go.mod h1:l97h4hym2hvWBVfmJDtrEHHCtkIKeTEb3TTJ4ZOB3wY=
-charm.land/bubbletea/v2 v2.0.2 h1:4CRtRnuZOdFDTWSff9r8QFt/9+z6Emubz3aDMnf/dx0=
-charm.land/bubbletea/v2 v2.0.2/go.mod h1:3LRff2U4WIYXy7MTxfbAQ+AdfM3D8Xuvz2wbsOD9OHQ=
-charm.land/fantasy v0.17.1 h1:SQzfnyJPDuQWt6e//KKmQmEEXdqHMC0IZz10XwkLcEM=
-charm.land/fantasy v0.17.1/go.mod h1:FF5ALCCHETacHJPBqU42CtwMInYQ0ul52fdzIHQMbQk=
+charm.land/bubbletea/v2 v2.0.6 h1:UHN/91OyuhaOFGSrBXQ/hMZD8IO1Uc4BvHlgHXL2WJo=
+charm.land/bubbletea/v2 v2.0.6/go.mod h1:MH/D8ZLlN3op37vQvijKuU29g3rqTp+aQapURFonF9g=
+charm.land/fantasy v0.25.0 h1:oXOWY1ivmTSnhYGzAolscF8zKtavWZyBWv0LHRSwN5Q=
+charm.land/fantasy v0.25.0/go.mod h1:8QrWUzIcKwZQP+aAnC9vLu3iID6hu9/Jt+rPMiieBkc=
 charm.land/huh/v2 v2.0.3 h1:2cJsMqEPwSywGHvdlKsJyQKPtSJLVnFKyFbsYZTlLkU=
 charm.land/huh/v2 v2.0.3/go.mod h1:93eEveeeqn47MwiC3tf+2atZ2l7Is88rAtmZNZ8x9Wc=
-charm.land/lipgloss/v2 v2.0.2 h1:xFolbF8JdpNkM2cEPTfXEcW1p6NRzOWTSamRfYEw8cs=
-charm.land/lipgloss/v2 v2.0.2/go.mod h1:KjPle2Qd3YmvP1KL5OMHiHysGcNwq6u83MUjYkFvEkM=
+charm.land/lipgloss/v2 v2.0.3 h1:yM2zJ4Cf5Y51b7RHIwioil4ApI/aypFXXVHSwlM6RzU=
+charm.land/lipgloss/v2 v2.0.3/go.mod h1:7myLU9iG/3xluAWzpY/fSxYYHCgoKTie7laxk6ATwXA=
 cloud.google.com/go v0.123.0 h1:2NAUJwPR47q+E35uaJeYoNhuNEM9kM8SjgRgdeOJUSE=
 cloud.google.com/go v0.123.0/go.mod h1:xBoMV08QcqUGuPW65Qfm1o9Y4zKZBpGS+7bImXLTAZU=
-cloud.google.com/go/auth v0.19.0 h1:DGYwtbcsGsT1ywuxsIoWi1u/vlks0moIblQHgSDgQkQ=
-cloud.google.com/go/auth v0.19.0/go.mod h1:2Aph7BT2KnaSFOM0JDPyiYgNh6PL9vGMiP8CUIXZ+IY=
+cloud.google.com/go/auth v0.20.0 h1:kXTssoVb4azsVDoUiF8KvxAqrsQcQtB53DcSgta74CA=
+cloud.google.com/go/auth v0.20.0/go.mod h1:942/yi/itH1SsmpyrbnTMDgGfdy2BUqIKyd0cyYLc5Q=
 cloud.google.com/go/auth/oauth2adapt v0.2.8 h1:keo8NaayQZ6wimpNSmW5OPc283g65QNIiLpZnkHRbnc=
 cloud.google.com/go/auth/oauth2adapt v0.2.8/go.mod h1:XQ9y31RkqZCcwJWNSx2Xvric3RrU88hAYYbjDWYDL+c=
 cloud.google.com/go/compute/metadata v0.9.0 h1:pDUj4QMoPejqq20dK0Pg2N4yG9zIkYGdBtwLoEkH9Zs=
 cloud.google.com/go/compute/metadata v0.9.0/go.mod h1:E0bWwX5wTnLPedCKqk3pJmVgCBSM6qQI1yTBdEb3C10=
-github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.0 h1:fou+2+WFTib47nS+nz/ozhEBnvU96bKHy6LjRsY4E28=
-github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.0/go.mod h1:t76Ruy8AHvUAC8GfMWJMa0ElSbuIcO03NLpynfbgsPA=
-github.com/Azure/azure-sdk-for-go/sdk/azidentity v1.10.1 h1:B+blDbyVIG3WaikNxPnhPiJ1MThR03b3vKGtER95TP4=
-github.com/Azure/azure-sdk-for-go/sdk/azidentity v1.10.1/go.mod h1:JdM5psgjfBf5fo2uWOZhflPWyDBZ/O/CNAH9CtsuZE4=
-github.com/Azure/azure-sdk-for-go/sdk/internal v1.11.2 h1:9iefClla7iYpfYWdzPCRDozdmndjTm8DXdpCzPajMgA=
-github.com/Azure/azure-sdk-for-go/sdk/internal v1.11.2/go.mod h1:XtLgD3ZD34DAaVIIAyG3objl5DynM3CQ/vMcbBNJZGI=
-github.com/AzureAD/microsoft-authentication-library-for-go v1.4.2 h1:oygO0locgZJe7PpYPXT5A29ZkwJaPqcva7BVeemZOZs=
-github.com/AzureAD/microsoft-authentication-library-for-go v1.4.2/go.mod h1:wP83P5OoQ5p6ip3ScPr0BAq0BvuPAvacpEuSzyouqAI=
+github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.1 h1:jHb/wfvRikGdxMXYV3QG/SzUOPYN9KEUUuC0Yd0/vC0=
+github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.1/go.mod h1:pzBXCYn05zvYIrwLgtK8Ap8QcjRg+0i76tMQdWN6wOk=
+github.com/Azure/azure-sdk-for-go/sdk/azidentity v1.13.1 h1:Hk5QBxZQC1jb2Fwj6mpzme37xbCDdNTxU7O9eb5+LB4=
+github.com/Azure/azure-sdk-for-go/sdk/azidentity v1.13.1/go.mod h1:IYus9qsFobWIc2YVwe/WPjcnyCkPKtnHAqUYeebc8z0=
+github.com/Azure/azure-sdk-for-go/sdk/internal v1.12.0 h1:fhqpLE3UEXi9lPaBRpQ6XuRW0nU7hgg4zlmZZa+a9q4=
+github.com/Azure/azure-sdk-for-go/sdk/internal v1.12.0/go.mod h1:7dCRMLwisfRH3dBupKeNCioWYUZ4SS09Z14H+7i8ZoY=
+github.com/AzureAD/microsoft-authentication-library-for-go v1.6.0 h1:XRzhVemXdgvJqCH0sFfrBUTnUJSBrBf7++ypk+twtRs=
+github.com/AzureAD/microsoft-authentication-library-for-go v1.6.0/go.mod h1:HKpQxkWaGLJ+D/5H8QRpyQXA1eKjxkFlOMwck5+33Jk=
 github.com/MakeNowJust/heredoc v1.0.0 h1:cXCdzVdstXyiTqTvfqk9SDHpKNjxuom+DOlyEeQ4pzQ=
 github.com/MakeNowJust/heredoc v1.0.0/go.mod h1:mG5amYoWBHf8vpLOuehzbGGw0EHxpZZ6lCpQ4fNJ8LE=
 github.com/alecthomas/assert/v2 v2.11.0 h1:2Q9r3ki8+JYXvGsDyBXwH3LcJ+WK5D0gc5E8vS6K3D0=
 github.com/alecthomas/assert/v2 v2.11.0/go.mod h1:Bze95FyfUr7x34QZrjL+XP+0qgp/zg8yS+TtBj1WA3k=
-github.com/alecthomas/chroma/v2 v2.23.1 h1:nv2AVZdTyClGbVQkIzlDm/rnhk1E9bU9nXwmZ/Vk/iY=
-github.com/alecthomas/chroma/v2 v2.23.1/go.mod h1:NqVhfBR0lte5Ouh3DcthuUCTUpDC9cxBOfyMbMQPs3o=
+github.com/alecthomas/chroma/v2 v2.26.1 h1:2X21EdxGZNv5GF9mG5u+uzc02GCFyGxbcBm3Grd9A78=
+github.com/alecthomas/chroma/v2 v2.26.1/go.mod h1:lxhRRa9H4hPmRLOOdYga4zkQIQjq3dtrrdwQeCfu78Y=
 github.com/alecthomas/repr v0.5.2 h1:SU73FTI9D1P5UNtvseffFSGmdNci/O6RsqzeXJtP0Qs=
 github.com/alecthomas/repr v0.5.2/go.mod h1:Fr0507jx4eOXV7AlPV6AVZLYrLIuIeSOWtW57eE/O/4=
 github.com/atotto/clipboard v0.1.4 h1:EH0zSVneZPSuFR11BlR9YppQTVDbh5+16AmcJi4g1z4=
 github.com/atotto/clipboard v0.1.4/go.mod h1:ZY9tmq7sm5xIbd9bOK4onWV4S6X0u6GY7Vn0Yu86PYI=
-github.com/aws/aws-sdk-go-v2 v1.41.5 h1:dj5kopbwUsVUVFgO4Fi5BIT3t4WyqIDjGKCangnV/yY=
-github.com/aws/aws-sdk-go-v2 v1.41.5/go.mod h1:mwsPRE8ceUUpiTgF7QmQIJ7lgsKUPQOUl3o72QBrE1o=
-github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.8 h1:eBMB84YGghSocM7PsjmmPffTa+1FBUeNvGvFou6V/4o=
-github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.8/go.mod h1:lyw7GFp3qENLh7kwzf7iMzAxDn+NzjXEAGjKS2UOKqI=
-github.com/aws/aws-sdk-go-v2/config v1.32.13 h1:5KgbxMaS2coSWRrx9TX/QtWbqzgQkOdEa3sZPhBhCSg=
-github.com/aws/aws-sdk-go-v2/config v1.32.13/go.mod h1:8zz7wedqtCbw5e9Mi2doEwDyEgHcEE9YOJp6a8jdSMY=
-github.com/aws/aws-sdk-go-v2/credentials v1.19.13 h1:mA59E3fokBvyEGHKFdnpNNrvaR351cqiHgRg+JzOSRI=
-github.com/aws/aws-sdk-go-v2/credentials v1.19.13/go.mod h1:yoTXOQKea18nrM69wGF9jBdG4WocSZA1h38A+t/MAsk=
-github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.18.21 h1:NUS3K4BTDArQqNu2ih7yeDLaS3bmHD0YndtA6UP884g=
-github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.18.21/go.mod h1:YWNWJQNjKigKY1RHVJCuupeWDrrHjRqHm0N9rdrWzYI=
-github.com/aws/aws-sdk-go-v2/internal/configsources v1.4.21 h1:Rgg6wvjjtX8bNHcvi9OnXWwcE0a2vGpbwmtICOsvcf4=
-github.com/aws/aws-sdk-go-v2/internal/configsources v1.4.21/go.mod h1:A/kJFst/nm//cyqonihbdpQZwiUhhzpqTsdbhDdRF9c=
-github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.7.21 h1:PEgGVtPoB6NTpPrBgqSE5hE/o47Ij9qk/SEZFbUOe9A=
-github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.7.21/go.mod h1:p+hz+PRAYlY3zcpJhPwXlLC4C+kqn70WIHwnzAfs6ps=
-github.com/aws/aws-sdk-go-v2/internal/ini v1.8.6 h1:qYQ4pzQ2Oz6WpQ8T3HvGHnZydA72MnLuFK9tJwmrbHw=
-github.com/aws/aws-sdk-go-v2/internal/ini v1.8.6/go.mod h1:O3h0IK87yXci+kg6flUKzJnWeziQUKciKrLjcatSNcY=
-github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.7 h1:5EniKhLZe4xzL7a+fU3C2tfUN4nWIqlLesfrjkuPFTY=
-github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.7/go.mod h1:x0nZssQ3qZSnIcePWLvcoFisRXJzcTVvYpAAdYX8+GI=
-github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.13.21 h1:c31//R3xgIJMSC8S6hEVq+38DcvUlgFY0FM6mSI5oto=
-github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.13.21/go.mod h1:r6+pf23ouCB718FUxaqzZdbpYFyDtehyZcmP5KL9FkA=
-github.com/aws/aws-sdk-go-v2/service/signin v1.0.9 h1:QKZH0S178gCmFEgst8hN0mCX1KxLgHBKKY/CLqwP8lg=
-github.com/aws/aws-sdk-go-v2/service/signin v1.0.9/go.mod h1:7yuQJoT+OoH8aqIxw9vwF+8KpvLZ8AWmvmUWHsGQZvI=
-github.com/aws/aws-sdk-go-v2/service/sso v1.30.14 h1:GcLE9ba5ehAQma6wlopUesYg/hbcOhFNWTjELkiWkh4=
-github.com/aws/aws-sdk-go-v2/service/sso v1.30.14/go.mod h1:WSvS1NLr7JaPunCXqpJnWk1Bjo7IxzZXrZi1QQCkuqM=
-github.com/aws/aws-sdk-go-v2/service/ssooidc v1.35.18 h1:mP49nTpfKtpXLt5SLn8Uv8z6W+03jYVoOSAl/c02nog=
-github.com/aws/aws-sdk-go-v2/service/ssooidc v1.35.18/go.mod h1:YO8TrYtFdl5w/4vmjL8zaBSsiNp3w0L1FfKVKenZT7w=
-github.com/aws/aws-sdk-go-v2/service/sts v1.41.10 h1:p8ogvvLugcR/zLBXTXrTkj0RYBUdErbMnAFFp12Lm/U=
-github.com/aws/aws-sdk-go-v2/service/sts v1.41.10/go.mod h1:60dv0eZJfeVXfbT1tFJinbHrDfSJ2GZl4Q//OSSNAVw=
-github.com/aws/smithy-go v1.24.2 h1:FzA3bu/nt/vDvmnkg+R8Xl46gmzEDam6mZ1hzmwXFng=
-github.com/aws/smithy-go v1.24.2/go.mod h1:YE2RhdIuDbA5E5bTdciG9KrW3+TiEONeUWCqxX9i1Fc=
+github.com/aws/aws-sdk-go-v2 v1.41.8 h1:sRs7nG6/RiEBZ/K5UO2sNw0w40U02Nmz1VtARloTZXk=
+github.com/aws/aws-sdk-go-v2 v1.41.8/go.mod h1:4LAfZOPHNVNQEckOACQx60Y8pSRjIkNZQz1w92xpMJc=
+github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.10 h1:gx1AwW1Iyk9Z9dD9F4akX5gnN3QZwUB20GGKH/I+Rho=
+github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.10/go.mod h1:qqY157uZoqm5OXq/amuaBJyC9hgBCBQnsaWnPe905GY=
+github.com/aws/aws-sdk-go-v2/config v1.32.19 h1:qRhIJMbevHUvIE7X4TK8N8zye5+5AhapcslPrvB+qKE=
+github.com/aws/aws-sdk-go-v2/config v1.32.19/go.mod h1:RbJ24nfoya63+Mf5VI+CGCGk9vEdv28xPeii+gojRYs=
+github.com/aws/aws-sdk-go-v2/credentials v1.19.18 h1:GcXQz2M/0ZvMo0v5DakUqbDBeBM1ZNaivkolEF4Esgw=
+github.com/aws/aws-sdk-go-v2/credentials v1.19.18/go.mod h1:sHJ06tMGcD3ZpmMyJqV+VBsGilhSIZPIN+ZFy5Dg0C4=
+github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.18.24 h1:FQm5ApnyzkuJdXLGskPce83CK1CQKC4RUnIHKVe4BU4=
+github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.18.24/go.mod h1:JsC7dqQc55MlZ5mvNsDMMge71u8pVcSzU3RNz2h/5yQ=
+github.com/aws/aws-sdk-go-v2/internal/configsources v1.4.24 h1:u6kJU2i0va1AgtJsH3RdWKWqHULlTh7zHwb35Womf74=
+github.com/aws/aws-sdk-go-v2/internal/configsources v1.4.24/go.mod h1:7GY+xLcXOFUpCkNwDReft9qOAVg54A4/AnjHIU7sSAY=
+github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.7.24 h1:Xhbcf3KugX6vX7SDyUK205Oicyfg7EGuvoVNyP5L6DM=
+github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.7.24/go.mod h1:rwDgb2HNOGZsnTHylOUedM7Vnl+bCfnXDqUNPsFWYfk=
+github.com/aws/aws-sdk-go-v2/internal/v4a v1.4.25 h1:54CTMmlJ71Rk2dYvM9qZOob+39wjlVja2zDLxCu69Ew=
+github.com/aws/aws-sdk-go-v2/internal/v4a v1.4.25/go.mod h1:BZaHqxsS9vN1fvV5EfEl0OBLOk5+AajWsMu6MjqnZB4=
+github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.9 h1:FLudkZLt5ci0ozzgkVo8BJGwvqNaZbTWb3UcucAateA=
+github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.9/go.mod h1:w7wZ/s9qK7c8g4al+UyoF1Sp/Z45UwMGcqIzLWVQHWk=
+github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.13.24 h1:CQW2FTrflfoslYWLf3fv7vG28Q219+v8YJS5QTQb2+Y=
+github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.13.24/go.mod h1:Xfx13T+u3nH6EEzgl9fBSO6nDRmze1FvnZNYkctQ2zw=
+github.com/aws/aws-sdk-go-v2/service/signin v1.1.0 h1:yQo3eZ5qFaL1sJWqs1nL6j3yPHA2/R7c6tQ4T+0IO10=
+github.com/aws/aws-sdk-go-v2/service/signin v1.1.0/go.mod h1:3Zzou41Qt/ueXfIzHvTEjDNuR5IjCUBVF01SNhrt1e8=
+github.com/aws/aws-sdk-go-v2/service/sso v1.30.18 h1:ApLTFdAZfDhZSiY5uskwECKHkSNNF83y2Ru2r7SezWA=
+github.com/aws/aws-sdk-go-v2/service/sso v1.30.18/go.mod h1:A9K9qx2l6nK89hp+a350FdGfRkrkH5HdiEjHbiy/Q/c=
+github.com/aws/aws-sdk-go-v2/service/ssooidc v1.36.1 h1:4VD7TIZOGzehrgQ8vDE+1c6BQW4ErZPGY8ohZT5LXEE=
+github.com/aws/aws-sdk-go-v2/service/ssooidc v1.36.1/go.mod h1:er0SFJfdV89Rit5hIJu/EXtv+qC2XMnxoksLmcUFkqM=
+github.com/aws/aws-sdk-go-v2/service/sts v1.42.2 h1:XKnxlM4KZH1gktcsh3zSWc7GW4KivEv/OkifmHOhCUY=
+github.com/aws/aws-sdk-go-v2/service/sts v1.42.2/go.mod h1:KJYmkQaFB3SUW2j3aBkPsxNmAb4ZsSOvbvCpuxzHJA0=
+github.com/aws/smithy-go v1.26.0 h1:9ouqbi+NyKP7fV3Te7UElCwdAb6Y8uk7LGwPE5tVe/s=
+github.com/aws/smithy-go v1.26.0/go.mod h1:YE2RhdIuDbA5E5bTdciG9KrW3+TiEONeUWCqxX9i1Fc=
 github.com/aymanbagabas/go-osc52/v2 v2.0.1 h1:HwpRHbFMcZLEVr42D4p7XBqjyuxQH5SMiErDT4WkJ2k=
 github.com/aymanbagabas/go-osc52/v2 v2.0.1/go.mod h1:uYgXzlJ7ZpABp8OJ+exZzJJhRNQ2ASbcXHWsFqH8hp8=
 github.com/aymanbagabas/go-udiff v0.4.1 h1:OEIrQ8maEeDBXQDoGCbbTTXYJMYRCRO1fnodZ12Gv5o=
@@ -86,24 +86,26 @@ github.com/charmbracelet/log v1.0.0 h1:HVVVMmfOorfj3BA9i8X8UL69Hoz9lI0PYwXfJvOdR
 github.com/charmbracelet/log v1.0.0/go.mod h1:uYgY3SmLpwJWxmlrPwXvzVYujxis1vAKRV/0VQB7yWA=
 github.com/charmbracelet/openai-go v0.0.0-20260319145158-d0740cc34266 h1:BW/sZtyd1JyYy0h5adMm3tzpNyL857LWjuTRET6OhpY=
 github.com/charmbracelet/openai-go v0.0.0-20260319145158-d0740cc34266/go.mod h1:1DahUaExbUZx/jD+FNT2PKP4L9rLE5+ZBRuI8mZjd/E=
-github.com/charmbracelet/ultraviolet v0.0.0-20260330092749-0f94982c930b h1:ASDO9RT6SNKTQN87jO2bRfxHFJq8cgeYdFzivY2gCeM=
-github.com/charmbracelet/ultraviolet v0.0.0-20260330092749-0f94982c930b/go.mod h1:Vo8TffMf0q7Uho/n8e6XpBZvOWtd3g39yX+9P5rRutA=
-github.com/charmbracelet/x/ansi v0.11.6 h1:GhV21SiDz/45W9AnV2R61xZMRri5NlLnl6CVF7ihZW8=
-github.com/charmbracelet/x/ansi v0.11.6/go.mod h1:2JNYLgQUsyqaiLovhU2Rv/pb8r6ydXKS3NIttu3VGZQ=
+github.com/charmbracelet/ultraviolet v0.0.0-20260525132238-948f4557a654 h1:FpSYhY28ucg9ZRr+2wj67FAQ0Ey5yiK0072PmRDJNek=
+github.com/charmbracelet/ultraviolet v0.0.0-20260525132238-948f4557a654/go.mod h1:hFpumms29Smx3LStRfku8vcCTBe1Kq8aCXtHUJa3mjY=
+github.com/charmbracelet/x/ansi v0.11.7 h1:kzv1kJvjg2S3r9KHo8hDdHFQLEqn4RBCb39dAYC84jI=
+github.com/charmbracelet/x/ansi v0.11.7/go.mod h1:9qGpnAVYz+8ACONkZBUWPtL7lulP9No6p1epAihUZwQ=
 github.com/charmbracelet/x/cellbuf v0.0.15 h1:ur3pZy0o6z/R7EylET877CBxaiE1Sp1GMxoFPAIztPI=
 github.com/charmbracelet/x/cellbuf v0.0.15/go.mod h1:J1YVbR7MUuEGIFPCaaZ96KDl5NoS0DAWkskup+mOY+Q=
 github.com/charmbracelet/x/conpty v0.1.1 h1:s1bUxjoi7EpqiXysVtC+a8RrvPPNcNvAjfi4jxsAuEs=
 github.com/charmbracelet/x/conpty v0.1.1/go.mod h1:OmtR77VODEFbiTzGE9G1XiRJAga6011PIm4u5fTNZpk=
+github.com/charmbracelet/x/editor v0.2.0 h1:7XLUKtaRaB8jN7bWU2p2UChiySyaAuIfYiIRg8gGWwk=
+github.com/charmbracelet/x/editor v0.2.0/go.mod h1:p3oQ28TSL3YPd+GKJ1fHWcp+7bVGpedHpXmo0D6t1dY=
 github.com/charmbracelet/x/errors v0.0.0-20240508181413-e8d8b6e2de86 h1:JSt3B+U9iqk37QUU2Rvb6DSBYRLtWqFqfxf8l5hOZUA=
 github.com/charmbracelet/x/errors v0.0.0-20240508181413-e8d8b6e2de86/go.mod h1:2P0UgXMEa6TsToMSuFqKFQR+fZTO9CNGUNokkPatT/0=
-github.com/charmbracelet/x/exp/charmtone v0.0.0-20260330094520-2dce04b6f8a4 h1:pIj18ZCZO4WOVj7jwjLoUb1lC7rS/I8oC3fZWXugNaY=
-github.com/charmbracelet/x/exp/charmtone v0.0.0-20260330094520-2dce04b6f8a4/go.mod h1:nsExn0DGyX0lh9LwLHTn2Gg+hafdzfSXnC+QmEJTZFY=
+github.com/charmbracelet/x/exp/charmtone v0.0.0-20260527151214-009e6338d40d h1:sMilwx1YIYTrQva6jsB522AoRYAerNaDIKP4ZPtUq0A=
+github.com/charmbracelet/x/exp/charmtone v0.0.0-20260527151214-009e6338d40d/go.mod h1:nsExn0DGyX0lh9LwLHTn2Gg+hafdzfSXnC+QmEJTZFY=
 github.com/charmbracelet/x/exp/golden v0.0.0-20250806222409-83e3a29d542f h1:pk6gmGpCE7F3FcjaOEKYriCvpmIN4+6OS/RD0vm4uIA=
 github.com/charmbracelet/x/exp/golden v0.0.0-20250806222409-83e3a29d542f/go.mod h1:IfZAMTHB6XkZSeXUqriemErjAWCCzT0LwjKFYCZyw0I=
 github.com/charmbracelet/x/exp/ordered v0.1.0 h1:55/qLwjIh0gL0Vni+QAWk7T/qRVP6sBf+2agPBgnOFE=
 github.com/charmbracelet/x/exp/ordered v0.1.0/go.mod h1:5UHwmG+is5THxMyCJHNPCn2/ecI07aKNrW+LcResjJ8=
-github.com/charmbracelet/x/exp/slice v0.0.0-20260330094520-2dce04b6f8a4 h1:VSd4zShIAf/4FgEDFJpapEcAPrc7h3dyyN7V9JlJpQw=
-github.com/charmbracelet/x/exp/slice v0.0.0-20260330094520-2dce04b6f8a4/go.mod h1:vqEfX6xzqW1pKKZUUiFOKg0OQ7bCh54Q2vR/tserrRA=
+github.com/charmbracelet/x/exp/slice v0.0.0-20260527151214-009e6338d40d h1:RxcAR+vJCoD8QqT1cqLtkQKw+1cqvjqnu5IpPqYzPco=
+github.com/charmbracelet/x/exp/slice v0.0.0-20260527151214-009e6338d40d/go.mod h1:vqEfX6xzqW1pKKZUUiFOKg0OQ7bCh54Q2vR/tserrRA=
 github.com/charmbracelet/x/exp/strings v0.1.0 h1:i69S2XI7uG1u4NLGeJPSYU++Nmjvpo9nwd6aoEm7gkA=
 github.com/charmbracelet/x/exp/strings v0.1.0/go.mod h1:/ehtMPNh9K4odGFkqYJKpIYyePhdp1hLBRvyY4bWkH8=
 github.com/charmbracelet/x/json v0.2.0 h1:DqB+ZGx2h+Z+1s98HOuOyli+i97wsFQIxP2ZQANTPrQ=
@@ -122,15 +124,17 @@ github.com/clipperhouse/uax29/v2 v2.7.0 h1:+gs4oBZ2gPfVrKPthwbMzWZDaAFPGYK72F0NJ
 github.com/clipperhouse/uax29/v2 v2.7.0/go.mod h1:EFJ2TJMRUaplDxHKj1qAEhCtQPW2tJSwu5BF98AuoVM=
 github.com/cncf/xds/go v0.0.0-20260202195803-dba9d589def2 h1:aBangftG7EVZoUb69Os8IaYg++6uMOdKK83QtkkvJik=
 github.com/cncf/xds/go v0.0.0-20260202195803-dba9d589def2/go.mod h1:qwXFYgsP6T7XnJtbKlf1HP8AjxZZyzxMmc+Lq5GjlU4=
-github.com/coder/acp-go-sdk v0.6.3 h1:LsXQytehdjKIYJnoVWON/nf7mqbiarnyuyE3rrjBsXQ=
-github.com/coder/acp-go-sdk v0.6.3/go.mod h1:yKzM/3R9uELp4+nBAwwtkS0aN1FOFjo11CNPy37yFko=
+github.com/coder/acp-go-sdk v0.13.0 h1:IAKBDIbe/iBfKAGikeIndzb8fowt4ioD+gCtSU4HwMA=
+github.com/coder/acp-go-sdk v0.13.0/go.mod h1:yKzM/3R9uELp4+nBAwwtkS0aN1FOFjo11CNPy37yFko=
 github.com/cpuguy83/go-md2man/v2 v2.0.6/go.mod h1:oOW0eioCTA6cOiMLiUPZOpcVxMig6NIQQ7OS05n1F4g=
 github.com/creack/pty v1.1.24 h1:bJrF4RRfyJnbTJqzRLHzcGaZK1NeM5kTC9jGgovnR1s=
 github.com/creack/pty v1.1.24/go.mod h1:08sCNb52WyoAwi2QDyzUCTgcvVFhUzewun7wtTfvcwE=
 github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc h1:U9qPSI2PIWSS1VwoXQT9A3Wy9MM3WgvqSxFWenqJduM=
 github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/dlclark/regexp2 v1.11.5 h1:Q/sSnsKerHeCkc/jSTNq1oCm7KiVgUMZRDUoRu0JQZQ=
-github.com/dlclark/regexp2 v1.11.5/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
+github.com/dlclark/regexp2 v1.12.0 h1:0j4c5qQmnC6XOWNjP3PIXURXN2gWx76rd3KvgdPkCz8=
+github.com/dlclark/regexp2 v1.12.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
+github.com/dlclark/regexp2/v2 v2.1.1 h1:LCUGyd9Wf+r+VVOl8Ny38JTpWJcAsdVnCIuhhtthmKw=
+github.com/dlclark/regexp2/v2 v2.1.1/go.mod h1:avUrQvPaLz2DrFNHJF0taWAFFX2C1GMSSoeiqFjcBmU=
 github.com/dnaeon/go-vcr v1.2.0 h1:zHCHvJYTMh1N7xnV7zf1m1GPBF9Ad0Jk/whtQ1663qI=
 github.com/dnaeon/go-vcr v1.2.0/go.mod h1:R4UdLID7HZT3taECzJs4YgbbH6PIGXB6W/sc5OLb6RQ=
 github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
@@ -144,10 +148,10 @@ github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2
 github.com/felixge/httpsnoop v1.0.4/go.mod h1:m8KPJKqk1gH5J9DgRY2ASl2lWCfGKXixSwevea8zH2U=
 github.com/frankban/quicktest v1.14.6 h1:7Xjx+VpznH+oBnejlPUj8oUpdxnVs4f8XU8WnHkI4W8=
 github.com/frankban/quicktest v1.14.6/go.mod h1:4ptaffx2x8+WTWXmUCuVU6aPUX1/Mz7zb5vbUoiM6w0=
-github.com/fsnotify/fsnotify v1.9.0 h1:2Ml+OJNzbYCTzsxtv8vKSFD9PbJjmhYF14k/jKC7S9k=
-github.com/fsnotify/fsnotify v1.9.0/go.mod h1:8jBTzvmWwFyi3Pb8djgCCO5IBqzKJ/Jwo8TRcHyHii0=
-github.com/go-json-experiment/json v0.0.0-20260214004413-d219187c3433 h1:vymEbVwYFP/L05h5TKQxvkXoKxNvTpjxYKdF1Nlwuao=
-github.com/go-json-experiment/json v0.0.0-20260214004413-d219187c3433/go.mod h1:tphK2c80bpPhMOI4v6bIc2xWywPfbqi1Z06+RcrMkDg=
+github.com/fsnotify/fsnotify v1.10.1 h1:b0/UzAf9yR5rhf3RPm9gf3ehBPpf0oZKIjtpKrx59Ho=
+github.com/fsnotify/fsnotify v1.10.1/go.mod h1:TLheqan6HD6GBK6PrDWyDPBaEV8LspOxvPSjC+bVfgo=
+github.com/go-json-experiment/json v0.0.0-20260520185125-572e7c383686 h1:NZBJxCpbHS1gzS6xAmyxbJznosZIIPk9IB42v62UvKA=
+github.com/go-json-experiment/json v0.0.0-20260520185125-572e7c383686/go.mod h1:tphK2c80bpPhMOI4v6bIc2xWywPfbqi1Z06+RcrMkDg=
 github.com/go-logfmt/logfmt v0.6.1 h1:4hvbpePJKnIzH1B+8OR/JPbTx37NktoI9LE2QZBBkvE=
 github.com/go-logfmt/logfmt v0.6.1/go.mod h1:EV2pOAQoZaT1ZXZbqDl5hrymndi4SY9ED9/z6CO0XAk=
 github.com/go-logr/logr v1.2.2/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A=
@@ -165,34 +169,34 @@ github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek
 github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
 github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
 github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
-github.com/google/jsonschema-go v0.4.2 h1:tmrUohrwoLZZS/P3x7ex0WAVknEkBZM46iALbcqoRA8=
-github.com/google/jsonschema-go v0.4.2/go.mod h1:r5quNTdLOYEz95Ru18zA0ydNbBuYoo9tgaYcxEYhJVE=
+github.com/google/jsonschema-go v0.4.3 h1:/DBOLZTfDow7pe2GmaJNhltueGTtDKICi8V8p+DQPd0=
+github.com/google/jsonschema-go v0.4.3/go.mod h1:r5quNTdLOYEz95Ru18zA0ydNbBuYoo9tgaYcxEYhJVE=
 github.com/google/s2a-go v0.1.9 h1:LGD7gtMgezd8a/Xak7mEWL0PjoTQFvpRudN895yqKW0=
 github.com/google/s2a-go v0.1.9/go.mod h1:YA0Ei2ZQL3acow2O62kdp9UlnvMmU7kA6Eutn0dXayM=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
-github.com/googleapis/enterprise-certificate-proxy v0.3.14 h1:yh8ncqsbUY4shRD5dA6RlzjJaT4hi3kII+zYw8wmLb8=
-github.com/googleapis/enterprise-certificate-proxy v0.3.14/go.mod h1:vqVt9yG9480NtzREnTlmGSBmFrA+bzb0yl0TxoBQXOg=
-github.com/googleapis/gax-go/v2 v2.20.0 h1:NIKVuLhDlIV74muWlsMM4CcQZqN6JJ20Qcxd9YMuYcs=
-github.com/googleapis/gax-go/v2 v2.20.0/go.mod h1:But/NJU6TnZsrLai/xBAQLLz+Hc7fHZJt/hsCz3Fih4=
+github.com/googleapis/enterprise-certificate-proxy v0.3.16 h1:F/VPrx0YPBdksZJQdCAp0WUsqnNmZpUZszzfYt0M5Dw=
+github.com/googleapis/enterprise-certificate-proxy v0.3.16/go.mod h1:9Yb0eAkH/Xqhvv3zbeKf/+wMJqCeocWc6KIhDvEAuYE=
+github.com/googleapis/gax-go/v2 v2.22.0 h1:PjIWBpgGIVKGoCXuiCoP64altEJCj3/Ei+kSU5vlZD4=
+github.com/googleapis/gax-go/v2 v2.22.0/go.mod h1:irWBbALSr0Sk3qlqb9SyJ1h68WjgeFuiOzI4Rqw5+aY=
 github.com/gorilla/websocket v1.5.3 h1:saDtZ6Pbx/0u+bgYQ3q96pZgCzfhKXGPqt7kZ72aNNg=
 github.com/gorilla/websocket v1.5.3/go.mod h1:YR8l580nyteQvAITg2hZ9XVh4b55+EU/adAjf1fMHhE=
 github.com/hexops/gotextdiff v1.0.3 h1:gitA9+qJrrTCsiCl7+kh75nPqQt1cx4ZkudSTLoUqJM=
 github.com/hexops/gotextdiff v1.0.3/go.mod h1:pSWU5MAI3yDq+fZBTazCSJysOMbxWL1BSow5/V2vxeg=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
-github.com/indaco/herald v0.10.0 h1:XzahEKX6cr50qZQrUdA3QrQBHg8uGm5jETD0UDi21BI=
-github.com/indaco/herald v0.10.0/go.mod h1:T5g1+XLYvpjouhzAGHnAHDCKizhESkoV6+QPZ3DhgWA=
-github.com/indaco/herald-md v0.1.0 h1:zmYudYo+uamzKTBcIffJVJYrqk9xDNnVrTh+de2zciw=
-github.com/indaco/herald-md v0.1.0/go.mod h1:Z1HxPCbSn+/+TFzOM/UbsmKeEk/28NNI6JOTileKXto=
-github.com/kaptinlin/go-i18n v0.3.0 h1:wP76dvYg04bvwTb+8NB+CmdZ2kL7lSSCQ9B/kFv7QHo=
-github.com/kaptinlin/go-i18n v0.3.0/go.mod h1:pVcu9qsW5pOIOoZFJXesRYmLos1vMQrby70JPAoWmJU=
-github.com/kaptinlin/jsonpointer v0.4.17 h1:mY9k8ciWncxbsECyaxKnR0MdmxamNdp2tLQkAKVrtSk=
-github.com/kaptinlin/jsonpointer v0.4.17/go.mod h1:SsfsjqnHG5zuKo1DTBzk1VknaHlL4osHw+X9kZKukpU=
-github.com/kaptinlin/jsonschema v0.7.7 h1:41BlQJ9dskH0oE5DSzBUrl/w4JQYIr6N6L0B5GNyDoM=
-github.com/kaptinlin/jsonschema v0.7.7/go.mod h1:rKjWfyySHSxAD7Li2ctYkPlOu960igoKBvZ2ADRtd5Q=
-github.com/kaptinlin/messageformat-go v0.4.19 h1:A5kuuZ1ybXDQ7kD1aoEWGAOemX7hLsMY0yolgSbgpRI=
-github.com/kaptinlin/messageformat-go v0.4.19/go.mod h1:utSDTfiXTxl66OC5RIEuObLH7Ue3YjbA2X86SYMBYWg=
+github.com/indaco/herald v0.13.0 h1:+xVG9Fx5NpuWhwku/9IlRL6I009NnX4VUGKvlZHTRxU=
+github.com/indaco/herald v0.13.0/go.mod h1:T5g1+XLYvpjouhzAGHnAHDCKizhESkoV6+QPZ3DhgWA=
+github.com/indaco/herald-md v0.3.0 h1:hN1cKyrexPPM9PeHBsKuaWvIizSi/iYvM9yzRgtdb8M=
+github.com/indaco/herald-md v0.3.0/go.mod h1:RUHVaDSG45ymJjKyxpDwBocLXrZo93FB4OeYMsw9B9s=
+github.com/kaptinlin/go-i18n v0.4.5 h1:9tIlo5A0RXth+yZJO2MG7Bhpu/X9PlzQnGz/qyYWNoY=
+github.com/kaptinlin/go-i18n v0.4.5/go.mod h1:mU/7BH4molY5lGZYBwBRKAaiJ70dWRHuqmQ0/pFLGno=
+github.com/kaptinlin/jsonpointer v0.4.25 h1:iJ197e8n+WwqaqBsa53FqG3rPJCg5oijyFXEXNWWC3E=
+github.com/kaptinlin/jsonpointer v0.4.25/go.mod h1:wVOBaXGGnP42YsMb6zev/3W5POTvspdNfh8DXzf8XS8=
+github.com/kaptinlin/jsonschema v0.7.13 h1:kahVXTy/rURL0XJjyQ9WELm59wEmXi6IY0TWswQEFvU=
+github.com/kaptinlin/jsonschema v0.7.13/go.mod h1:Uh0aUBusnhXDCEXJ2oimL/hx7YTo7F+sKniE+tM0ERc=
+github.com/kaptinlin/messageformat-go v0.6.0 h1:D6jiXFsKW4/JG2CMddv/F6Rev9KVbCRKEzzV5QOAcpc=
+github.com/kaptinlin/messageformat-go v0.6.0/go.mod h1:NKjwS6e9u7DRhAK+vydjDDwJ7UbdHhYjk/yk2WPuZPs=
 github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
 github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
 github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
@@ -201,12 +205,12 @@ github.com/kylelemons/godebug v1.1.0 h1:RPNrshWIDI6G2gRW9EHilWtl7Z6Sb1BR0xunSBf0
 github.com/kylelemons/godebug v1.1.0/go.mod h1:9/0rRGxNHcop5bhtWyNeEfOS8JIWk580+fNqagV/RAw=
 github.com/lucasb-eyer/go-colorful v1.4.0 h1:UtrWVfLdarDgc44HcS7pYloGHJUjHV/4FwW4TvVgFr4=
 github.com/lucasb-eyer/go-colorful v1.4.0/go.mod h1:R4dSotOR9KMtayYi1e77YzuveK+i7ruzyGqttikkLy0=
-github.com/mark3labs/mcp-go v0.46.0 h1:8KRibF4wcKejbLsHxCA/QBVUr5fQ9nwz/n8lGqmaALo=
-github.com/mark3labs/mcp-go v0.46.0/go.mod h1:JKTC7R2LLVagkEWK7Kwu7DbmA6iIvnNAod6yrHiQMag=
-github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
-github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
-github.com/mattn/go-runewidth v0.0.21 h1:jJKAZiQH+2mIinzCJIaIG9Be1+0NR+5sz/lYEEjdM8w=
-github.com/mattn/go-runewidth v0.0.21/go.mod h1:XBkDxAl56ILZc9knddidhrOlY5R/pDhgLpndooCuJAs=
+github.com/mark3labs/mcp-go v0.54.1 h1:Ap/ptEB9FtWzFKM8NDsTA7QDxerQOC06eZigrTldVj0=
+github.com/mark3labs/mcp-go v0.54.1/go.mod h1:+8WclSK1ZUweCP3hvktSji8n8ABG/95QaEkeVE/Uwas=
+github.com/mattn/go-isatty v0.0.22 h1:j8l17JJ9i6VGPUFUYoTUKPSgKe/83EYU2zBC7YNKMw4=
+github.com/mattn/go-isatty v0.0.22/go.mod h1:ZXfXG4SQHsB/w3ZeOYbR0PrPwLy+n6xiMrJlRFqopa4=
+github.com/mattn/go-runewidth v0.0.23 h1:7ykA0T0jkPpzSvMS5i9uoNn2Xy3R383f9HDx3RybWcw=
+github.com/mattn/go-runewidth v0.0.23/go.mod h1:XBkDxAl56ILZc9knddidhrOlY5R/pDhgLpndooCuJAs=
 github.com/mitchellh/hashstructure/v2 v2.0.2 h1:vGKWl0YJqUNxE8d+h8f6NJLcCJrgbhC4NcD46KavDd4=
 github.com/mitchellh/hashstructure/v2 v2.0.2/go.mod h1:MG3aRVU/N29oo/V/IhBX8GR/zz4kQkprJgF2EVszyDE=
 github.com/muesli/cancelreader v0.2.2 h1:3I4Kt4BQjOR54NavqnDogx/MIoWBFa0StPA8ELUXHmA=
@@ -221,8 +225,8 @@ github.com/muesli/roff v0.1.0 h1:YD0lalCotmYuF5HhZliKWlIx7IEhiXeSfq7hNjFqGF8=
 github.com/muesli/roff v0.1.0/go.mod h1:pjAHQM9hdUUwm/krAfrLGgJkXJ+YuhtsfZ42kieB2Ig=
 github.com/muesli/termenv v0.16.0 h1:S5AlUN9dENB57rsbnkPyfdGuWIlkmzJjbFf0Tf5FWUc=
 github.com/muesli/termenv v0.16.0/go.mod h1:ZRfOIKPFDYQoDFF4Olj7/QJbW60Ol/kL1pU3VfY/Cnk=
-github.com/pelletier/go-toml/v2 v2.3.0 h1:k59bC/lIZREW0/iVaQR8nDHxVq8OVlIzYCOJf421CaM=
-github.com/pelletier/go-toml/v2 v2.3.0/go.mod h1:2gIqNv+qfxSVS7cM2xJQKtLSTLUE9V8t9Stt+h56mCY=
+github.com/pelletier/go-toml/v2 v2.3.1 h1:MYEvvGnQjeNkRF1qUuGolNtNExTDwct51yp7olPtrEc=
+github.com/pelletier/go-toml/v2 v2.3.1/go.mod h1:2gIqNv+qfxSVS7cM2xJQKtLSTLUE9V8t9Stt+h56mCY=
 github.com/pkg/browser v0.0.0-20240102092130-5ac0b6a4141c h1:+mdjkGKdHQG3305AYmdv1U2eRNDiU2ErMBj1gwrq8eQ=
 github.com/pkg/browser v0.0.0-20240102092130-5ac0b6a4141c/go.mod h1:7rwL4CYBLnjLxUqIJNnCWiEdr3bn6IUYi15bNlnbCCU=
 github.com/planetscale/vtprotobuf v0.6.1-0.20240319094008-0393e58bdf10 h1:GFCKgmp0tecUJ0sJuv4pzYCqS9+RGSn52M3FUwPs+uo=
@@ -236,6 +240,8 @@ github.com/rogpeppe/go-internal v1.14.1/go.mod h1:MaRKkUm5W0goXpeCfT7UZI6fk/L7L7
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
 github.com/sagikazarmark/locafero v0.12.0 h1:/NQhBAkUb4+fH1jivKHWusDYFjMOOKU88eegjfxfHb4=
 github.com/sagikazarmark/locafero v0.12.0/go.mod h1:sZh36u/YSZ918v0Io+U9ogLYQJ9tLLBmM4eneO6WwsI=
+github.com/santhosh-tekuri/jsonschema/v6 v6.0.2 h1:KRzFb2m7YtdldCEkzs6KqmJw4nqEVZGK7IN2kJkjTuQ=
+github.com/santhosh-tekuri/jsonschema/v6 v6.0.2/go.mod h1:JXeL+ps8p7/KNMjDQk3TCwPpBy0wYklyWTfbkIzdIFU=
 github.com/spf13/afero v1.15.0 h1:b/YBCLWAJdFWJTN9cLhiXXcD7mzKn9Dm86dNnfyQw1I=
 github.com/spf13/afero v1.15.0/go.mod h1:NC2ByUVxtQs4b3sIUphxK0NioZnmxgyCrfzeuq8lxMg=
 github.com/spf13/cast v1.10.0 h1:h2x0u2shc1QuLHfxi+cTJvs30+ZAHOGRic8uyGTDWxY=
@@ -252,8 +258,8 @@ github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD
 github.com/subosito/gotenv v1.6.0 h1:9NlTDc1FTs4qu0DDq7AEtTPNw6SVm7uBMsUCUjABIf8=
 github.com/subosito/gotenv v1.6.0/go.mod h1:Dk4QP5c2W3ibzajGcXpNraDfq2IrhjMIvMSWPKKo0FU=
 github.com/tidwall/gjson v1.14.2/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
-github.com/tidwall/gjson v1.18.0 h1:FIDeeyB800efLX89e5a8Y0BNH+LOngJyGrIWxG2FKQY=
-github.com/tidwall/gjson v1.18.0/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/gjson v1.19.0 h1:xwxm7n691Uf3u5OFjzngavjGTh55KX5q/9w9xHW88JU=
+github.com/tidwall/gjson v1.19.0/go.mod h1:V37/opeE/JbLUOfH0QTXiNez2l0RUjYUhpT4szFQAfc=
 github.com/tidwall/match v1.1.1/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
 github.com/tidwall/match v1.2.0 h1:0pt8FlkOwjN2fPt4bIl4BoNxb98gGHN2ObFEDkrfZnM=
 github.com/tidwall/match v1.2.0/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
@@ -272,55 +278,54 @@ github.com/yuin/goldmark v1.8.2 h1:kEGpgqJXdgbkhcOgBxkC0X0PmoPG1ZyoZ117rDVp4zE=
 github.com/yuin/goldmark v1.8.2/go.mod h1:ip/1k0VRfGynBgxOz0yCqHrbZXhcjxyuS66Brc7iBKg=
 go.opentelemetry.io/auto/sdk v1.2.1 h1:jXsnJ4Lmnqd11kwkBV2LgLoFMZKizbCi5fNZ/ipaZ64=
 go.opentelemetry.io/auto/sdk v1.2.1/go.mod h1:KRTj+aOaElaLi+wW1kO/DZRXwkF4C5xPbEe3ZiIhN7Y=
-go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.67.0 h1:yI1/OhfEPy7J9eoa6Sj051C7n5dvpj0QX8g4sRchg04=
-go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.67.0/go.mod h1:NoUCKYWK+3ecatC4HjkRktREheMeEtrXoQxrqYFeHSc=
-go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.67.0 h1:OyrsyzuttWTSur2qN/Lm0m2a8yqyIjUVBZcxFPuXq2o=
-go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.67.0/go.mod h1:C2NGBr+kAB4bk3xtMXfZ94gqFDtg/GkI7e9zqGh5Beg=
-go.opentelemetry.io/otel v1.42.0 h1:lSQGzTgVR3+sgJDAU/7/ZMjN9Z+vUip7leaqBKy4sho=
-go.opentelemetry.io/otel v1.42.0/go.mod h1:lJNsdRMxCUIWuMlVJWzecSMuNjE7dOYyWlqOXWkdqCc=
-go.opentelemetry.io/otel/metric v1.42.0 h1:2jXG+3oZLNXEPfNmnpxKDeZsFI5o4J+nz6xUlaFdF/4=
-go.opentelemetry.io/otel/metric v1.42.0/go.mod h1:RlUN/7vTU7Ao/diDkEpQpnz3/92J9ko05BIwxYa2SSI=
-go.opentelemetry.io/otel/sdk v1.42.0 h1:LyC8+jqk6UJwdrI/8VydAq/hvkFKNHZVIWuslJXYsDo=
-go.opentelemetry.io/otel/sdk v1.42.0/go.mod h1:rGHCAxd9DAph0joO4W6OPwxjNTYWghRWmkHuGbayMts=
-go.opentelemetry.io/otel/sdk/metric v1.42.0 h1:D/1QR46Clz6ajyZ3G8SgNlTJKBdGp84q9RKCAZ3YGuA=
-go.opentelemetry.io/otel/sdk/metric v1.42.0/go.mod h1:Ua6AAlDKdZ7tdvaQKfSmnFTdHx37+J4ba8MwVCYM5hc=
-go.opentelemetry.io/otel/trace v1.42.0 h1:OUCgIPt+mzOnaUTpOQcBiM/PLQ/Op7oq6g4LenLmOYY=
-go.opentelemetry.io/otel/trace v1.42.0/go.mod h1:f3K9S+IFqnumBkKhRJMeaZeNk9epyhnCmQh/EysQCdc=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.69.0 h1:2yEATaop1/a1I4psnSLgWVPLWwCzkqWakgJy7xTDVy0=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.69.0/go.mod h1:D7J12YRapIekYyPWgGPlA/23pRmpSEZC5xJC/TTLI9U=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.69.0 h1:8tvICD4vSTOOsNrsI4Ljf6C+6UKvpTEH5XY3JMoyPoo=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.69.0/go.mod h1:z9+yiacE0IHRqM4qFfkbt/JYlmYXgss8GY/jXoNuPJI=
+go.opentelemetry.io/otel v1.44.0 h1:JjwHmHpA4iZ3wBxluu2fbbE7j4kqlE8jXyAyPXH7HqU=
+go.opentelemetry.io/otel v1.44.0/go.mod h1:BMgjTHL9WPRlRjL2oZCBTL4whCGtXch2H4BhOPIAyYc=
+go.opentelemetry.io/otel/metric v1.44.0 h1:1w0gILTcHdr3YI+ixLyjemwrVnsMURbTZFrSYCdDdmc=
+go.opentelemetry.io/otel/metric v1.44.0/go.mod h1:8O7hanEPBNgEMmybD3s2VBKcgWOCsA6tzHBPODAiquo=
+go.opentelemetry.io/otel/sdk v1.44.0 h1:nHYwb9lK+fJPU/dnT6s7W7Z8itMWyqrnVfbheVYrZ58=
+go.opentelemetry.io/otel/sdk v1.44.0/go.mod h1:Osuydd3Se74nqjAKxid74N5eC+jfEqfTegHRnq58oK0=
+go.opentelemetry.io/otel/sdk/metric v1.44.0 h1:3LlKgI+VjbVsjNRFZJZAJ30WjXC5VkNRks6si09iEfI=
+go.opentelemetry.io/otel/sdk/metric v1.44.0/go.mod h1:5B5pMARnXxKhltooO4xUuCBorl65a4EpnTalObqOigA=
+go.opentelemetry.io/otel/trace v1.44.0 h1:jxF5CsGYCe74MCRx2X4g7WsY/VBKRqqpNvXlX/6gtIk=
+go.opentelemetry.io/otel/trace v1.44.0/go.mod h1:oLl1jrMQAVo6v3GAggN+1VH9VIz9iUSvW53sW1Q8PIE=
 go.yaml.in/yaml/v3 v3.0.4 h1:tfq32ie2Jv2UxXFdLJdh3jXuOzWiL1fo0bu/FbuKpbc=
 go.yaml.in/yaml/v3 v3.0.4/go.mod h1:DhzuOOF2ATzADvBadXxruRBLzYTpT36CKvDb3+aBEFg=
-golang.org/x/crypto v0.49.0 h1:+Ng2ULVvLHnJ/ZFEq4KdcDd/cfjrrjjNSXNzxg0Y4U4=
-golang.org/x/crypto v0.49.0/go.mod h1:ErX4dUh2UM+CFYiXZRTcMpEcN8b/1gxEuv3nODoYtCA=
-golang.org/x/exp v0.0.0-20260312153236-7ab1446f8b90 h1:jiDhWWeC7jfWqR9c/uplMOqJ0sbNlNWv0UkzE0vX1MA=
-golang.org/x/exp v0.0.0-20260312153236-7ab1446f8b90/go.mod h1:xE1HEv6b+1SCZ5/uscMRjUBKtIxworgEcEi+/n9NQDQ=
-golang.org/x/net v0.52.0 h1:He/TN1l0e4mmR3QqHMT2Xab3Aj3L9qjbhRm78/6jrW0=
-golang.org/x/net v0.52.0/go.mod h1:R1MAz7uMZxVMualyPXb+VaqGSa3LIaUqk0eEt3w36Sw=
+golang.org/x/crypto v0.52.0 h1:RMs7fP2rXdep0CftQlK8Uf+kibLm7qkCcradZWYz988=
+golang.org/x/crypto v0.52.0/go.mod h1:1QgfPxDqh0T2M/elOJtp9RvuR95kVjir0e6/BvEmGbc=
+golang.org/x/exp v0.0.0-20260528193900-50dc527dd6c7 h1:cHpkPjp4TILjdZxz/O4ykwCpeS+dDqNuDGse4zgQDCk=
+golang.org/x/exp v0.0.0-20260528193900-50dc527dd6c7/go.mod h1:d2fgXJLVs4dYDHUk5lwMIfzRzSrWCfGZb0ZqeLa/Vcw=
+golang.org/x/net v0.55.0 h1:bcvxaJn3e1U6InsFWt1JUq1aSjnRxLzT2rtD2KfkDF8=
+golang.org/x/net v0.55.0/go.mod h1:L5U2KuzuOe1lY7Z+aWVIKK6qEeJXnXV9yzGA+WCHJww=
 golang.org/x/oauth2 v0.36.0 h1:peZ/1z27fi9hUOFCAZaHyrpWG5lwe0RJEEEeH0ThlIs=
 golang.org/x/oauth2 v0.36.0/go.mod h1:YDBUJMTkDnJS+A4BP4eZBjCqtokkg1hODuPjwiGPO7Q=
 golang.org/x/sync v0.20.0 h1:e0PTpb7pjO8GAtTs2dQ6jYa5BWYlMuX047Dco/pItO4=
 golang.org/x/sync v0.20.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
-golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.42.0 h1:omrd2nAlyT5ESRdCLYdm3+fMfNFE/+Rf4bDIQImRJeo=
-golang.org/x/sys v0.42.0/go.mod h1:4GL1E5IUh+htKOUEOaiffhrAeqysfVGipDYzABqnCmw=
-golang.org/x/term v0.41.0 h1:QCgPso/Q3RTJx2Th4bDLqML4W6iJiaXFq2/ftQF13YU=
-golang.org/x/term v0.41.0/go.mod h1:3pfBgksrReYfZ5lvYM0kSO0LIkAl4Yl2bXOkKP7Ec2A=
-golang.org/x/text v0.35.0 h1:JOVx6vVDFokkpaq1AEptVzLTpDe9KGpj5tR4/X+ybL8=
-golang.org/x/text v0.35.0/go.mod h1:khi/HExzZJ2pGnjenulevKNX1W67CUy0AsXcNubPGCA=
+golang.org/x/sys v0.45.0 h1:dO4czNzziLiiXplLQgBCEpCvXQ3dnkn0SdaZSYdQ+FY=
+golang.org/x/sys v0.45.0/go.mod h1:4GL1E5IUh+htKOUEOaiffhrAeqysfVGipDYzABqnCmw=
+golang.org/x/term v0.43.0 h1:S4RLU2sB31O/NCl+zFN9Aru9A/Cq2aqKpTZJ6B+DwT4=
+golang.org/x/term v0.43.0/go.mod h1:lrhlHNdQJHO+1qVYiHfFKVuVioJIheAc3fBSMFYEIsk=
+golang.org/x/text v0.37.0 h1:Cqjiwd9eSg8e0QAkyCaQTNHFIIzWtidPahFWR83rTrc=
+golang.org/x/text v0.37.0/go.mod h1:a5sjxXGs9hsn/AJVwuElvCAo9v8QYLzvavO5z2PiM38=
 golang.org/x/time v0.15.0 h1:bbrp8t3bGUeFOx08pvsMYRTCVSMk89u4tKbNOZbp88U=
 golang.org/x/time v0.15.0/go.mod h1:Y4YMaQmXwGQZoFaVFk4YpCt4FLQMYKZe9oeV/f4MSno=
-gonum.org/v1/gonum v0.16.0 h1:5+ul4Swaf3ESvrOnidPp4GZbzf0mxVQpDCYUQE7OJfk=
-gonum.org/v1/gonum v0.16.0/go.mod h1:fef3am4MQ93R2HHpKnLk4/Tbh/s0+wqD5nfa6Pnwy4E=
-google.golang.org/api v0.273.0 h1:r/Bcv36Xa/te1ugaN1kdJ5LoA5Wj/cL+a4gj6FiPBjQ=
-google.golang.org/api v0.273.0/go.mod h1:JbAt7mF+XVmWu6xNP8/+CTiGH30ofmCmk9nM8d8fHew=
-google.golang.org/genai v1.52.0 h1:ekVIxWHtLUNbt+v0WWi4j3JT4yrHDEbysMcHQcaCQoI=
-google.golang.org/genai v1.52.0/go.mod h1:A3kkl0nyBjyFlNjgxIwKq70julKbIxpSxqKO5gw/gmk=
-google.golang.org/genproto v0.0.0-20260319201613-d00831a3d3e7 h1:XzmzkmB14QhVhgnawEVsOn6OFsnpyxNPRY9QV01dNB0=
-google.golang.org/genproto v0.0.0-20260319201613-d00831a3d3e7/go.mod h1:L43LFes82YgSonw6iTXTxXUX1OlULt4AQtkik4ULL/I=
-google.golang.org/genproto/googleapis/api v0.0.0-20260319201613-d00831a3d3e7 h1:41r6JMbpzBMen0R/4TZeeAmGXSJC7DftGINUodzTkPI=
-google.golang.org/genproto/googleapis/api v0.0.0-20260319201613-d00831a3d3e7/go.mod h1:EIQZ5bFCfRQDV4MhRle7+OgjNtZ6P1PiZBgAKuxXu/Y=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20260319201613-d00831a3d3e7 h1:ndE4FoJqsIceKP2oYSnUZqhTdYufCYYkqwtFzfrhI7w=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20260319201613-d00831a3d3e7/go.mod h1:4Hqkh8ycfw05ld/3BWL7rJOSfebL2Q+DVDeRgYgxUU8=
-google.golang.org/grpc v1.79.3 h1:sybAEdRIEtvcD68Gx7dmnwjZKlyfuc61Dyo9pGXXkKE=
-google.golang.org/grpc v1.79.3/go.mod h1:KmT0Kjez+0dde/v2j9vzwoAScgEPx/Bw1CYChhHLrHQ=
+gonum.org/v1/gonum v0.17.0 h1:VbpOemQlsSMrYmn7T2OUvQ4dqxQXU+ouZFQsZOx50z4=
+gonum.org/v1/gonum v0.17.0/go.mod h1:El3tOrEuMpv2UdMrbNlKEh9vd86bmQ6vqIcDwxEOc1E=
+google.golang.org/api v0.282.0 h1:WmJiSVqUnKqJCpJOx7YADbXaC+9DDsnGSfllFSj7R2I=
+google.golang.org/api v0.282.0/go.mod h1:6Wssta4c5n9qHq5CBhmlai5h/PUa1djdDAIhYEHyvcM=
+google.golang.org/genai v1.58.0 h1:MNA3ZkRyr7MnRwZ9RNZ60p4+UMKV3yYRw6pyHq4pp0U=
+google.golang.org/genai v1.58.0/go.mod h1:A3kkl0nyBjyFlNjgxIwKq70julKbIxpSxqKO5gw/gmk=
+google.golang.org/genproto v0.0.0-20260504160031-60b97b32f348 h1:JjVGDZYWkJWZcxveJGzfkXC5myDVWAd4dZdgbzrDUv8=
+google.golang.org/genproto v0.0.0-20260504160031-60b97b32f348/go.mod h1:95PqD4xM+AdOcBGsmgfaofXsiA37uXDtDufVbntT3TU=
+google.golang.org/genproto/googleapis/api v0.0.0-20260504160031-60b97b32f348 h1:U8orV30l6KpDsi9dxU0CoJZGbjS8EEpw+6ba+XwGPQA=
+google.golang.org/genproto/googleapis/api v0.0.0-20260504160031-60b97b32f348/go.mod h1:Yzdzr5OOZFgSsEV2D/Xi9NL3bszpXFAg0hFJiRohcD8=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20260526163538-3dc84a4a5aaa h1:mZHHdPZl0dbGHCflZgAq/Q468DWVFcU2whhB2KAo8fk=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20260526163538-3dc84a4a5aaa/go.mod h1:4Hqkh8ycfw05ld/3BWL7rJOSfebL2Q+DVDeRgYgxUU8=
+google.golang.org/grpc v1.81.1 h1:VnnIIZ88UzOOKLukQi+ImGz8O1Wdp8nAGGnvOfEIWQQ=
+google.golang.org/grpc v1.81.1/go.mod h1:xGH9GfzOyMTGIOXBJmXt+BX/V0kcdQbdcuwQ/zNw42I=
 google.golang.org/protobuf v1.36.11 h1:fV6ZwhNocDyBLK0dj+fg8ektcVegBBuEolpbTQyBNVE=
 google.golang.org/protobuf v1.36.11/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=

internal/acpserver/agent.go

@@ -7,8 +7,11 @@ package acpserver
 
 import (
 	"context"
+	"encoding/base64"
 	"encoding/json"
 	"fmt"
+	"os"
+	"strings"
 	"sync/atomic"
 
 	"github.com/charmbracelet/log"
@@ -20,7 +23,6 @@ import (
 // Version is injected at build time; fallback to "dev".
 var Version = "dev"
 
-// Agent implements the acp.Agent interface, delegating to Kit for LLM
 // execution, tool calls, and session management.
 type Agent struct {
 	conn     *acp.AgentSideConnection
@@ -111,13 +113,20 @@ func (a *Agent) Prompt(ctx context.Context, params acp.PromptRequest) (acp.Promp
 		)
 	}
 
-	// Extract text from prompt content blocks.
-	promptText := extractPromptText(params.Prompt)
-	if promptText == "" {
+	// Extract text and file attachments from prompt content blocks.
+	promptText, files := extractPromptContent(params.Prompt)
+	if promptText == "" && len(files) == 0 {
 		return acp.PromptResponse{}, acp.NewInvalidParams("empty prompt")
 	}
 
-	log.Debug("acp: prompt", "session", sessionID, "prompt_len", len(promptText))
+	// If we have files but no text prompt, add a default prompt
+	// This is required because the underlying LLM library needs a non-empty prompt
+	// when there are no previous messages in the conversation.
+	if promptText == "" && len(files) > 0 {
+		promptText = "Please analyze the attached file."
+	}
+
+	log.Debug("acp: prompt", "session", sessionID, "prompt_len", len(promptText), "files", len(files))
 
 	// Create a cancellable context for this prompt turn.
 	promptCtx, cancel := context.WithCancel(ctx)
@@ -129,7 +138,13 @@ func (a *Agent) Prompt(ctx context.Context, params acp.PromptRequest) (acp.Promp
 	defer unsub()
 
 	// Run the prompt through Kit's full turn lifecycle.
-	_, err := sess.kit.PromptResult(promptCtx, promptText)
+	// Use PromptResultWithFiles when file attachments are present.
+	var err error
+	if len(files) > 0 {
+		_, err = sess.kit.PromptResultWithFiles(promptCtx, promptText, files)
+	} else {
+		_, err = sess.kit.PromptResult(promptCtx, promptText)
+	}
 	if err != nil {
 		if promptCtx.Err() != nil {
 			return acp.PromptResponse{
@@ -162,6 +177,77 @@ func (a *Agent) SetSessionMode(_ context.Context, _ acp.SetSessionModeRequest) (
 	return acp.SetSessionModeResponse{}, nil
 }
 
+// ListSessions returns an empty session list. Kit doesn't persist sessions
+// across restarts in ACP mode, so this is effectively a no-op.
+func (a *Agent) ListSessions(_ context.Context, _ acp.ListSessionsRequest) (acp.ListSessionsResponse, error) {
+	return acp.ListSessionsResponse{
+		Sessions: []acp.SessionInfo{},
+	}, nil
+}
+
+// CloseSession cancels any ongoing work for the session and frees its resources.
+func (a *Agent) CloseSession(_ context.Context, params acp.CloseSessionRequest) (acp.CloseSessionResponse, error) {
+	sessionID := string(params.SessionId)
+	sess, ok := a.registry.get(sessionID)
+	if !ok {
+		return acp.CloseSessionResponse{}, nil
+	}
+
+	log.Debug("acp: close session", "session", sessionID)
+	sess.cancelPrompt()
+	a.registry.remove(sessionID)
+	return acp.CloseSessionResponse{}, nil
+}
+
+// ResumeSession is not supported — Kit doesn't persist sessions across
+// restarts in ACP mode. Clients should use NewSession instead.
+func (a *Agent) ResumeSession(_ context.Context, _ acp.ResumeSessionRequest) (acp.ResumeSessionResponse, error) {
+	return acp.ResumeSessionResponse{}, fmt.Errorf("resume session not supported")
+}
+
+// SetSessionConfigOption handles session configuration changes. Currently
+// supports the "model" config option to change the active model for a session.
+func (a *Agent) SetSessionConfigOption(ctx context.Context, params acp.SetSessionConfigOptionRequest) (acp.SetSessionConfigOptionResponse, error) {
+	// Extract session ID and config ID from whichever variant is present.
+	var sessionID string
+	var configID string
+	var value string
+
+	switch {
+	case params.ValueId != nil:
+		sessionID = string(params.ValueId.SessionId)
+		configID = string(params.ValueId.ConfigId)
+		value = string(params.ValueId.Value)
+	case params.Boolean != nil:
+		sessionID = string(params.Boolean.SessionId)
+		configID = string(params.Boolean.ConfigId)
+		// Boolean config options are not used for model selection.
+		log.Debug("acp: set_session_config_option (boolean)", "session", sessionID, "config", configID, "value", params.Boolean.Value)
+		return acp.SetSessionConfigOptionResponse{}, nil
+	default:
+		return acp.SetSessionConfigOptionResponse{}, acp.NewInvalidParams("unsupported config option variant")
+	}
+
+	sess, ok := a.registry.get(sessionID)
+	if !ok {
+		return acp.SetSessionConfigOptionResponse{}, acp.NewInvalidParams(fmt.Sprintf("session not found: %s", sessionID))
+	}
+
+	log.Debug("acp: set_session_config_option", "session", sessionID, "config", configID, "value", value)
+
+	// Handle known config options.
+	switch configID {
+	case "model":
+		if err := sess.kit.SetModel(ctx, value); err != nil {
+			return acp.SetSessionConfigOptionResponse{}, fmt.Errorf("set model: %w", err)
+		}
+	default:
+		log.Debug("acp: unknown config option", "config", configID)
+	}
+
+	return acp.SetSessionConfigOptionResponse{}, nil
+}
+
 // ---------------------------------------------------------------------------
 // Event streaming: Kit events → ACP SessionUpdate notifications
 // ---------------------------------------------------------------------------
@@ -231,19 +317,196 @@ func (a *Agent) subscribeEvents(ctx context.Context, k *kit.Kit, sessionID acp.S
 // Helpers
 // ---------------------------------------------------------------------------
 
-// extractPromptText extracts the concatenated text content from ACP content
-// blocks. Non-text blocks are ignored for now.
-func extractPromptText(blocks []acp.ContentBlock) string {
-	var text string
-	for _, block := range blocks {
-		if block.Text != nil {
-			if text != "" {
-				text += "\n"
+// extractPromptContent extracts text and file attachments from ACP content blocks.
+// It converts supported content blocks (image, audio, resource) to Kit's LLMFilePart.
+func extractPromptContent(blocks []acp.ContentBlock) (string, []kit.LLMFilePart) {
+	var textParts []string
+	var files []kit.LLMFilePart
+
+	log.Debug("acp: extracting content", "blocks", len(blocks))
+
+	for i, block := range blocks {
+		switch {
+		// Text content
+		case block.Text != nil:
+			log.Debug("acp: content block", "index", i, "type", "text", "len", len(block.Text.Text))
+			textParts = append(textParts, block.Text.Text)
+
+		// Image data (base64)
+		case block.Image != nil:
+			mimeType := block.Image.MimeType
+			if mimeType == "" {
+				mimeType = "image/png" // Default fallback
 			}
-			text += block.Text.Text
+			log.Debug("acp: content block", "index", i, "type", "image", "mime", mimeType, "data_len", len(block.Image.Data))
+			if data, err := base64.StdEncoding.DecodeString(block.Image.Data); err == nil {
+				files = append(files, kit.LLMFilePart{
+					Filename:  "image.png",
+					Data:      data,
+					MediaType: mimeType,
+				})
+			} else {
+				log.Debug("acp: failed to decode image", "error", err)
+			}
+
+		// Audio data (base64)
+		case block.Audio != nil:
+			mimeType := block.Audio.MimeType
+			if mimeType == "" {
+				mimeType = "audio/wav" // Default fallback
+			}
+			log.Debug("acp: content block", "index", i, "type", "audio", "mime", mimeType)
+			if data, err := base64.StdEncoding.DecodeString(block.Audio.Data); err == nil {
+				files = append(files, kit.LLMFilePart{
+					Filename:  "audio.wav",
+					Data:      data,
+					MediaType: mimeType,
+				})
+			} else {
+				log.Debug("acp: failed to decode audio", "error", err)
+			}
+
+		// Embedded resource (text or binary file content)
+		case block.Resource != nil:
+			log.Debug("acp: content block", "index", i, "type", "resource")
+			res := block.Resource.Resource
+			// Text resource - append as text content with file reference
+			if res.TextResourceContents != nil {
+				uri := res.TextResourceContents.Uri
+				content := res.TextResourceContents.Text
+				mimeType := "text/plain"
+				if res.TextResourceContents.MimeType != nil {
+					mimeType = *res.TextResourceContents.MimeType
+				}
+				log.Debug("acp: text resource", "uri", uri, "mime", mimeType, "len", len(content))
+				// Text files are included as formatted text, NOT as FilePart
+				// FilePart is for binary files (images, audio, PDFs) only
+				textParts = append(textParts, fmt.Sprintf("[File: %s]\n```\n%s\n```", uri, content))
+			}
+			// Binary resource (base64 blob) - these become FilePart
+			if res.BlobResourceContents != nil {
+				uri := res.BlobResourceContents.Uri
+				mimeType := "application/octet-stream"
+				if res.BlobResourceContents.MimeType != nil {
+					mimeType = *res.BlobResourceContents.MimeType
+				}
+				log.Debug("acp: binary resource", "uri", uri, "mime", mimeType, "blob_len", len(res.BlobResourceContents.Blob))
+				if data, err := base64.StdEncoding.DecodeString(res.BlobResourceContents.Blob); err == nil {
+					files = append(files, kit.LLMFilePart{
+						Filename:  extractFilenameFromURI(uri),
+						Data:      data,
+						MediaType: mimeType,
+					})
+				} else {
+					log.Debug("acp: failed to decode binary resource", "error", err)
+				}
+			}
+
+		// Resource link (file reference without embedded content)
+		case block.ResourceLink != nil:
+			uri := block.ResourceLink.Uri
+			name := block.ResourceLink.Name
+			log.Debug("acp: content block", "index", i, "type", "resource_link", "uri", uri, "name", name)
+			// For resource links, we'll try to read the file from disk
+			// This requires the file URI to be accessible (file:// scheme)
+			if content, err := readResourceFromURI(uri); err == nil {
+				// Detect if it's a text file or binary file
+				mimeType := "text/plain"
+				if block.ResourceLink.MimeType != nil {
+					mimeType = *block.ResourceLink.MimeType
+				}
+				log.Debug("acp: resource link loaded", "uri", uri, "mime", mimeType, "size", len(content))
+
+				// Only create FilePart for binary files (images, audio, PDFs, etc.)
+				// Text files are included as formatted text in the message
+				if isTextMimeType(mimeType) || looksLikeText(content) {
+					textParts = append(textParts, fmt.Sprintf("[File: %s]\n```\n%s\n```", uri, string(content)))
+				} else {
+					// Binary file - create FilePart for models that support it
+					files = append(files, kit.LLMFilePart{
+						Filename:  extractFilenameFromURI(uri),
+						Data:      content,
+						MediaType: mimeType,
+					})
+				}
+			} else {
+				// If we can't read it, include as a text reference
+				log.Debug("acp: resource link failed to load", "uri", uri, "error", err)
+				textParts = append(textParts, fmt.Sprintf("[Referenced file: %s]", uri))
+			}
+
+		default:
+			log.Debug("acp: content block", "index", i, "type", "unknown/unhandled")
 		}
 	}
-	return text
+
+	// Debug log the extracted content
+	for i, f := range files {
+		log.Debug("acp: extracted file", "index", i, "filename", f.Filename, "mime", f.MediaType, "size", len(f.Data))
+	}
+
+	return strings.Join(textParts, "\n"), files
+}
+
+// isTextMimeType returns true if the MIME type indicates text content.
+func isTextMimeType(mimeType string) bool {
+	return strings.HasPrefix(mimeType, "text/") ||
+		mimeType == "application/json" ||
+		mimeType == "application/xml" ||
+		mimeType == "application/javascript" ||
+		mimeType == "application/typescript" ||
+		mimeType == "application/x-sh" ||
+		mimeType == "application/x-python" ||
+		mimeType == "application/x-yaml" ||
+		mimeType == "application/x-toml"
+}
+
+// looksLikeText checks if the content appears to be text (not binary).
+// It samples the first 512 bytes and checks for null bytes or high
+// concentration of non-printable characters.
+func looksLikeText(data []byte) bool {
+	if len(data) == 0 {
+		return true
+	}
+	// Check first 512 bytes (or less if file is smaller)
+	sampleSize := min(len(data), 512)
+	sample := data[:sampleSize]
+
+	// Count non-printable characters
+	nonPrintable := 0
+	for _, b := range sample {
+		// Null byte indicates binary
+		if b == 0 {
+			return false
+		}
+		// Count control characters (except common whitespace)
+		if b < 32 && b != '\n' && b != '\r' && b != '\t' {
+			nonPrintable++
+		}
+	}
+
+	// If more than 30% non-printable, consider it binary
+	return float64(nonPrintable)/float64(sampleSize) < 0.3
+}
+
+// extractFilenameFromURI extracts a filename from a file URI or path.
+func extractFilenameFromURI(uri string) string {
+	// Handle file:// URIs
+	uri = strings.TrimPrefix(uri, "file://")
+	// Extract basename
+	if idx := strings.LastIndex(uri, "/"); idx >= 0 {
+		return uri[idx+1:]
+	}
+	return uri
+}
+
+// readResourceFromURI attempts to read file content from a file:// URI.
+func readResourceFromURI(uri string) ([]byte, error) {
+	if !strings.HasPrefix(uri, "file://") {
+		return nil, fmt.Errorf("unsupported URI scheme: %s", uri)
+	}
+	path := uri[7:] // Remove file:// prefix
+	return os.ReadFile(path)
 }
 
 // parseToolArgs attempts to parse a JSON tool args string into a map for

internal/acpserver/session.go

@@ -7,7 +7,9 @@ import (
 	"sync"
 
 	"github.com/charmbracelet/log"
+	"github.com/spf13/viper"
 
+	"github.com/mark3labs/kit/internal/extbridge"
 	"github.com/mark3labs/kit/internal/extensions"
 	kit "github.com/mark3labs/kit/pkg/kit"
 )
@@ -37,10 +39,21 @@ func newSessionRegistry() *sessionRegistry {
 // given working directory. The Kit-generated session ID is used as the ACP
 // session ID so the mapping is 1:1.
 func (r *sessionRegistry) create(ctx context.Context, cwd string) (*acpSession, error) {
+	// Each ACP session gets its own isolated config store (CLI is left nil) so
+	// per-session SetModel / SetThinkingLevel calls cannot race or bleed across
+	// the sessionRegistry. We seed the relevant root-command flag values from
+	// the process-global store (which cobra populated from flags) so launching
+	// `kit acp -m <model> [--thinking-level ...] [--provider-url ...]` is still
+	// honored; .kit.yml and KIT_* env vars are loaded per session by kit.New.
+	streamOn := true
 	kitInstance, err := kit.New(ctx, &kit.Options{
-		SessionDir: cwd,
-		Quiet:      true,
-		Streaming:  true,
+		SessionDir:     cwd,
+		Quiet:          true,
+		Streaming:      &streamOn,
+		Model:          viper.GetString("model"),
+		ThinkingLevel:  viper.GetString("thinking-level"),
+		ProviderURL:    viper.GetString("provider-url"),
+		ProviderAPIKey: viper.GetString("provider-api-key"),
 	})
 	if err != nil {
 		// Provide actionable guidance for provider auth errors, which are
@@ -152,38 +165,7 @@ func (r *sessionRegistry) create(ctx context.Context, cwd string) (*acpSession,
 				return kitInstance.ExecuteCompletion(context.Background(), req)
 			},
 			SpawnSubagent: func(config extensions.SubagentConfig) (*extensions.SubagentHandle, *extensions.SubagentResult, error) {
-				sdkCfg := kit.SubagentConfig{
-					Prompt:       config.Prompt,
-					Model:        config.Model,
-					SystemPrompt: config.SystemPrompt,
-					Timeout:      config.Timeout,
-					NoSession:    config.NoSession,
-				}
-				if config.OnEvent != nil {
-					sdkCfg.OnEvent = func(e kit.Event) {
-						se := sdkEventToSubagentEvent(e)
-						if se.Type != "" {
-							config.OnEvent(se)
-						}
-					}
-				}
-				result, err := kitInstance.Subagent(context.Background(), sdkCfg)
-				if result == nil {
-					return nil, &extensions.SubagentResult{Error: err}, err
-				}
-				extResult := &extensions.SubagentResult{
-					Response:  result.Response,
-					Error:     err,
-					SessionID: result.SessionID,
-					Elapsed:   result.Elapsed,
-				}
-				if result.Usage != nil {
-					extResult.Usage = &extensions.SubagentUsage{
-						InputTokens:  result.Usage.InputTokens,
-						OutputTokens: result.Usage.OutputTokens,
-					}
-				}
-				return nil, extResult, err
+				return extbridge.SpawnSubagent(context.Background(), kitInstance, config)
 			},
 
 			// Render — fall back to logging.
@@ -232,6 +214,20 @@ func (r *sessionRegistry) closeAll() {
 	}
 }
 
+// remove closes and removes a single session by ID.
+func (r *sessionRegistry) remove(sessionID string) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	sess, ok := r.sessions[sessionID]
+	if !ok {
+		return
+	}
+	if sess.kit != nil {
+		_ = sess.kit.Close()
+	}
+	delete(r.sessions, sessionID)
+}
+
 // cancelPrompt cancels the current prompt for a session, if any.
 func (s *acpSession) cancelPrompt() {
 	s.cancelMu.Lock()
@@ -255,40 +251,3 @@ func (s *acpSession) clearCancel() {
 	defer s.cancelMu.Unlock()
 	s.cancelFn = nil
 }
-
-// sdkEventToSubagentEvent converts an SDK event to an extension SubagentEvent.
-func sdkEventToSubagentEvent(e kit.Event) extensions.SubagentEvent {
-	switch ev := e.(type) {
-	case kit.MessageUpdateEvent:
-		return extensions.SubagentEvent{Type: "text", Content: ev.Chunk}
-	case kit.ReasoningDeltaEvent:
-		return extensions.SubagentEvent{Type: "reasoning", Content: ev.Delta}
-	case kit.ToolCallEvent:
-		return extensions.SubagentEvent{
-			Type: "tool_call", ToolCallID: ev.ToolCallID,
-			ToolName: ev.ToolName, ToolKind: ev.ToolKind, ToolArgs: ev.ToolArgs,
-		}
-	case kit.ToolExecutionStartEvent:
-		return extensions.SubagentEvent{
-			Type: "tool_execution_start", ToolCallID: ev.ToolCallID,
-			ToolName: ev.ToolName, ToolKind: ev.ToolKind,
-		}
-	case kit.ToolExecutionEndEvent:
-		return extensions.SubagentEvent{
-			Type: "tool_execution_end", ToolCallID: ev.ToolCallID,
-			ToolName: ev.ToolName, ToolKind: ev.ToolKind,
-		}
-	case kit.ToolResultEvent:
-		return extensions.SubagentEvent{
-			Type: "tool_result", ToolCallID: ev.ToolCallID,
-			ToolName: ev.ToolName, ToolKind: ev.ToolKind,
-			ToolResult: ev.Result, IsError: ev.IsError,
-		}
-	case kit.TurnStartEvent:
-		return extensions.SubagentEvent{Type: "turn_start"}
-	case kit.TurnEndEvent:
-		return extensions.SubagentEvent{Type: "turn_end"}
-	default:
-		return extensions.SubagentEvent{}
-	}
-}

internal/agent/agent_mcp_test.go

@@ -0,0 +1,302 @@
+package agent
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"runtime"
+	"strings"
+	"testing"
+	"time"
+
+	"charm.land/fantasy"
+
+	"github.com/mark3labs/kit/internal/config"
+)
+
+// mockModel is a minimal LanguageModel that satisfies the interface
+// without making real API calls. Used to test tool management wiring.
+type mockModel struct{}
+
+func (m *mockModel) Generate(_ context.Context, _ fantasy.Call) (*fantasy.Response, error) {
+	return &fantasy.Response{}, nil
+}
+func (m *mockModel) Stream(_ context.Context, _ fantasy.Call) (fantasy.StreamResponse, error) {
+	return nil, nil
+}
+func (m *mockModel) GenerateObject(_ context.Context, _ fantasy.ObjectCall) (*fantasy.ObjectResponse, error) {
+	return &fantasy.ObjectResponse{}, nil
+}
+func (m *mockModel) StreamObject(_ context.Context, _ fantasy.ObjectCall) (fantasy.ObjectStreamResponse, error) {
+	return nil, nil
+}
+func (m *mockModel) Provider() string { return "mock" }
+func (m *mockModel) Model() string    { return "mock-model" }
+
+// testdataDir returns the absolute path to the tools testdata directory.
+func testdataDir(t *testing.T) string {
+	t.Helper()
+	_, file, _, ok := runtime.Caller(0)
+	if !ok {
+		t.Fatal("cannot determine test file path")
+	}
+	return filepath.Join(filepath.Dir(file), "..", "tools", "testdata")
+}
+
+// echoServerConfig returns an MCPServerConfig for the test echo MCP server.
+func echoServerConfig(t *testing.T) config.MCPServerConfig {
+	t.Helper()
+	script := filepath.Join(testdataDir(t), "echo_server.py")
+	if _, err := os.Stat(script); err != nil {
+		t.Skipf("echo_server.py not found: %v", err)
+	}
+	return config.MCPServerConfig{
+		Command: []string{"python3", script},
+	}
+}
+
+// mockAuthHandler is a minimal MCPAuthHandler for testing that auth handler
+// propagation works without requiring a real OAuth server.
+type mockAuthHandler struct {
+	redirectURI string
+}
+
+func (h *mockAuthHandler) RedirectURI() string { return h.redirectURI }
+func (h *mockAuthHandler) HandleAuth(_ context.Context, _ string, _ string) (string, error) {
+	return "", nil
+}
+
+// newTestAgent creates a minimal Agent with a mock model and no core tools,
+// suitable for testing MCP server management without an API key.
+func newTestAgent() *Agent {
+	model := &mockModel{}
+	a := &Agent{
+		model:        model,
+		coreTools:    nil,
+		extraTools:   nil,
+		maxSteps:     10,
+		systemPrompt: "test",
+		fantasyAgent: fantasy.NewAgent(model),
+	}
+	return a
+}
+
+func TestAgent_AddMCPServer(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping integration test in short mode")
+	}
+
+	a := newTestAgent()
+	defer func() { _ = a.Close() }()
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	cfg := echoServerConfig(t)
+
+	// Initially no MCP tools.
+	if a.GetMCPToolCount() != 0 {
+		t.Fatalf("Expected 0 MCP tools initially, got %d", a.GetMCPToolCount())
+	}
+
+	// Add a server.
+	count, err := a.AddMCPServer(ctx, "echo", cfg)
+	if err != nil {
+		t.Fatalf("AddMCPServer failed: %v", err)
+	}
+	if count != 2 {
+		t.Errorf("Expected 2 tools, got %d", count)
+	}
+
+	// Verify tools are in the agent's tool list.
+	if a.GetMCPToolCount() != 2 {
+		t.Errorf("Expected 2 MCP tools, got %d", a.GetMCPToolCount())
+	}
+
+	allTools := a.GetTools()
+	toolNames := make(map[string]bool)
+	for _, tool := range allTools {
+		toolNames[tool.Info().Name] = true
+	}
+	if !toolNames["echo__echo"] {
+		t.Error("Expected tool 'echo__echo' in agent tools")
+	}
+	if !toolNames["echo__greet"] {
+		t.Error("Expected tool 'echo__greet' in agent tools")
+	}
+
+	// Verify loaded server names.
+	names := a.GetLoadedServerNames()
+	found := false
+	for _, n := range names {
+		if n == "echo" {
+			found = true
+		}
+	}
+	if !found {
+		t.Errorf("Expected 'echo' in loaded server names: %v", names)
+	}
+}
+
+func TestAgent_RemoveMCPServer(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping integration test in short mode")
+	}
+
+	a := newTestAgent()
+	defer func() { _ = a.Close() }()
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	cfg := echoServerConfig(t)
+
+	// Add then remove.
+	_, err := a.AddMCPServer(ctx, "echo", cfg)
+	if err != nil {
+		t.Fatalf("AddMCPServer failed: %v", err)
+	}
+
+	err = a.RemoveMCPServer("echo")
+	if err != nil {
+		t.Fatalf("RemoveMCPServer failed: %v", err)
+	}
+
+	// Verify tools removed.
+	if a.GetMCPToolCount() != 0 {
+		t.Errorf("Expected 0 MCP tools after removal, got %d", a.GetMCPToolCount())
+	}
+
+	// Verify agent's tool list has no MCP tools.
+	for _, tool := range a.GetTools() {
+		if strings.Contains(tool.Info().Name, "echo__") {
+			t.Errorf("Found leftover tool after removal: %s", tool.Info().Name)
+		}
+	}
+}
+
+func TestAgent_RemoveMCPServer_NoToolManager(t *testing.T) {
+	a := newTestAgent()
+	defer func() { _ = a.Close() }()
+
+	err := a.RemoveMCPServer("nonexistent")
+	if err == nil {
+		t.Fatal("Expected error when no tool manager exists")
+	}
+	if !strings.Contains(err.Error(), "no MCP servers loaded") {
+		t.Errorf("Expected 'no MCP servers loaded' error, got: %v", err)
+	}
+}
+
+func TestAgent_AddMCPServer_CreatesToolManager(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping integration test in short mode")
+	}
+
+	a := newTestAgent()
+	defer func() { _ = a.Close() }()
+
+	// Initially no tool manager.
+	if a.GetMCPToolManager() != nil {
+		t.Fatal("Expected nil tool manager initially")
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	cfg := echoServerConfig(t)
+	_, err := a.AddMCPServer(ctx, "echo", cfg)
+	if err != nil {
+		t.Fatalf("AddMCPServer failed: %v", err)
+	}
+
+	// Tool manager should now exist.
+	if a.GetMCPToolManager() == nil {
+		t.Fatal("Expected tool manager to be created by AddMCPServer")
+	}
+}
+
+func TestAgent_AddRemoveAdd_MCP(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping integration test in short mode")
+	}
+
+	a := newTestAgent()
+	defer func() { _ = a.Close() }()
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	cfg := echoServerConfig(t)
+
+	// Add → Remove → Add cycle.
+	_, err := a.AddMCPServer(ctx, "echo", cfg)
+	if err != nil {
+		t.Fatalf("First add failed: %v", err)
+	}
+
+	err = a.RemoveMCPServer("echo")
+	if err != nil {
+		t.Fatalf("Remove failed: %v", err)
+	}
+
+	count, err := a.AddMCPServer(ctx, "echo", cfg)
+	if err != nil {
+		t.Fatalf("Re-add failed: %v", err)
+	}
+	if count != 2 {
+		t.Errorf("Expected 2 tools on re-add, got %d", count)
+	}
+	if a.GetMCPToolCount() != 2 {
+		t.Errorf("Expected 2 MCP tools after re-add, got %d", a.GetMCPToolCount())
+	}
+}
+
+// TestAgent_AddMCPServer_InheritsAuthHandler verifies that AddMCPServer()
+// propagates the agent's authHandler and tokenStoreFactory to a newly created
+// MCPToolManager (fix for issue #3).
+func TestAgent_AddMCPServer_InheritsAuthHandler(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping integration test in short mode")
+	}
+
+	handler := &mockAuthHandler{redirectURI: "http://localhost:9999/oauth/callback"}
+
+	model := &mockModel{}
+	a := &Agent{
+		model:             model,
+		coreTools:         nil,
+		extraTools:        nil,
+		maxSteps:          10,
+		systemPrompt:      "test",
+		fantasyAgent:      fantasy.NewAgent(model),
+		authHandler:       handler,
+		tokenStoreFactory: nil, // nil is fine; we just test authHandler propagation
+	}
+	defer func() { _ = a.Close() }()
+
+	// Initially no tool manager.
+	if a.GetMCPToolManager() != nil {
+		t.Fatal("Expected nil tool manager initially")
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	cfg := echoServerConfig(t)
+	_, err := a.AddMCPServer(ctx, "echo", cfg)
+	if err != nil {
+		t.Fatalf("AddMCPServer failed: %v", err)
+	}
+
+	// Tool manager should now exist and have the auth handler set.
+	tm := a.GetMCPToolManager()
+	if tm == nil {
+		t.Fatal("Expected tool manager to be created by AddMCPServer")
+	}
+
+	// Verify the auth handler was propagated by checking the field directly.
+	if tm.GetAuthHandler() == nil {
+		t.Fatal("Expected auth handler to be propagated to tool manager")
+	}
+}

internal/agent/agent_parallel_tool_args_test.go

@@ -0,0 +1,109 @@
+package agent
+
+import (
+	"context"
+	"sync"
+	"testing"
+
+	"charm.land/fantasy"
+)
+
+// fakeParallelAgent simulates a provider that emits two parallel tool_use
+// blocks in a single step. It invokes the streaming callbacks in the order:
+//
+//	OnToolCall(A) -> OnToolCall(B) -> OnToolResult(A) -> OnToolResult(B)
+//
+// Before the fix in #33 the agent-layer wrapper recorded a single
+// `currentToolArgs` variable that was clobbered by the second OnToolCall, so
+// both OnToolResult callbacks received B's args instead of their own.
+type fakeParallelAgent struct {
+	calls   []fantasy.ToolCallContent
+	results []fantasy.ToolResultContent
+}
+
+func (f *fakeParallelAgent) Generate(_ context.Context, _ fantasy.AgentCall) (*fantasy.AgentResult, error) {
+	return &fantasy.AgentResult{}, nil
+}
+
+func (f *fakeParallelAgent) Stream(_ context.Context, opts fantasy.AgentStreamCall) (*fantasy.AgentResult, error) {
+	for _, tc := range f.calls {
+		if opts.OnToolCall != nil {
+			if err := opts.OnToolCall(tc); err != nil {
+				return nil, err
+			}
+		}
+	}
+	for _, tr := range f.results {
+		if opts.OnToolResult != nil {
+			if err := opts.OnToolResult(tr); err != nil {
+				return nil, err
+			}
+		}
+	}
+	return &fantasy.AgentResult{}, nil
+}
+
+// TestGenerateWithCallbacks_ParallelToolArgs is the regression test for #33.
+// It drives the streaming-callback wiring inside GenerateWithCallbacks with a
+// fake fantasy.Agent that emits two parallel tool calls before either result.
+// Each OnToolResult must receive the args of its own tool call (matched by
+// ToolCallID), not the args of the last OnToolCall in the step.
+func TestGenerateWithCallbacks_ParallelToolArgs(t *testing.T) {
+	t.Parallel()
+
+	argsA := `{"name":"scheduled_jobs"}`
+	argsB := `{"name":"gmail_trigger"}`
+
+	fake := &fakeParallelAgent{
+		calls: []fantasy.ToolCallContent{
+			{ToolCallID: "kit-A", ToolName: "load_skill", Input: argsA},
+			{ToolCallID: "kit-B", ToolName: "load_skill", Input: argsB},
+		},
+		results: []fantasy.ToolResultContent{
+			{ToolCallID: "kit-A", ToolName: "load_skill", Result: fantasy.ToolResultOutputContentText{Text: "ok-A"}},
+			{ToolCallID: "kit-B", ToolName: "load_skill", Result: fantasy.ToolResultOutputContentText{Text: "ok-B"}},
+		},
+	}
+
+	a := &Agent{
+		fantasyAgent:     fake,
+		streamingEnabled: false, // exercise the "hasCallbacks" branch
+	}
+
+	var mu sync.Mutex
+	resultArgs := map[string]string{}
+	executionArgs := map[string]string{} // captured when running == false
+
+	cb := GenerateCallbacks{
+		OnToolExecution: func(id, _, args string, running bool) {
+			if running {
+				return
+			}
+			mu.Lock()
+			defer mu.Unlock()
+			executionArgs[id] = args
+		},
+		OnToolResult: func(id, _, args, _, _ string, _ bool) {
+			mu.Lock()
+			defer mu.Unlock()
+			resultArgs[id] = args
+		},
+	}
+
+	if _, err := a.GenerateWithCallbacks(context.Background(), nil, cb); err != nil {
+		t.Fatalf("GenerateWithCallbacks returned error: %v", err)
+	}
+
+	if got, want := resultArgs["kit-A"], argsA; got != want {
+		t.Errorf("OnToolResult for kit-A: args = %q, want %q", got, want)
+	}
+	if got, want := resultArgs["kit-B"], argsB; got != want {
+		t.Errorf("OnToolResult for kit-B: args = %q, want %q", got, want)
+	}
+	if got, want := executionArgs["kit-A"], argsA; got != want {
+		t.Errorf("OnToolExecution(finish) for kit-A: args = %q, want %q", got, want)
+	}
+	if got, want := executionArgs["kit-B"], argsB; got != want {
+		t.Errorf("OnToolExecution(finish) for kit-B: args = %q, want %q", got, want)
+	}
+}

internal/agent/factory.go

@@ -36,13 +36,28 @@ type AgentCreationOptions struct {
 	SpinnerFunc SpinnerFunc // Function to show spinner (provided by caller)
 	// DebugLogger is an optional logger for debugging MCP communications
 	DebugLogger tools.DebugLogger // Optional debug logger
+	// AuthHandler handles OAuth authorization for remote MCP servers
+	AuthHandler tools.MCPAuthHandler
+	// TokenStoreFactory, if non-nil, creates a custom token store for each
+	// remote MCP server's OAuth tokens. When nil, the default file-based
+	// token store is used.
+	TokenStoreFactory tools.TokenStoreFactory
 	// CoreTools overrides the default core tool set. If empty, core.AllTools()
 	// is used.
 	CoreTools []fantasy.AgentTool
+	// DisableCoreTools, when true, prevents loading any core tools.
+	// If both DisableCoreTools is true and CoreTools is empty, the agent
+	// will have no tools (useful for simple chat completions).
+	DisableCoreTools bool
 	// ToolWrapper wraps the combined tool list before agent creation.
 	ToolWrapper func([]fantasy.AgentTool) []fantasy.AgentTool
 	// ExtraTools are additional tools to include (e.g. from extensions).
 	ExtraTools []fantasy.AgentTool
+	// OnMCPServerLoaded, if non-nil, is called when each MCP server finishes
+	// loading (successfully or with error). Called from the background goroutine.
+	OnMCPServerLoaded func(serverName string, toolCount int, err error)
+	// MCPTaskConfig configures task-augmented tools/call execution.
+	MCPTaskConfig tools.MCPTaskConfig
 }
 
 // CreateAgent creates an agent with optional spinner for Ollama models.
@@ -50,15 +65,20 @@ type AgentCreationOptions struct {
 // Returns the created agent or an error if creation fails.
 func CreateAgent(ctx context.Context, opts *AgentCreationOptions) (*Agent, error) {
 	agentConfig := &AgentConfig{
-		ModelConfig:      opts.ModelConfig,
-		MCPConfig:        opts.MCPConfig,
-		SystemPrompt:     opts.SystemPrompt,
-		MaxSteps:         opts.MaxSteps,
-		StreamingEnabled: opts.StreamingEnabled,
-		DebugLogger:      opts.DebugLogger,
-		CoreTools:        opts.CoreTools,
-		ToolWrapper:      opts.ToolWrapper,
-		ExtraTools:       opts.ExtraTools,
+		ModelConfig:       opts.ModelConfig,
+		MCPConfig:         opts.MCPConfig,
+		SystemPrompt:      opts.SystemPrompt,
+		MaxSteps:          opts.MaxSteps,
+		StreamingEnabled:  opts.StreamingEnabled,
+		DebugLogger:       opts.DebugLogger,
+		AuthHandler:       opts.AuthHandler,
+		TokenStoreFactory: opts.TokenStoreFactory,
+		CoreTools:         opts.CoreTools,
+		DisableCoreTools:  opts.DisableCoreTools,
+		ToolWrapper:       opts.ToolWrapper,
+		ExtraTools:        opts.ExtraTools,
+		OnMCPServerLoaded: opts.OnMCPServerLoaded,
+		MCPTaskConfig:     opts.MCPTaskConfig,
 	}
 
 	var agent *Agent

internal/agent/mcp_adapter.go

@@ -0,0 +1,88 @@
+package agent
+
+import (
+	"context"
+	"fmt"
+
+	"charm.land/fantasy"
+
+	"github.com/mark3labs/kit/internal/tools"
+)
+
+// mcpExecutor is the subset of *tools.MCPToolManager that the adapter
+// actually uses. Extracted as an interface so the adapter is unit-testable
+// without constructing a full manager + connection pool.
+type mcpExecutor interface {
+	ExecuteTool(ctx context.Context, prefixedName, inputJSON string) (*tools.MCPToolResult, error)
+}
+
+// mcpAgentTool adapts an tools.MCPTool to the fantasy.AgentTool interface.
+// This keeps the fantasy dependency confined to the agent layer — the tools
+// package is a pure MCP client library with no LLM framework dependency.
+type mcpAgentTool struct {
+	tool            tools.MCPTool
+	exec            mcpExecutor
+	providerOptions fantasy.ProviderOptions
+}
+
+// Info returns the fantasy tool info including name, description, and parameter schema.
+func (t *mcpAgentTool) Info() fantasy.ToolInfo {
+	return fantasy.ToolInfo{
+		Name:        t.tool.Name,
+		Description: t.tool.Description,
+		Parameters:  t.tool.Parameters,
+		Required:    t.tool.Required,
+	}
+}
+
+// Run executes the MCP tool by delegating to the MCPToolManager.
+//
+// MCP-side failures (JSON-RPC protocol errors, transport failures, schema
+// validation rejections from the server) are surfaced to the model as soft
+// tool errors rather than escalated to a critical agent error. This matches
+// the contract that native Kit tools follow via kit.ErrorResult(...) and
+// lets the model self-correct (e.g. retry with a fixed argument shape) or
+// give up gracefully rather than aborting the turn mid-run.
+//
+// Context cancellation is the one exception: if the caller cancelled the
+// context the turn was aborted intentionally, so we propagate the ctx error
+// to let the agent loop unwind cleanly.
+func (t *mcpAgentTool) Run(ctx context.Context, call fantasy.ToolCall) (fantasy.ToolResponse, error) {
+	result, err := t.exec.ExecuteTool(ctx, t.tool.Name, call.Input)
+	if err != nil {
+		if ctxErr := ctx.Err(); ctxErr != nil {
+			return fantasy.ToolResponse{}, ctxErr
+		}
+		return fantasy.NewTextErrorResponse(
+			fmt.Sprintf("MCP tool %q failed: %s", t.tool.Name, err.Error()),
+		), nil
+	}
+
+	if result.IsError {
+		return fantasy.NewTextErrorResponse(result.Content), nil
+	}
+	return fantasy.NewTextResponse(result.Content), nil
+}
+
+// ProviderOptions returns provider-specific options for this tool.
+func (t *mcpAgentTool) ProviderOptions() fantasy.ProviderOptions {
+	return t.providerOptions
+}
+
+// SetProviderOptions sets provider-specific options for this tool.
+func (t *mcpAgentTool) SetProviderOptions(opts fantasy.ProviderOptions) {
+	t.providerOptions = opts
+}
+
+// mcpToolsToAgentTools converts a slice of MCPTool to fantasy.AgentTool
+// implementations that route execution through the MCPToolManager.
+func mcpToolsToAgentTools(mcpTools []tools.MCPTool, manager *tools.MCPToolManager) []fantasy.AgentTool {
+	agentTools := make([]fantasy.AgentTool, len(mcpTools))
+	for i, t := range mcpTools {
+		agentTools[i] = &mcpAgentTool{
+			tool: t,
+			exec: manager,
+		}
+	}
+	return agentTools
+}

internal/agent/mcp_adapter_test.go

@@ -0,0 +1,158 @@
+package agent
+
+import (
+	"context"
+	"errors"
+	"strings"
+	"testing"
+	"time"
+
+	"charm.land/fantasy"
+
+	"github.com/mark3labs/kit/internal/tools"
+)
+
+// stubExecutor lets each test script the (result, err) pair returned by
+// ExecuteTool. The adapter holds an mcpExecutor interface, so this is the
+// only seam the tests need.
+type stubExecutor struct {
+	result *tools.MCPToolResult
+	err    error
+	// called records the last invocation for assertion.
+	called bool
+	name   string
+	input  string
+}
+
+func (s *stubExecutor) ExecuteTool(_ context.Context, prefixedName, inputJSON string) (*tools.MCPToolResult, error) {
+	s.called = true
+	s.name = prefixedName
+	s.input = inputJSON
+	return s.result, s.err
+}
+
+func newMCPAgentTool(exec mcpExecutor, name string) *mcpAgentTool {
+	return &mcpAgentTool{
+		tool: tools.MCPTool{Name: name},
+		exec: exec,
+	}
+}
+
+// Manager-side Go errors (JSON-RPC protocol errors, transport failures,
+// schema validation rejections from the MCP server) must be surfaced to
+// the model as soft tool errors so the agent loop can keep going. Aborting
+// the turn would discard all prior tool results — see issue #N.
+func TestMCPAgentTool_RPCErrorBecomesSoftError(t *testing.T) {
+	exec := &stubExecutor{
+		err: errors.New("MCP error -32602: Invalid params: missing field \"task\""),
+	}
+	tool := newMCPAgentTool(exec, "pubmed__search")
+
+	resp, err := tool.Run(context.Background(), fantasy.ToolCall{
+		ID:    "call-1",
+		Name:  "pubmed__search",
+		Input: `{"query":"foo"}`,
+	})
+
+	if err != nil {
+		t.Fatalf("expected nil error (soft), got %v", err)
+	}
+	if !resp.IsError {
+		t.Fatalf("expected IsError=true, got false")
+	}
+	if !strings.Contains(resp.Content, "pubmed__search") {
+		t.Errorf("expected tool name in error content, got %q", resp.Content)
+	}
+	if !strings.Contains(resp.Content, "-32602") {
+		t.Errorf("expected underlying error text in content, got %q", resp.Content)
+	}
+}
+
+// Context cancellation is the one error that must remain critical: it
+// means the caller intentionally aborted, and the agent loop needs to
+// unwind cleanly rather than burning more steps.
+func TestMCPAgentTool_CtxCancelStaysCritical(t *testing.T) {
+	exec := &stubExecutor{
+		// Real managers typically return ctx.Err() (or a wrapper) when the
+		// context is cancelled mid-call.
+		err: context.Canceled,
+	}
+	tool := newMCPAgentTool(exec, "slow__tool")
+
+	ctx, cancel := context.WithCancel(context.Background())
+	cancel()
+
+	resp, err := tool.Run(ctx, fantasy.ToolCall{Name: "slow__tool"})
+
+	if !errors.Is(err, context.Canceled) {
+		t.Fatalf("expected context.Canceled, got %v", err)
+	}
+	if resp.IsError || resp.Content != "" {
+		t.Errorf("expected empty response on critical error, got IsError=%v Content=%q", resp.IsError, resp.Content)
+	}
+}
+
+// Deadline-exceeded behaves the same as cancellation: ctx.Err() is
+// non-nil, so the adapter must propagate the critical error rather than
+// converting the executor's error into a soft response.
+func TestMCPAgentTool_CtxDeadlineStaysCritical(t *testing.T) {
+	exec := &stubExecutor{err: context.DeadlineExceeded}
+	tool := newMCPAgentTool(exec, "slow__tool")
+
+	ctx, cancel := context.WithDeadline(context.Background(), time.Now().Add(-time.Second))
+	defer cancel()
+
+	resp, err := tool.Run(ctx, fantasy.ToolCall{Name: "slow__tool"})
+	if !errors.Is(err, context.DeadlineExceeded) {
+		t.Fatalf("expected context.DeadlineExceeded, got %v", err)
+	}
+	if resp.IsError || resp.Content != "" {
+		t.Errorf("expected empty response on critical error, got IsError=%v Content=%q", resp.IsError, resp.Content)
+	}
+}
+
+// Server-side soft errors (CallToolResult{ isError: true }) must continue
+// to flow through as soft errors — this was the existing behavior and
+// must not regress.
+func TestMCPAgentTool_ServerIsErrorRemainsSoftError(t *testing.T) {
+	exec := &stubExecutor{
+		result: &tools.MCPToolResult{
+			IsError: true,
+			Content: "search service is rate limited; try again in 30s",
+		},
+	}
+	tool := newMCPAgentTool(exec, "pubmed__search")
+
+	resp, err := tool.Run(context.Background(), fantasy.ToolCall{Name: "pubmed__search"})
+	if err != nil {
+		t.Fatalf("expected nil error, got %v", err)
+	}
+	if !resp.IsError {
+		t.Fatalf("expected IsError=true, got false")
+	}
+	if resp.Content != "search service is rate limited; try again in 30s" {
+		t.Errorf("expected pass-through content, got %q", resp.Content)
+	}
+}
+
+// Happy path: ordinary successful tool result is passed through unchanged.
+func TestMCPAgentTool_SuccessIsPassthrough(t *testing.T) {
+	exec := &stubExecutor{
+		result: &tools.MCPToolResult{
+			IsError: false,
+			Content: `{"hits":3}`,
+		},
+	}
+	tool := newMCPAgentTool(exec, "pubmed__search")
+
+	resp, err := tool.Run(context.Background(), fantasy.ToolCall{Name: "pubmed__search"})
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if resp.IsError {
+		t.Fatalf("expected IsError=false")
+	}
+	if resp.Content != `{"hits":3}` {
+		t.Errorf("expected pass-through content, got %q", resp.Content)
+	}
+}

internal/agent/steer.go

@@ -1,6 +1,17 @@
 package agent
 
-import "context"
+import (
+	"context"
+
+	"charm.land/fantasy"
+)
+
+// SteerMessage carries a steering prompt and optional file attachments
+// (e.g. clipboard images) through the steer channel.
+type SteerMessage struct {
+	Text  string
+	Files []fantasy.FilePart
+}
 
 // steerChKey is the context key for the steer channel.
 type steerChKey struct{}
@@ -11,7 +22,7 @@ type steerConsumedKey struct{}
 // ContextWithSteerCh returns a new context with the steer channel attached.
 // The agent's PrepareStep function checks this channel between steps and
 // injects any pending steer messages as user messages before the next LLM call.
-func ContextWithSteerCh(ctx context.Context, ch <-chan string) context.Context {
+func ContextWithSteerCh(ctx context.Context, ch <-chan SteerMessage) context.Context {
 	return context.WithValue(ctx, steerChKey{}, ch)
 }
 
@@ -23,8 +34,8 @@ func ContextWithSteerConsumed(ctx context.Context, fn func(count int)) context.C
 }
 
 // steerChFromContext extracts the steer channel from the context, or nil.
-func steerChFromContext(ctx context.Context) <-chan string {
-	ch, _ := ctx.Value(steerChKey{}).(<-chan string)
+func steerChFromContext(ctx context.Context) <-chan SteerMessage {
+	ch, _ := ctx.Value(steerChKey{}).(<-chan SteerMessage)
 	return ch
 }
 

internal/app/app.go

@@ -20,7 +20,7 @@ import (
 // queueItem holds a prompt and optional image attachments for the execution queue.
 type queueItem struct {
 	Prompt string
-	Files  []fantasy.FilePart
+	Files  []kit.LLMFilePart
 }
 
 // App is the application-layer orchestrator. It owns the agentic loop,
@@ -70,19 +70,29 @@ type App struct {
 	rootCtx    context.Context
 	rootCancel context.CancelFunc
 
-	// widgetUpdatePending is set to true when a WidgetUpdateEvent has been
-	// sent to the TUI but not yet consumed by its event loop. While the flag
-	// is set, subsequent NotifyWidgetUpdate calls are coalesced (dropped) to
-	// prevent fast extension tickers from flooding the BubbleTea mailbox with
-	// redundant re-render triggers. The flag is cleared after a short debounce
-	// (~1 frame) so new updates are always let through once the TUI has had a
-	// chance to process the pending event.
-	widgetUpdatePending atomic.Bool
+	// widgetUpdatePending is set to true while a WidgetUpdateEvent burst is
+	// being coalesced. The leading edge fires immediately; subsequent calls
+	// within the debounce window set widgetUpdateTrailing so a final event
+	// is delivered with the latest runner state at the end of the window.
+	// Without the trailing send, a rapid SetWidget→RemoveWidget pair (e.g.
+	// SubagentEnd pushing a final frame then removing the widget) would let
+	// the second call get silently dropped, leaving the TUI's layout stuck
+	// on the pre-removal widget height — visible as empty rows below the
+	// status bar after the widget disappears.
+	widgetUpdatePending  atomic.Bool
+	widgetUpdateTrailing atomic.Bool
+
+	// steerDrainFn is the test seam used by releaseBusyAfterCompact to pull
+	// any steer messages that arrived during compaction. In production it is
+	// nil and the helper falls back to a.opts.Kit.DrainSteer(); tests that
+	// need to exercise the steer-drain path without standing up a full
+	// *kit.Kit can set this field directly to inject fake items.
+	steerDrainFn func() []queueItem
 }
 
 // New creates a new App with the provided options and pre-loaded messages.
 // initialMessages may be nil or empty for a fresh session.
-func New(opts Options, initialMessages []fantasy.Message) *App {
+func New(opts Options, initialMessages []kit.LLMMessage) *App {
 	rootCtx, rootCancel := context.WithCancel(context.Background())
 	return &App{
 		opts:       opts,
@@ -126,9 +136,8 @@ func (a *App) Run(prompt string) int {
 // If the app is idle the prompt executes immediately; otherwise it is queued.
 // Returns the current queue depth (0 = started immediately, >0 = queued).
 //
-// Satisfies ui.AppController (via RunWithImages which converts ImageAttachment
-// to fantasy.FilePart).
-func (a *App) RunWithFiles(prompt string, files []fantasy.FilePart) int {
+// Satisfies ui.AppController.
+func (a *App) RunWithFiles(prompt string, files []kit.LLMFilePart) int {
 	a.mu.Lock()
 
 	if a.closed {
@@ -163,6 +172,24 @@ func (a *App) CancelCurrentStep() {
 	cancel()
 }
 
+// IsBusy returns true when the agent is currently processing a turn.
+func (a *App) IsBusy() bool {
+	a.mu.Lock()
+	defer a.mu.Unlock()
+	return a.busy
+}
+
+// Abort cancels the current agent step (if running) and clears the queue.
+// Unlike InterruptAndSend, no new message is injected — the agent simply
+// stops. Safe to call when idle (no-op).
+func (a *App) Abort() {
+	a.mu.Lock()
+	a.queue = a.queue[:0]
+	cancel := a.cancelStep
+	a.mu.Unlock()
+	cancel()
+}
+
 // QueueLength returns the number of prompts currently waiting in the queue.
 //
 // Satisfies ui.AppController.
@@ -188,6 +215,15 @@ func (a *App) QueueLength() int {
 //
 // Satisfies ui.AppController.
 func (a *App) Steer(prompt string) int {
+	return a.SteerWithFiles(prompt, nil)
+}
+
+// SteerWithFiles injects a steering message with optional file attachments
+// (e.g. pasted images) into the currently running agent turn. Behaves like
+// Steer but includes file parts alongside the text.
+//
+// Satisfies ui.AppController.
+func (a *App) SteerWithFiles(prompt string, files []kit.LLMFilePart) int {
 	a.mu.Lock()
 
 	if a.closed {
@@ -196,8 +232,8 @@ func (a *App) Steer(prompt string) int {
 	}
 
 	if !a.busy {
-		// Not busy — start immediately, same as Run().
-		item := queueItem{Prompt: prompt}
+		// Not busy — start immediately, same as RunWithFiles().
+		item := queueItem{Prompt: prompt, Files: files}
 		a.busy = true
 		a.wg.Add(1)
 		a.mu.Unlock()
@@ -212,7 +248,7 @@ func (a *App) Steer(prompt string) int {
 	// execution, before next LLM call). If PrepareStep doesn't fire
 	// (text-only response), drainQueue will pick it up after the turn.
 	if a.opts.Kit != nil {
-		a.opts.Kit.InjectSteer(prompt)
+		a.opts.Kit.InjectSteerWithFiles(prompt, files)
 	}
 	return 1
 }
@@ -314,12 +350,12 @@ func (a *App) SwitchTreeSession(ts *session.TreeManager) {
 //
 // Satisfies ui.AppController.
 func (a *App) AddContextMessage(text string) {
-	msg := fantasy.NewUserMessage(text)
-	a.store.Add(msg)
+	kitMsg := fantasy.NewUserMessage(text)
+	a.store.Add(kitMsg)
 
 	// Persist to tree session if active.
 	if ts := a.opts.TreeSession; ts != nil {
-		_, _ = ts.AppendLLMMessage(msg)
+		_, _ = ts.AppendLLMMessage(fantasy.NewUserMessage(text))
 	}
 }
 
@@ -330,6 +366,10 @@ func (a *App) AddContextMessage(text string) {
 // tea.Program. customInstructions is optional text appended to the summary
 // prompt (e.g. "Focus on the API design decisions").
 //
+// Any prompts queued via Run/RunWithFiles or steering messages injected via
+// Steer/SteerWithFiles while compaction is running are flushed automatically
+// once compaction completes (see releaseBusyAfterCompact).
+//
 // Satisfies ui.AppController.
 func (a *App) CompactConversation(customInstructions string) error {
 	a.mu.Lock()
@@ -351,11 +391,16 @@ func (a *App) CompactConversation(customInstructions string) error {
 
 	go func() {
 		defer a.wg.Done()
-		defer func() {
-			a.mu.Lock()
-			a.busy = false
-			a.mu.Unlock()
-		}()
+		defer a.releaseBusyAfterCompact()
+
+		// Subscribe to SDK events for streaming compaction summary to the TUI.
+		sendFn := func(msg tea.Msg) {
+			if a.program != nil {
+				a.program.Send(msg)
+			}
+		}
+		unsub := a.subscribeSDKEvents(sendFn, nil)
+		defer unsub()
 
 		result, err := a.opts.Kit.Compact(a.rootCtx, nil, customInstructions)
 		if err != nil {
@@ -382,6 +427,152 @@ func (a *App) CompactConversation(customInstructions string) error {
 	return nil
 }
 
+// CompactAsync is like CompactConversation but calls onComplete/onError
+// callbacks instead of sending TUI events. Used by the extension API's
+// ctx.Compact() which needs callback-based notification.
+//
+// Like CompactConversation, any prompts/steer messages received during
+// compaction are flushed automatically once compaction finishes.
+func (a *App) CompactAsync(customInstructions string, onComplete func(), onError func(string)) error {
+	a.mu.Lock()
+	if a.closed {
+		a.mu.Unlock()
+		return fmt.Errorf("app is closed")
+	}
+	if a.busy {
+		a.mu.Unlock()
+		return fmt.Errorf("cannot compact while the agent is working")
+	}
+	if a.opts.Kit == nil {
+		a.mu.Unlock()
+		return fmt.Errorf("SDK instance not available")
+	}
+	a.busy = true
+	a.wg.Add(1)
+	a.mu.Unlock()
+
+	go func() {
+		defer a.wg.Done()
+		defer a.releaseBusyAfterCompact()
+
+		// Subscribe to SDK events for streaming compaction summary to the TUI.
+		sendFn := func(msg tea.Msg) {
+			if a.program != nil {
+				a.program.Send(msg)
+			}
+		}
+		unsub := a.subscribeSDKEvents(sendFn, nil)
+		defer unsub()
+
+		result, err := a.opts.Kit.Compact(a.rootCtx, nil, customInstructions)
+		if err != nil {
+			a.sendEvent(CompactErrorEvent{Err: err})
+			if onError != nil {
+				onError(err.Error())
+			}
+			return
+		}
+		if result == nil {
+			a.sendEvent(CompactErrorEvent{Err: fmt.Errorf("nothing to compact")})
+			if onError != nil {
+				onError("nothing to compact")
+			}
+			return
+		}
+
+		// Sync in-memory store with the compacted session.
+		if a.opts.TreeSession != nil {
+			a.store.Replace(a.opts.TreeSession.GetLLMMessages())
+		}
+
+		a.sendEvent(CompactCompleteEvent{
+			Summary:         result.Summary,
+			OriginalTokens:  result.OriginalTokens,
+			CompactedTokens: result.CompactedTokens,
+			MessagesRemoved: result.MessagesRemoved,
+		})
+		if onComplete != nil {
+			onComplete()
+		}
+	}()
+	return nil
+}
+
+// releaseBusyAfterCompact is the deferred tail that runs at the end of every
+// compaction goroutine (success, error, or panic-after-recover paths). It
+// flips a.busy back to false, but before doing so it checks whether any
+// prompts piled up while compaction was running:
+//
+//   - Run/RunWithFiles append to a.queue when a.busy is set.
+//   - Steer/SteerWithFiles deposit messages into the SDK steer channel via
+//     Kit.InjectSteerWithFiles when a.busy is set.
+//
+// Without this hand-off the queue would sit idle until the user submits
+// another prompt — see issue #27. If we find anything pending we keep busy
+// set, splice the steer messages to the front of the queue, and start a
+// fresh drainQueue goroutine to deliver them as a single batched turn.
+func (a *App) releaseBusyAfterCompact() {
+	// Pull steer messages outside the app mutex; DrainSteer takes its own
+	// internal lock and we don't want to nest the two. The test seam
+	// (a.steerDrainFn) takes precedence so unit tests can inject fake
+	// steer items without a real *kit.Kit.
+	var steerItems []queueItem
+	switch {
+	case a.steerDrainFn != nil:
+		steerItems = a.steerDrainFn()
+	case a.opts.Kit != nil:
+		if leftover := a.opts.Kit.DrainSteer(); len(leftover) > 0 {
+			steerItems = make([]queueItem, len(leftover))
+			for i, sm := range leftover {
+				steerItems[i] = queueItem{Prompt: sm.Text, Files: sm.Files}
+			}
+		}
+	}
+
+	a.mu.Lock()
+	// If the app was closed while compaction was running, drop everything
+	// and just clear busy. Run/Steer would have rejected new items already
+	// after Close(), but this guards against in-flight items that slipped
+	// in just before closed was set.
+	if a.closed {
+		a.queue = a.queue[:0]
+		a.busy = false
+		a.mu.Unlock()
+		return
+	}
+
+	// Combine steer-channel items (front) with the in-memory queue (back).
+	// Steer messages are placed first so they retain their "act now"
+	// semantics relative to ordinary queued prompts that arrived later.
+	pending := append(steerItems, a.queue...)
+	a.queue = a.queue[:0]
+
+	if len(pending) == 0 {
+		a.busy = false
+		a.mu.Unlock()
+		return
+	}
+
+	// Hand off to drainQueue: it will pick up the first item directly and
+	// scoop the rest from a.queue on its first iteration.
+	first := pending[0]
+	if len(pending) > 1 {
+		a.queue = append(a.queue, pending[1:]...)
+	}
+	// Stay busy across the goroutine swap.
+	a.wg.Add(1)
+	a.mu.Unlock()
+
+	// Notify the UI that steer-channel messages were consumed so the
+	// steering badge can clear; ordinary queued prompts will be reflected
+	// by the QueueUpdatedEvent that drainQueue emits as it picks them up.
+	if len(steerItems) > 0 {
+		a.sendEvent(SteerConsumedEvent{})
+	}
+
+	go a.drainQueue(first)
+}
+
 // --------------------------------------------------------------------------
 // Non-interactive execution
 // --------------------------------------------------------------------------
@@ -390,6 +581,12 @@ func (a *App) CompactConversation(customInstructions string) error {
 // response text to stdout. No intermediate events are emitted. Blocks until
 // the step completes or ctx is cancelled.
 func (a *App) RunOnce(ctx context.Context, prompt string) error {
+	return a.RunOnceWithFiles(ctx, prompt, nil)
+}
+
+// RunOnceWithFiles executes a single agent step synchronously with optional
+// multimodal file attachments. Prints the response to stdout and returns.
+func (a *App) RunOnceWithFiles(ctx context.Context, prompt string, files []kit.LLMFilePart) error {
 	stepCtx, cancel := context.WithCancel(ctx)
 	defer cancel()
 
@@ -397,7 +594,7 @@ func (a *App) RunOnce(ctx context.Context, prompt string) error {
 	a.cancelStep = cancel
 	a.mu.Unlock()
 
-	result, err := a.executeStep(stepCtx, prompt, nil, nil)
+	result, err := a.executeStep(stepCtx, prompt, nil, files)
 	if err != nil {
 		return err
 	}
@@ -412,6 +609,12 @@ func (a *App) RunOnce(ctx context.Context, prompt string) error {
 // full TurnResult without printing anything. This is used by --json mode to
 // capture structured output for serialization.
 func (a *App) RunOnceResult(ctx context.Context, prompt string) (*kit.TurnResult, error) {
+	return a.RunOnceResultWithFiles(ctx, prompt, nil)
+}
+
+// RunOnceResultWithFiles executes a single agent step synchronously with
+// optional multimodal file attachments and returns the full TurnResult.
+func (a *App) RunOnceResultWithFiles(ctx context.Context, prompt string, files []kit.LLMFilePart) (*kit.TurnResult, error) {
 	stepCtx, cancel := context.WithCancel(ctx)
 	defer cancel()
 
@@ -419,7 +622,7 @@ func (a *App) RunOnceResult(ctx context.Context, prompt string) (*kit.TurnResult
 	a.cancelStep = cancel
 	a.mu.Unlock()
 
-	return a.executeStep(stepCtx, prompt, nil, nil)
+	return a.executeStep(stepCtx, prompt, nil, files)
 }
 
 // RunOnceWithDisplay executes a single agent step synchronously, sending
@@ -433,6 +636,12 @@ func (a *App) RunOnceResult(ctx context.Context, prompt string) (*kit.TurnResult
 //
 // Blocks until the step completes or ctx is cancelled.
 func (a *App) RunOnceWithDisplay(ctx context.Context, prompt string, eventFn func(tea.Msg)) error {
+	return a.RunOnceWithDisplayAndFiles(ctx, prompt, eventFn, nil)
+}
+
+// RunOnceWithDisplayAndFiles executes a single agent step synchronously with
+// optional multimodal file attachments, sending intermediate display events.
+func (a *App) RunOnceWithDisplayAndFiles(ctx context.Context, prompt string, eventFn func(tea.Msg), files []kit.LLMFilePart) error {
 	stepCtx, cancel := context.WithCancel(ctx)
 	defer cancel()
 
@@ -440,7 +649,7 @@ func (a *App) RunOnceWithDisplay(ctx context.Context, prompt string, eventFn fun
 	a.cancelStep = cancel
 	a.mu.Unlock()
 
-	result, err := a.executeStep(stepCtx, prompt, eventFn, nil)
+	result, err := a.executeStep(stepCtx, prompt, eventFn, files)
 	if err != nil {
 		return err
 	}
@@ -506,11 +715,10 @@ func (a *App) drainQueue(first queueItem) {
 		a.mu.Lock()
 		items = append(items, a.queue...)
 		a.queue = a.queue[:0] // Clear the queue
-		queueLen := len(a.queue)
 		a.mu.Unlock()
 
-		// Send queue updated event (queue is now empty)
-		a.sendEvent(QueueUpdatedEvent{Length: queueLen})
+		// Notify UI: all queued messages have been consumed into this batch.
+		a.sendEvent(QueueUpdatedEvent{Length: 0})
 
 		// Process all collected items as a single batch
 		a.runQueueBatch(items)
@@ -523,8 +731,8 @@ func (a *App) drainQueue(first queueItem) {
 			if leftover := a.opts.Kit.DrainSteer(); len(leftover) > 0 {
 				a.mu.Lock()
 				steerItems := make([]queueItem, len(leftover))
-				for i, text := range leftover {
-					steerItems[i] = queueItem{Prompt: text}
+				for i, sm := range leftover {
+					steerItems[i] = queueItem{Prompt: sm.Text, Files: sm.Files}
 				}
 				a.queue = append(steerItems, a.queue...)
 				a.mu.Unlock()
@@ -543,6 +751,11 @@ func (a *App) drainQueue(first queueItem) {
 		}
 		a.mu.Unlock()
 
+		if hasMore {
+			// Notify UI: these newly queued messages have been consumed into the next batch.
+			a.sendEvent(QueueUpdatedEvent{Length: 0})
+		}
+
 		if !hasMore {
 			// No more items, we're done
 			break
@@ -609,7 +822,7 @@ func (a *App) runQueueBatch(items []queueItem) {
 // executeStep runs a single agentic step by delegating to the SDK's
 // PromptResult() (or PromptResultWithFiles for multimodal), which handles
 // session persistence, hooks, extension events, and the generation loop.
-func (a *App) executeStep(ctx context.Context, prompt string, eventFn func(tea.Msg), files []fantasy.FilePart) (*kit.TurnResult, error) {
+func (a *App) executeStep(ctx context.Context, prompt string, eventFn func(tea.Msg), files []kit.LLMFilePart) (*kit.TurnResult, error) {
 	// Test hook: bypass SDK entirely.
 	if a.opts.PromptFunc != nil {
 		return a.opts.PromptFunc(ctx, prompt)
@@ -759,6 +972,12 @@ func (a *App) subscribeSDKEvents(sendFn func(tea.Msg), stepUsageSeen *atomic.Boo
 		switch ev := e.(type) {
 		case kit.ToolCallEvent:
 			sendFn(ToolCallStartedEvent{ToolCallID: ev.ToolCallID, ToolName: ev.ToolName, ToolArgs: ev.ToolArgs})
+		case kit.ToolCallStartEvent:
+			sendFn(ToolCallInputStartEvent{ToolCallID: ev.ToolCallID, ToolName: ev.ToolName, ToolKind: ev.ToolKind})
+		case kit.ToolCallDeltaEvent:
+			sendFn(ToolCallInputDeltaEvent{ToolCallID: ev.ToolCallID, Delta: ev.Delta})
+		case kit.ToolCallEndEvent:
+			sendFn(ToolCallInputEndEvent{ToolCallID: ev.ToolCallID})
 		case kit.ToolExecutionStartEvent:
 			sendFn(ToolExecutionEvent{ToolCallID: ev.ToolCallID, ToolName: ev.ToolName, ToolArgs: ev.ToolArgs, IsStarting: true})
 		case kit.ToolExecutionEndEvent:
@@ -776,6 +995,8 @@ func (a *App) subscribeSDKEvents(sendFn func(tea.Msg), stepUsageSeen *atomic.Boo
 			sendFn(StreamChunkEvent{Content: ev.Chunk})
 		case kit.ReasoningDeltaEvent:
 			sendFn(ReasoningChunkEvent{Delta: ev.Delta})
+		case kit.ReasoningCompleteEvent:
+			sendFn(ReasoningCompleteEvent{})
 		case kit.ToolOutputEvent:
 			sendFn(ToolOutputEvent{
 				ToolCallID: ev.ToolCallID,
@@ -786,7 +1007,23 @@ func (a *App) subscribeSDKEvents(sendFn func(tea.Msg), stepUsageSeen *atomic.Boo
 		case kit.SteerConsumedEvent:
 			sendFn(SteerConsumedEvent{})
 		case kit.StepUsageEvent:
-			a.recordStepUsage(ev, stepUsageSeen)
+			a.recordStepUsage(ev, stepUsageSeen, sendFn)
+		case kit.PasswordPromptEvent:
+			// Convert SDK PasswordPromptEvent to app PasswordPromptEvent
+			// The TUI will handle this and send the response back
+			responseCh := make(chan PasswordPromptResponse, 1)
+			sendFn(PasswordPromptEvent{
+				Prompt:     ev.Prompt,
+				ResponseCh: responseCh,
+			})
+			// Wait for TUI response and forward to SDK
+			resp := <-responseCh
+			ev.ResponseCh <- kit.PasswordPromptResponse{
+				Password:  resp.Password,
+				Cancelled: resp.Cancelled,
+			}
+		case kit.TurnEndEvent:
+			a.handleTurnEnd(ev, sendFn)
 		}
 	}))
 
@@ -797,6 +1034,64 @@ func (a *App) subscribeSDKEvents(sendFn func(tea.Msg), stepUsageSeen *atomic.Boo
 	}
 }
 
+// handleTurnEnd inspects a turn's final StopReason and surfaces actionable
+// feedback to the user when the turn ended in a state they can act on.
+//
+// Today the only surfaced case is FinishReasonLength — the model hit its
+// configured max_output_tokens budget and the reply was truncated. Without
+// this banner the TUI used to swallow the truncation silently, leading to
+// "ghost" cut-offs with no indication of why.
+//
+// Separated from subscribeSDKEvents so tests can exercise it directly via a
+// stubbed sendFn without standing up a full Kit.
+func (a *App) handleTurnEnd(ev kit.TurnEndEvent, sendFn func(tea.Msg)) {
+	if sendFn == nil {
+		return
+	}
+	if ev.StopReason != kit.FinishReasonLength {
+		return
+	}
+	sendFn(ExtensionPrintEvent{
+		Level: "info",
+		Text:  a.formatMaxTokensTruncatedMessage(),
+	})
+}
+
+// formatMaxTokensTruncatedMessage builds the user-facing explanation for a
+// truncated turn. It reports the active max_output_tokens budget and, when
+// known, the model's catalog output ceiling so the user can judge how much
+// headroom is available.
+func (a *App) formatMaxTokensTruncatedMessage() string {
+	k := a.opts.Kit
+	if k == nil {
+		// Extremely early / test-stub case: still emit a useful generic hint.
+		return "⚠ Response truncated: the model hit the configured max_output_tokens limit. " +
+			"Raise it with --max-tokens N, KIT_MAX_TOKENS=N, or per-model " +
+			"modelSettings[provider/model].maxTokens in config."
+	}
+	current := k.MaxTokens()
+	ceiling := k.MaxOutputLimit()
+	model := k.GetModelString()
+
+	msg := "⚠ Response truncated: "
+	if model != "" {
+		msg += fmt.Sprintf("%s hit the configured max_output_tokens limit", model)
+	} else {
+		msg += "the model hit the configured max_output_tokens limit"
+	}
+	if current > 0 {
+		msg += fmt.Sprintf(" (%d)", current)
+	}
+	msg += "."
+	if ceiling > 0 && current > 0 && ceiling > current {
+		msg += fmt.Sprintf(" This model supports up to %d output tokens.", ceiling)
+	}
+	msg += "\n\nRaise it with --max-tokens N, KIT_MAX_TOKENS=N, " +
+		"or per-model modelSettings[provider/model].maxTokens in your config. " +
+		"Re-run the last prompt after raising it to get the full response."
+	return msg
+}
+
 // QuitFromExtension triggers a graceful shutdown. In interactive mode it
 // sends a tea.QuitMsg to the program so the TUI exits cleanly. In
 // non-interactive mode it cancels the root context, stopping any in-flight
@@ -817,7 +1112,8 @@ func (a *App) QuitFromExtension() {
 // controls styling: "" for plain text, "info" for a system message block,
 // "error" for an error block. In interactive mode it sends an
 // ExtensionPrintEvent through the program so the TUI can render it with the
-// appropriate renderer. In non-interactive mode it falls back to stdout.
+// appropriate renderer. In non-interactive mode it falls back to stderr with
+// a level prefix so errors are distinguishable from plain output.
 func (a *App) PrintFromExtension(level, text string) {
 	a.mu.Lock()
 	prog := a.program
@@ -826,8 +1122,16 @@ func (a *App) PrintFromExtension(level, text string) {
 		prog.Send(ExtensionPrintEvent{Text: text, Level: level})
 		return
 	}
-	// Non-interactive fallback: write directly to stdout.
-	fmt.Println(text)
+	// Non-interactive fallback: write to stderr with a level prefix so that
+	// errors and info messages are distinguishable from plain output.
+	switch level {
+	case "error":
+		fmt.Fprintf(os.Stderr, "[ERROR] %s\n", text)
+	case "info":
+		fmt.Fprintf(os.Stderr, "[INFO] %s\n", text)
+	default:
+		fmt.Println(text)
+	}
 }
 
 // SetEditorTextFromExtension sends an EditorTextSetEvent to the TUI to
@@ -856,31 +1160,87 @@ func (a *App) NotifyModelChanged(provider, model string) {
 // extension widgets. Called from the extension context's SetWidget/RemoveWidget
 // closures. In non-interactive mode this is a no-op (widgets are TUI-only).
 //
-// Coalescing: if a WidgetUpdateEvent is already queued and not yet consumed
-// by the TUI event loop, additional calls within the same ~16 ms window are
-// dropped. This prevents fast extension tickers from flooding BubbleTea's
-// mailbox with redundant re-render triggers.
+// Coalescing (leading + trailing edge): the first call in an idle period
+// fires immediately for responsiveness. Subsequent calls within a ~16 ms
+// debounce window are batched into a single trailing event delivered at
+// the end of the window. The trailing send is essential for correctness:
+// extensions routinely make tight SetWidget→RemoveWidget pairs (e.g. on
+// SubagentEnd) and silently dropping the second call would leave the TUI's
+// layout stuck on stale widget dimensions until some other event happens
+// to trigger a re-render.
 func (a *App) NotifyWidgetUpdate() {
-	// Coalesce: only one pending update at a time.
 	if !a.widgetUpdatePending.CompareAndSwap(false, true) {
+		// A leading-edge event is already in flight — mark that the runner
+		// state has changed again so the trailing send below picks it up.
+		a.widgetUpdateTrailing.Store(true)
 		return
 	}
 	a.mu.Lock()
 	prog := a.program
 	a.mu.Unlock()
-	if prog != nil {
-		prog.Send(WidgetUpdateEvent{})
-		// Reset the pending flag after a short debounce so subsequent calls
-		// within the same render cycle are also coalesced, but new updates
-		// after the cycle are allowed through.
-		go func() {
-			time.Sleep(16 * time.Millisecond) // ~1 frame at 60 fps
-			a.widgetUpdatePending.Store(false)
-		}()
-	} else {
+	if prog == nil {
 		// No program registered (non-interactive mode); clear the flag so
 		// future calls are never permanently blocked.
 		a.widgetUpdatePending.Store(false)
+		return
+	}
+	prog.Send(WidgetUpdateEvent{})
+	go func() {
+		time.Sleep(16 * time.Millisecond) // ~1 frame at 60 fps
+		// If any extra calls came in during the debounce window, deliver
+		// one trailing event so the TUI sees the latest widget state. We
+		// swap-and-test instead of plain-load so concurrent calls after
+		// the trailing send still race correctly with the pending reset.
+		if a.widgetUpdateTrailing.Swap(false) {
+			a.mu.Lock()
+			p := a.program
+			a.mu.Unlock()
+			if p != nil {
+				p.Send(WidgetUpdateEvent{})
+			}
+		}
+		a.widgetUpdatePending.Store(false)
+	}()
+}
+
+// NotifyContentReload sends a ContentReloadEvent to the TUI so it refreshes
+// prompt templates and skills from their provider callbacks. Called by file
+// watchers when .md/.txt files change in prompt or skill directories.
+// In non-interactive mode this is a no-op.
+func (a *App) NotifyContentReload() {
+	a.mu.Lock()
+	prog := a.program
+	a.mu.Unlock()
+	if prog != nil {
+		prog.Send(ContentReloadEvent{})
+	}
+}
+
+// NotifyMCPToolsReady sends an MCPToolsReadyEvent to the TUI so it refreshes
+// tool names and MCP tool count from provider callbacks. Called when background
+// MCP tool loading completes. In non-interactive mode this is a no-op.
+func (a *App) NotifyMCPToolsReady() {
+	a.mu.Lock()
+	prog := a.program
+	a.mu.Unlock()
+	if prog != nil {
+		prog.Send(MCPToolsReadyEvent{})
+	}
+}
+
+// NotifyMCPServerLoaded sends an MCPServerLoadedEvent to the TUI so it can
+// display a system message when a single MCP server finishes loading. Called
+// per server as background MCP tool loading progresses.
+func (a *App) NotifyMCPServerLoaded(serverName string, toolCount int, err error) {
+	a.mu.Lock()
+	prog := a.program
+	a.mu.Unlock()
+	if prog != nil {
+		prog.Send(MCPServerLoadedEvent{
+			ServerName: serverName,
+			ToolCount:  toolCount,
+			Error:      err,
+		})
 	}
 }
 
@@ -968,18 +1328,28 @@ func (a *App) PrintBlockFromExtension(opts extensions.PrintBlockOpts) {
 		})
 		return
 	}
-	// Non-interactive fallback.
+	// Non-interactive fallback: render a simple framed block to stderr so
+	// it is visually distinct from plain stdout output.
 	if opts.Subtitle != "" {
-		fmt.Printf("%s\n  — %s\n", opts.Text, opts.Subtitle)
+		fmt.Fprintf(os.Stderr, "--- %s ---\n%s\n", opts.Subtitle, opts.Text)
 	} else {
-		fmt.Println(opts.Text)
+		fmt.Fprintf(os.Stderr, "---\n%s\n---\n", opts.Text)
 	}
 }
 
 // recordStepUsage applies token/cost usage reported for a completed step.
 // Step usage events arrive even when a turn is later cancelled, so this keeps
 // the usage widget accurate on all stop paths.
-func (a *App) recordStepUsage(ev kit.StepUsageEvent, stepUsageSeen *atomic.Bool) {
+//
+// Both session totals (cost, token counts) and the context window fill level
+// are updated here so the status bar reflects progress after every LLM call,
+// not just at the end of the full turn. Context fill monotonically increases
+// across steps because each step re-sends the entire conversation plus any
+// new tool results, so the numbers only go up.
+//
+// sendFn is called with a UsageUpdatedEvent to trigger a TUI re-render so
+// the updated values are visible immediately.
+func (a *App) recordStepUsage(ev kit.StepUsageEvent, stepUsageSeen *atomic.Bool, sendFn func(tea.Msg)) {
 	hasUsage := ev.InputTokens > 0 || ev.OutputTokens > 0 || ev.CacheReadTokens > 0 || ev.CacheWriteTokens > 0
 	if a.opts.Debug {
 		log.Printf("[DEBUG] recordStepUsage: hasUsage=%v input=%d output=%d cacheRead=%d cacheWrite=%d",
@@ -1000,10 +1370,21 @@ func (a *App) recordStepUsage(ev kit.StepUsageEvent, stepUsageSeen *atomic.Bool)
 		int(ev.CacheReadTokens),
 		int(ev.CacheWriteTokens),
 	)
-	// NOTE: We do NOT call SetContextTokens here. Context fill is set once
-	// at turn completion via updateUsageFromTurnResult using FinalUsage.InputTokens,
-	// which reflects the full accumulated context. Per-step context tokens would
-	// cause the display to jump around during multi-step tool calls.
+	// Update context window fill from this step's usage. Each step sends
+	// the full conversation to the LLM, so the reported token counts
+	// represent the actual context utilization at that point.
+	contextFill := int(ev.InputTokens) + int(ev.CacheReadTokens) + int(ev.CacheWriteTokens) + int(ev.OutputTokens)
+	if contextFill > 0 {
+		if a.opts.Debug {
+			log.Printf("[DEBUG] recordStepUsage: SetContextTokens=%d (Input=%d + CacheRead=%d + CacheWrite=%d + Output=%d)",
+				contextFill, ev.InputTokens, ev.CacheReadTokens, ev.CacheWriteTokens, ev.OutputTokens)
+		}
+		a.opts.UsageTracker.SetContextTokens(contextFill)
+	}
+	// Notify the TUI so it re-renders the status bar with updated values.
+	if sendFn != nil {
+		sendFn(UsageUpdatedEvent{})
+	}
 }
 
 // updateUsageFromTurnResult records token usage from an SDK TurnResult into the
@@ -1067,15 +1448,30 @@ func (a *App) updateUsageFromTurnResult(result *kit.TurnResult, userPrompt strin
 	}
 
 	// --- Context window fill (drives the % bar) ---
-	// Use FinalUsage.InputTokens as the context window fill. The API's InputTokens
-	// already includes the full conversation history (system prompt + all previous
-	// messages + current user message). Adding OutputTokens would double-count since
-	// the output becomes part of the input for the next turn.
-	if result.FinalUsage != nil && result.FinalUsage.InputTokens > 0 {
-		if a.opts.Debug {
-			log.Printf("[DEBUG] updateUsageFromTurnResult: calling SetContextTokens=%d (FinalUsage.InputTokens)",
-				result.FinalUsage.InputTokens)
+	// Calculate context fill from the LAST API call's usage. The context
+	// window is filled by everything sent to and received from the model:
+	//
+	//   InputTokens       — non-cached input (may be small with prompt caching)
+	//   CacheReadTokens   — input tokens served from cache
+	//   CacheCreationTokens — input tokens written to cache this call
+	//   OutputTokens      — assistant output (becomes input next turn)
+	//
+	// With Anthropic prompt caching, InputTokens can drop to near-zero while
+	// CacheReadTokens holds the bulk of the context. We must sum all four to
+	// get the true context window utilization.
+	//
+	// We use FinalUsage (last step only), NOT TotalUsage, because TotalUsage
+	// sums across all tool-calling steps — and each step re-sends the full
+	// conversation, so TotalUsage massively overstates the actual window fill.
+	if result.FinalUsage != nil {
+		u := result.FinalUsage
+		contextFill := int(u.InputTokens) + int(u.CacheReadTokens) + int(u.CacheCreationTokens) + int(u.OutputTokens)
+		if contextFill > 0 {
+			if a.opts.Debug {
+				log.Printf("[DEBUG] updateUsageFromTurnResult: SetContextTokens=%d (Input=%d + CacheRead=%d + CacheCreate=%d + Output=%d)",
+					contextFill, u.InputTokens, u.CacheReadTokens, u.CacheCreationTokens, u.OutputTokens)
+			}
+			a.opts.UsageTracker.SetContextTokens(contextFill)
 		}
-		a.opts.UsageTracker.SetContextTokens(int(result.FinalUsage.InputTokens))
 	}
 }

internal/app/app_test.go

@@ -3,12 +3,12 @@ package app
 import (
 	"context"
 	"errors"
+	"strings"
 	"sync"
 	"testing"
 	"time"
 
-	"charm.land/fantasy"
-
+	tea "charm.land/bubbletea/v2"
 	kit "github.com/mark3labs/kit/pkg/kit"
 )
 
@@ -534,9 +534,9 @@ func TestQueueLength_reflects(t *testing.T) {
 }
 
 // TestRecordStepUsage_updatesTracker verifies that per-step usage updates are
-// recorded immediately for cost tracking. Context tokens are NOT updated here
-// (only via updateUsageFromTurnResult) to avoid display jumps during multi-step
-// tool calls.
+// recorded immediately for cost tracking. Context tokens are also updated so
+// the status bar reflects context fill after every LLM call in a multi-step
+// turn, not just at the end.
 func TestRecordStepUsage_updatesTracker(t *testing.T) {
 	usage := &usageUpdaterStub{}
 	app := New(Options{UsageTracker: usage}, nil)
@@ -547,7 +547,7 @@ func TestRecordStepUsage_updatesTracker(t *testing.T) {
 		OutputTokens:     45,
 		CacheReadTokens:  5,
 		CacheWriteTokens: 2,
-	}, nil)
+	}, nil, nil)
 
 	usage.mu.Lock()
 	defer usage.mu.Unlock()
@@ -559,9 +559,13 @@ func TestRecordStepUsage_updatesTracker(t *testing.T) {
 		t.Fatalf("unexpected usage update payload: in=%d out=%d cache_read=%d cache_write=%d",
 			usage.lastUpdateInput, usage.lastUpdateOutput, usage.lastUpdateCacheRead, usage.lastUpdateCacheWrite)
 	}
-	// Context tokens should NOT be updated by recordStepUsage (only by updateUsageFromTurnResult)
-	if usage.contextCalls != 0 {
-		t.Fatalf("expected 0 context token updates from recordStepUsage, got %d", usage.contextCalls)
+	// Context tokens should now be updated per-step (Input + CacheRead + CacheWrite + Output).
+	if usage.contextCalls != 1 {
+		t.Fatalf("expected 1 context token update from recordStepUsage, got %d", usage.contextCalls)
+	}
+	expectedContext := 120 + 45 + 5 + 2
+	if usage.lastContextTokens != expectedContext {
+		t.Fatalf("expected context tokens %d, got %d", expectedContext, usage.lastContextTokens)
 	}
 }
 
@@ -574,13 +578,13 @@ func TestUpdateUsageFromTurnResult_skipsTotalsWhenStepUsageSeen(t *testing.T) {
 
 	app.updateUsageFromTurnResult(&kit.TurnResult{
 		Response: "ok",
-		TotalUsage: &fantasy.Usage{
+		TotalUsage: &kit.LLMUsage{
 			InputTokens:         999,
 			OutputTokens:        111,
 			CacheReadTokens:     7,
 			CacheCreationTokens: 3,
 		},
-		FinalUsage: &fantasy.Usage{InputTokens: 456},
+		FinalUsage: &kit.LLMUsage{InputTokens: 456},
 	}, "prompt", true)
 
 	usage.mu.Lock()
@@ -608,13 +612,13 @@ func TestUpdateUsageFromTurnResult_recordsWhenInputTokensZero(t *testing.T) {
 	// Simulate OpenAI-compatible behavior: all prompt tokens cached, InputTokens=0
 	app.updateUsageFromTurnResult(&kit.TurnResult{
 		Response: "ok",
-		TotalUsage: &fantasy.Usage{
+		TotalUsage: &kit.LLMUsage{
 			InputTokens:         0,   // All cached - subtracted from prompt
 			OutputTokens:        150, // Actual generated tokens
 			CacheReadTokens:     500, // Cache hit
 			CacheCreationTokens: 0,
 		},
-		FinalUsage: &fantasy.Usage{InputTokens: 0, OutputTokens: 150},
+		FinalUsage: &kit.LLMUsage{InputTokens: 0, OutputTokens: 150},
 	}, "prompt", false)
 
 	usage.mu.Lock()
@@ -632,33 +636,336 @@ func TestUpdateUsageFromTurnResult_recordsWhenInputTokensZero(t *testing.T) {
 	}
 }
 
-// TestUpdateUsageFromTurnResult_contextTokensUsesInputOnly verifies that context
-// window fill uses InputTokens only (not input+output). The API's InputTokens
-// already includes the full conversation history; adding output would double-count.
-func TestUpdateUsageFromTurnResult_contextTokensUsesInputOnly(t *testing.T) {
+// TestUpdateUsageFromTurnResult_contextTokensUsesAllCategories verifies that
+// context window fill uses all token categories from the final API call:
+// InputTokens + CacheReadTokens + CacheCreationTokens + OutputTokens.
+// With Anthropic prompt caching, InputTokens can be near-zero while
+// CacheReadTokens holds the bulk of the context.
+func TestUpdateUsageFromTurnResult_contextTokensUsesAllCategories(t *testing.T) {
 	usage := &usageUpdaterStub{}
 	app := New(Options{UsageTracker: usage}, nil)
 	defer app.Close()
 
 	app.updateUsageFromTurnResult(&kit.TurnResult{
 		Response: "ok",
-		TotalUsage: &fantasy.Usage{
-			InputTokens:  1000,
-			OutputTokens: 200,
+		TotalUsage: &kit.LLMUsage{
+			InputTokens:         3,
+			OutputTokens:        5,
+			CacheReadTokens:     0,
+			CacheCreationTokens: 4317,
 		},
-		FinalUsage: &fantasy.Usage{
-			InputTokens:  1000, // Full context including history
-			OutputTokens: 200,
+		FinalUsage: &kit.LLMUsage{
+			InputTokens:         3,    // Non-cached input (small with caching)
+			OutputTokens:        5,    // Assistant output
+			CacheReadTokens:     0,    // No cache reads on first call
+			CacheCreationTokens: 4317, // System prompt + tools written to cache
 		},
 	}, "prompt", false)
 
 	usage.mu.Lock()
 	defer usage.mu.Unlock()
 
-	// Context tokens should be InputTokens only (1000), not input+output (1200)
-	// because InputTokens already includes the full conversation history
-	if usage.contextCalls != 1 || usage.lastContextTokens != 1000 {
-		t.Fatalf("expected context tokens=1000 (InputTokens only), got calls=%d tokens=%d",
-			usage.contextCalls, usage.lastContextTokens)
+	// Context tokens should be Input + CacheRead + CacheCreate + Output = 4325
+	expected := 3 + 0 + 4317 + 5
+	if usage.contextCalls != 1 || usage.lastContextTokens != expected {
+		t.Fatalf("expected context tokens=%d (all categories), got calls=%d tokens=%d",
+			expected, usage.contextCalls, usage.lastContextTokens)
+	}
+}
+
+// TestHandleTurnEnd_LengthEmitsWarning verifies that when the SDK reports a
+// FinishReasonLength (max_output_tokens hit), the app surfaces a user-visible
+// ExtensionPrintEvent with Level="info" so the TUI can render a banner
+// instead of silently showing a truncated reply.
+func TestHandleTurnEnd_LengthEmitsWarning(t *testing.T) {
+	app := New(Options{}, nil)
+	defer app.Close()
+
+	var mu sync.Mutex
+	var received []tea.Msg
+	sendFn := func(m tea.Msg) {
+		mu.Lock()
+		defer mu.Unlock()
+		received = append(received, m)
+	}
+
+	app.handleTurnEnd(kit.TurnEndEvent{StopReason: kit.FinishReasonLength}, sendFn)
+
+	mu.Lock()
+	defer mu.Unlock()
+	if len(received) != 1 {
+		t.Fatalf("expected 1 event on length stop, got %d", len(received))
+	}
+	ev, ok := received[0].(ExtensionPrintEvent)
+	if !ok {
+		t.Fatalf("expected ExtensionPrintEvent, got %T", received[0])
+	}
+	if ev.Level != "info" {
+		t.Errorf("expected Level=info, got %q", ev.Level)
+	}
+	if ev.Text == "" {
+		t.Error("expected non-empty warning text")
+	}
+	if !strings.Contains(ev.Text, "max_output_tokens") {
+		t.Errorf("warning text should mention max_output_tokens, got: %s", ev.Text)
+	}
+}
+
+// TestHandleTurnEnd_NonLengthIgnored verifies that ordinary stop reasons
+// (stop, tool-calls, error, unknown, "") do not produce a warning banner.
+func TestHandleTurnEnd_NonLengthIgnored(t *testing.T) {
+	app := New(Options{}, nil)
+	defer app.Close()
+
+	reasons := []string{
+		kit.FinishReasonStop,
+		kit.FinishReasonToolCalls,
+		kit.FinishReasonError,
+		kit.FinishReasonContentFilter,
+		kit.FinishReasonOther,
+		kit.FinishReasonUnknown,
+		"",
+	}
+	for _, r := range reasons {
+		var called bool
+		app.handleTurnEnd(kit.TurnEndEvent{StopReason: r}, func(m tea.Msg) {
+			called = true
+		})
+		if called {
+			t.Errorf("stop reason %q unexpectedly emitted a warning", r)
+		}
+	}
+}
+
+// TestHandleTurnEnd_NilSendFn guards against panics when no TUI listener is
+// attached (e.g. early init or headless teardown).
+func TestHandleTurnEnd_NilSendFn(t *testing.T) {
+	app := New(Options{}, nil)
+	defer app.Close()
+
+	// Should not panic with a nil sendFn.
+	app.handleTurnEnd(kit.TurnEndEvent{StopReason: kit.FinishReasonLength}, nil)
+}
+
+// TestFormatMaxTokensTruncatedMessage_NoKit verifies the fallback message
+// when Options.Kit is nil (test/stub path).
+func TestFormatMaxTokensTruncatedMessage_NoKit(t *testing.T) {
+	app := New(Options{}, nil)
+	defer app.Close()
+
+	msg := app.formatMaxTokensTruncatedMessage()
+	if msg == "" {
+		t.Fatal("expected non-empty fallback message")
+	}
+	for _, needle := range []string{"max_output_tokens", "--max-tokens", "KIT_MAX_TOKENS", "modelSettings"} {
+		if !strings.Contains(msg, needle) {
+			t.Errorf("fallback message missing %q:\n%s", needle, msg)
+		}
+	}
+}
+
+// --------------------------------------------------------------------------
+// releaseBusyAfterCompact (issue #27)
+// --------------------------------------------------------------------------
+
+// TestReleaseBusyAfterCompact_flushesQueuedMessages is a regression test for
+// issue #27: messages queued via Run() while /compact is running used to sit
+// in a.queue indefinitely until the user typed another prompt. After the fix
+// the deferred releaseBusyAfterCompact tail picks up any pending items and
+// dispatches drainQueue automatically.
+//
+// We simulate the compaction completion path directly (bypassing the SDK)
+// by toggling busy=true, populating the queue exactly as Run() would have
+// during compaction, and then invoking releaseBusyAfterCompact.
+func TestReleaseBusyAfterCompact_flushesQueuedMessages(t *testing.T) {
+	stub := newStubWithFuncs(
+		func(ctx context.Context) (*kit.TurnResult, error) {
+			return turnResult("compacted then drained"), nil
+		},
+	)
+	app := newTestApp(stub)
+	defer app.Close()
+
+	// Simulate the state at the start of the compaction tail: busy is set
+	// and a couple of prompts have piled up in the queue while we were
+	// summarising. (Run() would have appended them and returned a queue
+	// length > 0 to the caller.)
+	app.mu.Lock()
+	app.busy = true
+	app.queue = append(app.queue,
+		queueItem{Prompt: "queued during compact #1"},
+		queueItem{Prompt: "queued during compact #2"},
+	)
+	app.mu.Unlock()
+
+	// Invoke the deferred tail directly. It should kick off drainQueue.
+	app.releaseBusyAfterCompact()
+
+	// drainQueue runs in a goroutine. Wait for the app to come back to idle.
+	ok := waitForCondition(2*time.Second, func() bool {
+		app.mu.Lock()
+		defer app.mu.Unlock()
+		return !app.busy
+	})
+	if !ok {
+		t.Fatal("app did not become idle after releaseBusyAfterCompact: queue not drained")
+	}
+
+	// Wait for any in-flight goroutine to finish before reading state.
+	app.wg.Wait()
+
+	if got := app.QueueLength(); got != 0 {
+		t.Fatalf("expected empty queue after drain, got %d", got)
+	}
+	if n := stub.callCount(); n == 0 {
+		t.Fatalf("expected stub PromptFunc to fire at least once after compact, got %d calls", n)
+	}
+}
+
+// TestReleaseBusyAfterCompact_idleWhenQueueEmpty verifies that with no
+// pending messages the helper just clears busy and does NOT spawn a
+// drainQueue goroutine (no spurious agent turn).
+func TestReleaseBusyAfterCompact_idleWhenQueueEmpty(t *testing.T) {
+	stub := newStub()
+	app := newTestApp(stub)
+	defer app.Close()
+
+	app.mu.Lock()
+	app.busy = true
+	app.mu.Unlock()
+
+	app.releaseBusyAfterCompact()
+
+	app.mu.Lock()
+	busy := app.busy
+	app.mu.Unlock()
+	if busy {
+		t.Fatal("expected busy=false after releaseBusyAfterCompact with empty queue")
+	}
+
+	// Give any rogue goroutine a moment to (incorrectly) call PromptFunc.
+	time.Sleep(50 * time.Millisecond)
+	if n := stub.callCount(); n != 0 {
+		t.Fatalf("expected 0 PromptFunc calls when queue empty, got %d", n)
+	}
+}
+
+// TestReleaseBusyAfterCompact_splicesSteerAheadOfQueue exercises the SDK
+// steer-drain branch of releaseBusyAfterCompact (issue #27 follow-up).
+//
+// Production wires a.opts.Kit.DrainSteer() to pull messages that arrived via
+// Steer/SteerWithFiles during compaction, but Options.Kit is *kit.Kit (a
+// concrete struct) so unit tests cannot stand up a real instance without a
+// full LLM backend. The test uses the unexported steerDrainFn seam to inject
+// fake steer items, then asserts that:
+//
+//   - Steer items are dispatched ahead of any prompts that piled up in
+//     a.queue (steer retains "act now" priority over ordinary queued
+//     prompts), and
+//   - the helper still hands off to drainQueue so the steer item actually
+//     fires (the previous behaviour left them stranded — see #27).
+func TestReleaseBusyAfterCompact_splicesSteerAheadOfQueue(t *testing.T) {
+	var pmu sync.Mutex
+	var firstPrompt string
+	stub := newStubWithFuncs(
+		func(ctx context.Context) (*kit.TurnResult, error) {
+			return turnResult("steer dispatched"), nil
+		},
+	)
+	// Wrap PromptFunc so we can capture the prompt text the stub receives
+	// (newStubWithFuncs's fns ignore prompt; we need it to verify ordering).
+	capturingPrompt := func(ctx context.Context, prompt string) (*kit.TurnResult, error) {
+		pmu.Lock()
+		if firstPrompt == "" {
+			firstPrompt = prompt
+		}
+		pmu.Unlock()
+		return stub.fn(ctx, prompt)
+	}
+	app := New(Options{PromptFunc: capturingPrompt}, nil)
+	defer app.Close()
+
+	// Inject fake steer items via the test seam. In production the same
+	// items would have been delivered through Kit.InjectSteerWithFiles
+	// during /compact and pulled by DrainSteer here.
+	app.steerDrainFn = func() []queueItem {
+		return []queueItem{
+			{Prompt: "steer-1"},
+			{Prompt: "steer-2"},
+		}
+	}
+
+	// Simulate the state at the end of compaction: busy is set and a couple
+	// of regular Run() prompts have piled up after the steer messages.
+	app.mu.Lock()
+	app.busy = true
+	app.queue = append(app.queue,
+		queueItem{Prompt: "queued-1"},
+		queueItem{Prompt: "queued-2"},
+	)
+	app.mu.Unlock()
+
+	app.releaseBusyAfterCompact()
+
+	// Wait for the dispatched batch to complete.
+	ok := waitForCondition(2*time.Second, func() bool {
+		app.mu.Lock()
+		defer app.mu.Unlock()
+		return !app.busy
+	})
+	if !ok {
+		t.Fatal("app did not become idle after steer-spliced releaseBusyAfterCompact")
+	}
+	app.wg.Wait()
+
+	// drainQueue picks up `first` directly and batches the rest. With
+	// PromptFunc set, executeBatch invokes us with items[0] only — that
+	// item must be the first steer message, proving steer items were
+	// spliced ahead of the previously queued prompts.
+	pmu.Lock()
+	got := firstPrompt
+	pmu.Unlock()
+	if got != "steer-1" {
+		t.Fatalf("expected first dispatched prompt to be steer item %q (steer items must come before queued prompts), got %q",
+			"steer-1", got)
+	}
+
+	// Queue should be fully drained and PromptFunc must have actually fired.
+	if n := app.QueueLength(); n != 0 {
+		t.Fatalf("expected empty queue after drain, got %d entries", n)
+	}
+	if n := stub.callCount(); n == 0 {
+		t.Fatal("expected stub PromptFunc to fire at least once after splice")
+	}
+}
+
+// TestReleaseBusyAfterCompact_dropsQueueWhenClosed verifies that if the app
+// was closed during compaction the helper discards any pending items rather
+// than spawning drainQueue against a torn-down App.
+func TestReleaseBusyAfterCompact_dropsQueueWhenClosed(t *testing.T) {
+	stub := newStub()
+	app := newTestApp(stub)
+
+	app.mu.Lock()
+	app.busy = true
+	app.queue = append(app.queue, queueItem{Prompt: "would have run"})
+	app.closed = true
+	app.mu.Unlock()
+
+	app.releaseBusyAfterCompact()
+
+	app.mu.Lock()
+	busy := app.busy
+	qLen := len(app.queue)
+	app.mu.Unlock()
+	if busy {
+		t.Fatal("expected busy=false even when closed")
+	}
+	if qLen != 0 {
+		t.Fatalf("expected queue cleared on closed app, got %d entries", qLen)
+	}
+	time.Sleep(20 * time.Millisecond)
+	if n := stub.callCount(); n != 0 {
+		t.Fatalf("expected 0 PromptFunc calls on closed app, got %d", n)
 	}
 }

internal/app/events.go

@@ -1,6 +1,6 @@
 package app
 
-import "charm.land/fantasy"
+import kit "github.com/mark3labs/kit/pkg/kit"
 
 // StreamChunkEvent is sent by the app layer when a streaming text delta arrives
 // from the LLM. Each chunk contains an incremental portion of the response.
@@ -16,6 +16,11 @@ type ReasoningChunkEvent struct {
 	Delta string
 }
 
+// ReasoningCompleteEvent is sent when reasoning/thinking is finished, after
+// the last reasoning token has been processed. The TUI uses this to freeze
+// the reasoning duration counter.
+type ReasoningCompleteEvent struct{}
+
 // ToolCallStartedEvent is sent when a tool call has been parsed and is about to execute.
 // It carries the tool name and its arguments for display purposes.
 type ToolCallStartedEvent struct {
@@ -27,6 +32,36 @@ type ToolCallStartedEvent struct {
 	ToolArgs string
 }
 
+// ToolCallInputStartEvent is sent when the LLM begins generating tool call
+// arguments. The tool name is known but the full argument JSON is still being
+// streamed. UIs can use this to show a "running" indicator immediately instead
+// of waiting for the full argument JSON to finish streaming.
+type ToolCallInputStartEvent struct {
+	// ToolCallID is the stable identifier for correlating tool lifecycle events.
+	ToolCallID string
+	// ToolName is the name of the tool being called.
+	ToolName string
+	// ToolKind classifies the tool: "execute", "edit", "read", "search", "agent".
+	ToolKind string
+}
+
+// ToolCallInputDeltaEvent is sent for each streamed fragment of tool call
+// arguments as they arrive from the LLM. Useful for live-previewing content
+// or showing a progress indicator with byte count.
+type ToolCallInputDeltaEvent struct {
+	// ToolCallID is the stable identifier for correlating tool lifecycle events.
+	ToolCallID string
+	// Delta is a JSON fragment of tool call arguments.
+	Delta string
+}
+
+// ToolCallInputEndEvent is sent when tool argument streaming is complete,
+// before the tool call is parsed and execution begins.
+type ToolCallInputEndEvent struct {
+	// ToolCallID is the stable identifier for correlating tool lifecycle events.
+	ToolCallID string
+}
+
 // ToolExecutionEvent is sent when a tool starts or finishes executing.
 // The IsStarting flag distinguishes between the start and end of execution.
 type ToolExecutionEvent struct {
@@ -74,6 +109,24 @@ type ToolCallContentEvent struct {
 	Content string
 }
 
+// PasswordPromptEvent is sent when a sudo command needs a password.
+// The TUI should display a password prompt overlay and send the result back.
+type PasswordPromptEvent struct {
+	// Prompt is the message to display to the user.
+	Prompt string
+	// ResponseCh receives the password from the TUI.
+	// The TUI must send exactly one value.
+	ResponseCh chan<- PasswordPromptResponse
+}
+
+// PasswordPromptResponse carries the user's password input.
+type PasswordPromptResponse struct {
+	// Password is the entered password.
+	Password string
+	// Cancelled is true if the user cancelled the prompt.
+	Cancelled bool
+}
+
 // ResponseCompleteEvent is sent when the LLM produces a final (non-streaming) response.
 // In streaming mode, this may be empty if all content was delivered via StreamChunkEvents.
 type ResponseCompleteEvent struct {
@@ -118,8 +171,8 @@ type SpinnerEvent struct {
 // MessageCreatedEvent is sent when a new message is added to the message store.
 // This allows the TUI to stay in sync with the conversation history.
 type MessageCreatedEvent struct {
-	// Message is the fantasy message that was added to the store.
-	Message fantasy.Message
+	// Message is the message that was added to the store.
+	Message kit.LLMMessage
 }
 
 // CompactCompleteEvent is sent when a /compact operation finishes successfully.
@@ -157,11 +210,36 @@ type ModelChangedEvent struct {
 	ModelName string
 }
 
+// UsageUpdatedEvent is sent after each completed LLM step to notify the TUI
+// that token counts and costs have changed. The UsageTracker is updated
+// in-place before this event is sent; the TUI just needs to re-render to
+// reflect the new values in the status bar.
+type UsageUpdatedEvent struct{}
+
 // WidgetUpdateEvent is sent when an extension adds, updates, or removes a
 // widget via ctx.SetWidget or ctx.RemoveWidget. The TUI re-reads widget state
 // from its WidgetProvider on the next render cycle.
 type WidgetUpdateEvent struct{}
 
+// ContentReloadEvent is sent when prompt templates or skills are reloaded
+// from disk (e.g. by a file watcher detecting changes). The TUI refreshes
+// its autocomplete entries and internal state from the provider callbacks.
+type ContentReloadEvent struct{}
+
+// MCPToolsReadyEvent is sent when background MCP tool loading completes.
+// The TUI refreshes its tool names and MCP tool count from provider callbacks
+// so that /tools and the startup info bar reflect the loaded MCP tools.
+type MCPToolsReadyEvent struct{}
+
+// MCPServerLoadedEvent is sent when a single MCP server finishes loading
+// (successfully or with error). The TUI displays a system message so users
+// see real-time progress as each server initializes.
+type MCPServerLoadedEvent struct {
+	ServerName string
+	ToolCount  int
+	Error      error // nil on success
+}
+
 // EditorTextSetEvent is sent when an extension calls ctx.SetEditorText to
 // pre-fill the input editor with text. The TUI handles this by setting the
 // textarea content and moving the cursor to the end.

internal/app/messages.go

@@ -3,14 +3,14 @@ package app
 import (
 	"sync"
 
-	"charm.land/fantasy"
+	kit "github.com/mark3labs/kit/pkg/kit"
 )
 
 // MessageStore is a thread-safe in-memory store for the conversation history.
 // On-disk persistence is handled by the TreeManager at the app/SDK layer.
 type MessageStore struct {
 	mu       sync.RWMutex
-	messages []fantasy.Message
+	messages []kit.LLMMessage
 }
 
 // NewMessageStore creates an empty MessageStore.
@@ -20,14 +20,14 @@ func NewMessageStore() *MessageStore {
 
 // NewMessageStoreWithMessages creates a MessageStore pre-populated with the
 // given messages. This is used when loading an existing session at startup.
-func NewMessageStoreWithMessages(msgs []fantasy.Message) *MessageStore {
-	cp := make([]fantasy.Message, len(msgs))
+func NewMessageStoreWithMessages(msgs []kit.LLMMessage) *MessageStore {
+	cp := make([]kit.LLMMessage, len(msgs))
 	copy(cp, msgs)
 	return &MessageStore{messages: cp}
 }
 
 // Add appends a single message to the store.
-func (s *MessageStore) Add(msg fantasy.Message) {
+func (s *MessageStore) Add(msg kit.LLMMessage) {
 	s.mu.Lock()
 	defer s.mu.Unlock()
 	s.messages = append(s.messages, msg)
@@ -36,22 +36,22 @@ func (s *MessageStore) Add(msg fantasy.Message) {
 // Replace replaces the entire message history with the given slice. This is
 // used after an agent step returns the full updated conversation (including
 // tool calls and results).
-func (s *MessageStore) Replace(msgs []fantasy.Message) {
+func (s *MessageStore) Replace(msgs []kit.LLMMessage) {
 	s.mu.Lock()
 	defer s.mu.Unlock()
 
-	cp := make([]fantasy.Message, len(msgs))
+	cp := make([]kit.LLMMessage, len(msgs))
 	copy(cp, msgs)
 	s.messages = cp
 }
 
 // GetAll returns a snapshot copy of the current message slice.
 // The returned slice is safe to modify without affecting the store.
-func (s *MessageStore) GetAll() []fantasy.Message {
+func (s *MessageStore) GetAll() []kit.LLMMessage {
 	s.mu.RLock()
 	defer s.mu.RUnlock()
 
-	cp := make([]fantasy.Message, len(s.messages))
+	cp := make([]kit.LLMMessage, len(s.messages))
 	copy(cp, s.messages)
 	return cp
 }

internal/app/messages_test.go

@@ -3,17 +3,27 @@ package app
 import (
 	"testing"
 
-	"charm.land/fantasy"
+	kit "github.com/mark3labs/kit/pkg/kit"
 )
 
-// makeTextMsg builds a minimal fantasy.Message with a single TextPart.
-func makeTextMsg(role, text string) fantasy.Message {
-	return fantasy.Message{
-		Role:    fantasy.MessageRole(role),
-		Content: []fantasy.MessagePart{fantasy.TextPart{Text: text}},
+// makeTextMsg builds a minimal kit.LLMMessage with the given role and text.
+func makeTextMsg(role, text string) kit.LLMMessage {
+	return kit.LLMMessage{
+		Role:    kit.LLMMessageRole(role),
+		Content: []kit.LLMMessagePart{kit.LLMTextPart{Text: text}},
 	}
 }
 
+// textOf extracts the plain text from an LLMMessage for assertions.
+func textOf(msg kit.LLMMessage) string {
+	for _, part := range msg.Content {
+		if tp, ok := part.(kit.LLMTextPart); ok {
+			return tp.Text
+		}
+	}
+	return ""
+}
+
 // --------------------------------------------------------------------------
 // NewMessageStore / NewMessageStoreWithMessages
 // --------------------------------------------------------------------------
@@ -29,7 +39,7 @@ func TestNewMessageStore_empty(t *testing.T) {
 }
 
 func TestNewMessageStoreWithMessages_preloaded(t *testing.T) {
-	msgs := []fantasy.Message{
+	msgs := []kit.LLMMessage{
 		makeTextMsg("user", "hello"),
 		makeTextMsg("assistant", "hi"),
 	}
@@ -42,7 +52,7 @@ func TestNewMessageStoreWithMessages_preloaded(t *testing.T) {
 // NewMessageStoreWithMessages must deep-copy the slice so that external
 // modifications don't affect the store.
 func TestNewMessageStoreWithMessages_isolatesInput(t *testing.T) {
-	msgs := []fantasy.Message{makeTextMsg("user", "hello")}
+	msgs := []kit.LLMMessage{makeTextMsg("user", "hello")}
 	s := NewMessageStoreWithMessages(msgs)
 
 	// Mutate the source slice.
@@ -52,9 +62,8 @@ func TestNewMessageStoreWithMessages_isolatesInput(t *testing.T) {
 	if len(got) != 1 {
 		t.Fatalf("expected 1 message, got %d", len(got))
 	}
-	tp, ok := got[0].Content[0].(fantasy.TextPart)
-	if !ok || tp.Text != "hello" {
-		t.Fatalf("store was mutated by external slice change; got %q", tp.Text)
+	if textOf(got[0]) != "hello" {
+		t.Fatalf("store was mutated by external slice change; got %q", textOf(got[0]))
 	}
 }
 
@@ -80,9 +89,8 @@ func TestAdd_preservesOrder(t *testing.T) {
 	}
 	got := s.GetAll()
 	for i, expected := range texts {
-		tp, ok := got[i].Content[0].(fantasy.TextPart)
-		if !ok || tp.Text != expected {
-			t.Fatalf("message[%d]: expected %q, got %q", i, expected, tp.Text)
+		if textOf(got[i]) != expected {
+			t.Fatalf("message[%d]: expected %q, got %q", i, expected, textOf(got[i]))
 		}
 	}
 }
@@ -95,7 +103,7 @@ func TestReplace_swapsHistory(t *testing.T) {
 	s := NewMessageStore()
 	s.Add(makeTextMsg("user", "old"))
 
-	replacement := []fantasy.Message{
+	replacement := []kit.LLMMessage{
 		makeTextMsg("user", "new1"),
 		makeTextMsg("assistant", "new2"),
 	}
@@ -105,25 +113,22 @@ func TestReplace_swapsHistory(t *testing.T) {
 		t.Fatalf("expected 2 messages after replace, got %d", s.Len())
 	}
 	got := s.GetAll()
-	tp0, _ := got[0].Content[0].(fantasy.TextPart)
-	tp1, _ := got[1].Content[0].(fantasy.TextPart)
-	if tp0.Text != "new1" || tp1.Text != "new2" {
-		t.Fatalf("unexpected messages after replace: %q %q", tp0.Text, tp1.Text)
+	if textOf(got[0]) != "new1" || textOf(got[1]) != "new2" {
+		t.Fatalf("unexpected messages after replace: %q %q", textOf(got[0]), textOf(got[1]))
 	}
 }
 
 // Replace must deep-copy the incoming slice.
 func TestReplace_isolatesInput(t *testing.T) {
 	s := NewMessageStore()
-	replacement := []fantasy.Message{makeTextMsg("user", "original")}
+	replacement := []kit.LLMMessage{makeTextMsg("user", "original")}
 	s.Replace(replacement)
 
 	replacement[0] = makeTextMsg("user", "mutated")
 
 	got := s.GetAll()
-	tp, _ := got[0].Content[0].(fantasy.TextPart)
-	if tp.Text != "original" {
-		t.Fatalf("store was mutated by external slice change after Replace; got %q", tp.Text)
+	if textOf(got[0]) != "original" {
+		t.Fatalf("store was mutated by external slice change after Replace; got %q", textOf(got[0]))
 	}
 }
 
@@ -140,9 +145,8 @@ func TestGetAll_returnsCopy(t *testing.T) {
 	got[0] = makeTextMsg("user", "mutated")
 
 	internal := s.GetAll()
-	tp, _ := internal[0].Content[0].(fantasy.TextPart)
-	if tp.Text != "hello" {
-		t.Fatalf("GetAll returned non-copy; store was mutated to %q", tp.Text)
+	if textOf(internal[0]) != "hello" {
+		t.Fatalf("GetAll returned non-copy; store was mutated to %q", textOf(internal[0]))
 	}
 }
 
@@ -179,9 +183,8 @@ func TestClear_allowsSubsequentAdds(t *testing.T) {
 		t.Fatalf("expected 1 message after Clear+Add, got %d", s.Len())
 	}
 	got := s.GetAll()
-	tp, _ := got[0].Content[0].(fantasy.TextPart)
-	if tp.Text != "after" {
-		t.Fatalf("expected %q, got %q", "after", tp.Text)
+	if textOf(got[0]) != "after" {
+		t.Fatalf("expected %q, got %q", "after", textOf(got[0]))
 	}
 }
 

internal/app/options.go

@@ -21,8 +21,10 @@ type UsageUpdater interface {
 	// the provider does not return exact counts.
 	EstimateAndUpdateUsage(inputText, outputText string)
 	// SetContextTokens records the approximate current context window fill
-	// level. This should be the final API call's input+output tokens (from
-	// FinalResponse.Usage), NOT the aggregate TotalUsage.
+	// level. This should be the sum of ALL token categories from the last
+	// API call: InputTokens + CacheReadTokens + CacheCreationTokens +
+	// OutputTokens. With Anthropic prompt caching, InputTokens can be
+	// near-zero while CacheReadTokens holds the bulk of the context.
 	SetContextTokens(tokens int)
 }
 
@@ -67,10 +69,6 @@ type Options struct {
 	// Debug enables verbose debug logging.
 	Debug bool
 
-	// CompactMode selects the compact renderer instead of the block renderer for
-	// message formatting.
-	CompactMode bool
-
 	// UsageTracker is an optional callback for recording token usage after each
 	// agent step. When non-nil, the app layer calls UpdateUsage (or
 	// EstimateAndUpdateUsage as a fallback) using the usage data returned by the

internal/auth/credentials.go

@@ -255,29 +255,6 @@ func (cm *CredentialManager) HasAnthropicCredentials() (bool, error) {
 	}
 }
 
-// SetOpenAICredentials stores OpenAI API key credentials. It validates the
-// API key format before storing. The API key must start with "sk-" and be
-// at least 20 characters long. Returns an error if the API key is invalid or
-// if storage fails.
-func (cm *CredentialManager) SetOpenAICredentials(apiKey string) error {
-	if err := validateOpenAIAPIKey(apiKey); err != nil {
-		return err
-	}
-
-	store, err := cm.LoadCredentials()
-	if err != nil {
-		return err
-	}
-
-	store.OpenAI = &OpenAICredentials{
-		Type:      "api_key",
-		APIKey:    apiKey,
-		CreatedAt: time.Now(),
-	}
-
-	return cm.SaveCredentials(store)
-}
-
 // GetOpenAICredentials retrieves stored OpenAI credentials. Returns nil if
 // no credentials are stored. The returned credentials may be either OAuth or API
 // key type, check the Type field to determine which.
@@ -417,26 +394,6 @@ func validateAnthropicAPIKey(apiKey string) error {
 	return nil
 }
 
-// validateOpenAIAPIKey validates the format of an OpenAI API key
-func validateOpenAIAPIKey(apiKey string) error {
-	apiKey = strings.TrimSpace(apiKey)
-
-	if apiKey == "" {
-		return fmt.Errorf("API key cannot be empty")
-	}
-
-	// OpenAI API keys typically start with "sk-" and are quite long
-	if !strings.HasPrefix(apiKey, "sk-") {
-		return fmt.Errorf("invalid OpenAI API key format (should start with 'sk-')")
-	}
-
-	if len(apiKey) < 20 {
-		return fmt.Errorf("API key appears to be too short")
-	}
-
-	return nil
-}
-
 // GetAnthropicAPIKey retrieves an Anthropic API key from multiple sources in priority order:
 // 1. Command-line flag value (highest priority)
 // 2. Stored credentials (OAuth or API key)
@@ -471,5 +428,13 @@ func GetAnthropicAPIKey(flagValue string) (string, string, error) {
 		return envKey, "ANTHROPIC_API_KEY environment variable", nil
 	}
 
+	// Check if OpenAI credentials exist to provide a helpful suggestion
+	if cm != nil {
+		hasOpenAI, _ := cm.HasOpenAICredentials()
+		if hasOpenAI {
+			return "", "", fmt.Errorf("no Anthropic API key found. Use 'kit auth login anthropic', set ANTHROPIC_API_KEY environment variable, or use --provider-api-key flag\n\nNote: OpenAI credentials were detected. To use OpenAI, run with --model openai/gpt-5.4 or set it as default:\n  kit auth login openai --set-default")
+		}
+	}
+
 	return "", "", fmt.Errorf("no Anthropic API key found. Use 'kit auth login anthropic', set ANTHROPIC_API_KEY environment variable, or use --provider-api-key flag")
 }

internal/compaction/compaction.go

@@ -428,6 +428,10 @@ type PreviousCompaction struct {
 	ModifiedFiles []string
 }
 
+// StreamCallback is called for each chunk of text during streaming compaction.
+// Return a non-nil error to cancel the stream.
+type StreamCallback func(delta string) error
+
 // Compact summarises older messages using the LLM, returning the compaction
 // result and a new message slice (summary message + preserved recent
 // messages).
@@ -442,6 +446,8 @@ type PreviousCompaction struct {
 //
 // prev carries file tracking from a previous compaction for cumulative
 // tracking. Pass nil if there is no prior compaction.
+// onChunk is an optional callback for streaming summary text. Pass nil for
+// non-streaming compaction.
 func Compact(
 	ctx context.Context,
 	model fantasy.LanguageModel,
@@ -449,6 +455,7 @@ func Compact(
 	opts CompactionOptions,
 	customInstructions string,
 	prev *PreviousCompaction,
+	onChunk StreamCallback,
 ) (*CompactionResult, []fantasy.Message, error) {
 	opts.defaults()
 
@@ -487,9 +494,9 @@ func Compact(
 	var err error
 
 	if IsSplitTurn(messages, cutPoint) {
-		summaryText, err = compactSplitTurn(ctx, model, oldMessages, messages, cutPoint, opts, customInstructions)
+		summaryText, err = compactSplitTurn(ctx, model, oldMessages, messages, cutPoint, opts, customInstructions, onChunk)
 	} else {
-		summaryText, err = compactNormal(ctx, model, oldMessages, opts, customInstructions)
+		summaryText, err = compactNormal(ctx, model, oldMessages, opts, customInstructions, onChunk)
 	}
 	if err != nil {
 		return nil, nil, err
@@ -527,15 +534,17 @@ func Compact(
 }
 
 // compactNormal generates a summary for a clean turn-boundary cut.
+// If onChunk is provided, text deltas are streamed to it.
 func compactNormal(
 	ctx context.Context,
 	model fantasy.LanguageModel,
 	oldMessages []fantasy.Message,
 	opts CompactionOptions,
 	customInstructions string,
+	onChunk StreamCallback,
 ) (string, error) {
 	conversationText := serializeMessages(oldMessages)
-	return generateSummary(ctx, model, conversationText, opts, customInstructions)
+	return generateSummary(ctx, model, conversationText, opts, customInstructions, onChunk)
 }
 
 // compactSplitTurn handles the case where the cut point lands mid-turn.
@@ -546,6 +555,7 @@ func compactNormal(
 //
 // The merged result preserves context from both the older history and the
 // beginning of the current long turn.
+// If onChunk is provided, both summaries and the separator are streamed.
 func compactSplitTurn(
 	ctx context.Context,
 	model fantasy.LanguageModel,
@@ -554,6 +564,7 @@ func compactSplitTurn(
 	cutPoint int,
 	opts CompactionOptions,
 	customInstructions string,
+	onChunk StreamCallback,
 ) (string, error) {
 	// Find where the split turn starts.
 	turnStart := findTurnStart(allMessages, cutPoint)
@@ -573,12 +584,19 @@ func compactSplitTurn(
 	// Generate history summary if there are complete turns before the split.
 	if len(historyMessages) >= 2 {
 		historySummary, err = generateSummary(ctx, model,
-			serializeMessages(historyMessages), opts, "")
+			serializeMessages(historyMessages), opts, "", onChunk)
 		if err != nil {
 			return "", fmt.Errorf("split turn history summary failed: %w", err)
 		}
 	}
 
+	// Stream the separator between history and turn prefix summaries.
+	if onChunk != nil && historySummary != "" {
+		if err := onChunk("\n\n---\n\n## Current Turn (in progress)\n\n"); err != nil {
+			return "", fmt.Errorf("streaming separator failed: %w", err)
+		}
+	}
+
 	// Generate turn prefix summary.
 	turnPrefixText := serializeMessages(turnPrefixMessages)
 	turnPrefixPrompt := "The messages above are the BEGINNING of a long turn that was split. " +
@@ -588,16 +606,10 @@ func compactSplitTurn(
 		turnPrefixPrompt += "\n\nAdditional instructions: " + customInstructions
 	}
 
-	summaryAgent := fantasy.NewAgent(model,
-		fantasy.WithSystemPrompt(defaultSystemPrompt),
-	)
-	result, err := summaryAgent.Generate(ctx, fantasy.AgentCall{
-		Prompt: turnPrefixText + "\n\n" + turnPrefixPrompt,
-	})
+	turnPrefixSummary, err := generateSummary(ctx, model, turnPrefixText, opts, turnPrefixPrompt, onChunk)
 	if err != nil {
 		return "", fmt.Errorf("split turn prefix summary failed: %w", err)
 	}
-	turnPrefixSummary := result.Response.Content.Text()
 
 	// Merge the two summaries.
 	if historySummary != "" && turnPrefixSummary != "" {
@@ -610,12 +622,14 @@ func compactSplitTurn(
 }
 
 // generateSummary calls the LLM to produce a structured summary.
+// If onChunk is provided, the summary is streamed using Agent.Stream().
 func generateSummary(
 	ctx context.Context,
 	model fantasy.LanguageModel,
 	conversationText string,
 	opts CompactionOptions,
 	customInstructions string,
+	onChunk StreamCallback,
 ) (string, error) {
 	userPrompt := opts.SummaryPrompt
 	if userPrompt == "" {
@@ -628,8 +642,31 @@ func generateSummary(
 	summaryAgent := fantasy.NewAgent(model,
 		fantasy.WithSystemPrompt(defaultSystemPrompt),
 	)
+
+	prompt := conversationText + "\n\n" + userPrompt
+
+	// Use streaming if onChunk is provided.
+	if onChunk != nil {
+		var fullText strings.Builder
+		_, err := summaryAgent.Stream(ctx, fantasy.AgentStreamCall{
+			Prompt: prompt,
+			OnTextDelta: func(_, delta string) error {
+				if delta != "" {
+					fullText.WriteString(delta)
+					return onChunk(delta)
+				}
+				return nil
+			},
+		})
+		if err != nil {
+			return "", fmt.Errorf("compaction summarisation (streaming) failed: %w", err)
+		}
+		return fullText.String(), nil
+	}
+
+	// Non-streaming path.
 	result, err := summaryAgent.Generate(ctx, fantasy.AgentCall{
-		Prompt: conversationText + "\n\n" + userPrompt,
+		Prompt: prompt,
 	})
 	if err != nil {
 		return "", fmt.Errorf("compaction summarisation failed: %w", err)

internal/compaction/compaction_test.go

@@ -243,7 +243,7 @@ func TestCompact_TooFewMessages(t *testing.T) {
 		makeTextMessageN(fantasy.MessageRoleUser, 400),
 	}
 
-	result, newMsgs, err := Compact(context.TODO(), nil, msgs, CompactionOptions{}, "", nil)
+	result, newMsgs, err := Compact(context.TODO(), nil, msgs, CompactionOptions{}, "", nil, nil)
 	if err != nil {
 		t.Fatalf("unexpected error: %v", err)
 	}
@@ -262,7 +262,7 @@ func TestCompact_WithinBudget(t *testing.T) {
 		makeTextMessageN(fantasy.MessageRoleAssistant, 400),
 	}
 
-	result, newMsgs, err := Compact(context.TODO(), nil, msgs, CompactionOptions{}, "", nil)
+	result, newMsgs, err := Compact(context.TODO(), nil, msgs, CompactionOptions{}, "", nil, nil)
 	if err != nil {
 		t.Fatalf("unexpected error: %v", err)
 	}

internal/config/config.go

@@ -22,6 +22,45 @@ type MCPServerConfig struct {
 	AllowedTools  []string          `json:"allowedTools,omitempty" yaml:"allowedTools,omitempty"`
 	ExcludedTools []string          `json:"excludedTools,omitempty" yaml:"excludedTools,omitempty"`
 
+	// OAuth configuration for remote servers that don't support dynamic
+	// client registration (e.g. GitHub). When OAuthClientID is set, it is
+	// passed directly to the transport's OAuthConfig instead of relying on
+	// dynamic registration.
+	OAuthClientID     string   `json:"oauthClientId,omitempty" yaml:"oauthClientId,omitempty"`
+	OAuthClientSecret string   `json:"oauthClientSecret,omitempty" yaml:"oauthClientSecret,omitempty"`
+	OAuthScopes       []string `json:"oauthScopes,omitempty" yaml:"oauthScopes,omitempty"`
+
+	// NoOAuth disables OAuth transport configuration for this server, even
+	// when the connection pool has an auth handler. Use this for public MCP
+	// servers (e.g. PubMed) that don't require authentication. Without this
+	// flag, the pool would attach OAuth transport to every remote server,
+	// causing proactive dynamic-client-registration attempts that fail on
+	// servers that don't support it.
+	NoOAuth bool `json:"noOAuth,omitempty" yaml:"noOAuth,omitempty"`
+
+	// TasksMode controls when this server's tools/call requests are augmented
+	// with MCP task metadata (turning a synchronous call into an asynchronous,
+	// pollable job — see https://modelcontextprotocol.io/specification/2025-11-25/basic/utilities/tasks).
+	//
+	// Valid values:
+	//   - "" or "auto": (default) augment requests with task metadata only
+	//     when the server advertises tasks/toolCalls capability during initialize.
+	//   - "never":      never augment — every tool call is synchronous, regardless
+	//     of server capability.
+	//   - "always":     always augment, even when the server didn't advertise
+	//     task support. The server may still respond synchronously; this just
+	//     opts in unconditionally on the client side.
+	//
+	// In all modes, when the server returns a CreateTaskResult the client polls
+	// tasks/get / tasks/result until the task reaches a terminal state.
+	TasksMode string `json:"tasksMode,omitempty" yaml:"tasksMode,omitempty"`
+
+	// InProcessServer holds a live *server.MCPServer for in-process transport.
+	// When set (and Type is "inprocess"), the connection pool creates an
+	// in-process client instead of spawning a subprocess or making HTTP calls.
+	// This field is never serialized — it is only used programmatically via the SDK.
+	InProcessServer any `json:"-" yaml:"-"`
+
 	// Legacy fields for backward compatibility
 	Transport string         `json:"transport,omitempty"`
 	Args      []string       `json:"args,omitempty"`
@@ -35,13 +74,18 @@ type MCPServerConfig struct {
 func (s *MCPServerConfig) UnmarshalJSON(data []byte) error {
 	// First try to unmarshal as the new format
 	type newFormat struct {
-		Type          string            `json:"type"`
-		Command       []string          `json:"command,omitempty"`
-		Environment   map[string]string `json:"environment,omitempty"`
-		URL           string            `json:"url,omitempty"`
-		Headers       []string          `json:"headers,omitempty"`
-		AllowedTools  []string          `json:"allowedTools,omitempty" yaml:"allowedTools,omitempty"`
-		ExcludedTools []string          `json:"excludedTools,omitempty" yaml:"excludedTools,omitempty"`
+		Type              string            `json:"type"`
+		Command           []string          `json:"command,omitempty"`
+		Environment       map[string]string `json:"environment,omitempty"`
+		URL               string            `json:"url,omitempty"`
+		Headers           []string          `json:"headers,omitempty"`
+		AllowedTools      []string          `json:"allowedTools,omitempty" yaml:"allowedTools,omitempty"`
+		ExcludedTools     []string          `json:"excludedTools,omitempty" yaml:"excludedTools,omitempty"`
+		OAuthClientID     string            `json:"oauthClientId,omitempty" yaml:"oauthClientId,omitempty"`
+		OAuthClientSecret string            `json:"oauthClientSecret,omitempty" yaml:"oauthClientSecret,omitempty"`
+		OAuthScopes       []string          `json:"oauthScopes,omitempty" yaml:"oauthScopes,omitempty"`
+		NoOAuth           bool              `json:"noOAuth,omitempty" yaml:"noOAuth,omitempty"`
+		TasksMode         string            `json:"tasksMode,omitempty" yaml:"tasksMode,omitempty"`
 	}
 
 	// Also try legacy format
@@ -54,6 +98,7 @@ func (s *MCPServerConfig) UnmarshalJSON(data []byte) error {
 		Headers       []string       `json:"headers,omitempty"`
 		AllowedTools  []string       `json:"allowedTools,omitempty" yaml:"allowedTools,omitempty"`
 		ExcludedTools []string       `json:"excludedTools,omitempty" yaml:"excludedTools,omitempty"`
+		TasksMode     string         `json:"tasksMode,omitempty" yaml:"tasksMode,omitempty"`
 	}
 
 	// Try new format first
@@ -66,6 +111,11 @@ func (s *MCPServerConfig) UnmarshalJSON(data []byte) error {
 		s.Headers = newConfig.Headers
 		s.AllowedTools = newConfig.AllowedTools
 		s.ExcludedTools = newConfig.ExcludedTools
+		s.OAuthClientID = newConfig.OAuthClientID
+		s.OAuthClientSecret = newConfig.OAuthClientSecret
+		s.OAuthScopes = newConfig.OAuthScopes
+		s.NoOAuth = newConfig.NoOAuth
+		s.TasksMode = newConfig.TasksMode
 		return nil
 	}
 
@@ -86,6 +136,7 @@ func (s *MCPServerConfig) UnmarshalJSON(data []byte) error {
 	s.Headers = legacyConfig.Headers
 	s.AllowedTools = legacyConfig.AllowedTools
 	s.ExcludedTools = legacyConfig.ExcludedTools
+	s.TasksMode = legacyConfig.TasksMode
 
 	// Infer type from legacy format for better compatibility
 	// Only set Type when it doesn't change existing transport behavior
@@ -157,11 +208,28 @@ type Theme struct {
 	Markdown MarkdownThemeConfig `json:"markdown,omitzero" yaml:"markdown,omitempty"`
 }
 
+// GenerationParams defines generation parameter defaults that can be attached
+// to individual models. These act as model-level defaults — CLI flags and
+// global config values take precedence when explicitly set.
+type GenerationParams struct {
+	MaxTokens        *int     `json:"maxTokens,omitempty" yaml:"maxTokens,omitempty"`
+	Temperature      *float32 `json:"temperature,omitempty" yaml:"temperature,omitempty"`
+	TopP             *float32 `json:"topP,omitempty" yaml:"topP,omitempty"`
+	TopK             *int32   `json:"topK,omitempty" yaml:"topK,omitempty"`
+	FrequencyPenalty *float32 `json:"frequencyPenalty,omitempty" yaml:"frequencyPenalty,omitempty"`
+	PresencePenalty  *float32 `json:"presencePenalty,omitempty" yaml:"presencePenalty,omitempty"`
+	StopSequences    []string `json:"stopSequences,omitempty" yaml:"stopSequences,omitempty"`
+	ThinkingLevel    string   `json:"thinkingLevel,omitempty" yaml:"thinkingLevel,omitempty"`
+	SystemPrompt     string   `json:"systemPrompt,omitempty" yaml:"systemPrompt,omitempty"`
+}
+
 // CustomModelConfig defines a custom model that can be used with custom/custom
 // or other custom/ prefixed models. These models are loaded from the config file
 // and merged into the custom provider in the model registry.
 type CustomModelConfig struct {
 	Name        string      `json:"name" yaml:"name"`
+	BaseURL     string      `json:"baseUrl,omitempty" yaml:"baseUrl,omitempty"`
+	APIKey      string      `json:"apiKey,omitempty" yaml:"apiKey,omitempty"`
 	Family      string      `json:"family,omitempty" yaml:"family,omitempty"`
 	Attachment  bool        `json:"attachment,omitempty" yaml:"attachment,omitempty"`
 	Reasoning   bool        `json:"reasoning,omitempty" yaml:"reasoning,omitempty"`
@@ -169,6 +237,11 @@ type CustomModelConfig struct {
 	Knowledge   string      `json:"knowledge,omitempty" yaml:"knowledge,omitempty"`
 	Cost        CostConfig  `json:"cost" yaml:"cost"`
 	Limit       LimitConfig `json:"limit" yaml:"limit"`
+
+	// Generation parameter defaults for this model.
+	// These are applied when the user hasn't explicitly set the corresponding
+	// CLI flag or global config value.
+	Params GenerationParams `json:"params,omitzero" yaml:"params,omitempty"`
 }
 
 // CostConfig defines the pricing for a custom model.
@@ -191,18 +264,19 @@ type Config struct {
 	Model          string                     `json:"model,omitempty" yaml:"model,omitempty"`
 	MaxSteps       int                        `json:"max-steps,omitempty" yaml:"max-steps,omitempty"`
 	Debug          bool                       `json:"debug,omitempty" yaml:"debug,omitempty"`
-	Compact        bool                       `json:"compact,omitempty" yaml:"compact,omitempty"`
 	SystemPrompt   string                     `json:"system-prompt,omitempty" yaml:"system-prompt,omitempty"`
 	ProviderAPIKey string                     `json:"provider-api-key,omitempty" yaml:"provider-api-key,omitempty"`
 	ProviderURL    string                     `json:"provider-url,omitempty" yaml:"provider-url,omitempty"`
 	Stream         *bool                      `json:"stream,omitempty" yaml:"stream,omitempty"`
 	Theme          any                        `json:"theme" yaml:"theme"`
 	// Model generation parameters
-	MaxTokens     int      `json:"max-tokens,omitempty" yaml:"max-tokens,omitempty"`
-	Temperature   *float32 `json:"temperature,omitempty" yaml:"temperature,omitempty"`
-	TopP          *float32 `json:"top-p,omitempty" yaml:"top-p,omitempty"`
-	TopK          *int32   `json:"top-k,omitempty" yaml:"top-k,omitempty"`
-	StopSequences []string `json:"stop-sequences,omitempty" yaml:"stop-sequences,omitempty"`
+	MaxTokens        int      `json:"max-tokens,omitempty" yaml:"max-tokens,omitempty"`
+	Temperature      *float32 `json:"temperature,omitempty" yaml:"temperature,omitempty"`
+	TopP             *float32 `json:"top-p,omitempty" yaml:"top-p,omitempty"`
+	TopK             *int32   `json:"top-k,omitempty" yaml:"top-k,omitempty"`
+	FrequencyPenalty *float32 `json:"frequency-penalty,omitempty" yaml:"frequency-penalty,omitempty"`
+	PresencePenalty  *float32 `json:"presence-penalty,omitempty" yaml:"presence-penalty,omitempty"`
+	StopSequences    []string `json:"stop-sequences,omitempty" yaml:"stop-sequences,omitempty"`
 
 	// Thinking / extended reasoning
 	ThinkingLevel string `json:"thinking-level,omitempty" yaml:"thinking-level,omitempty"`
@@ -216,6 +290,12 @@ type Config struct {
 
 	// Custom model definitions (under custom/ provider)
 	CustomModels map[string]CustomModelConfig `json:"customModels,omitempty" yaml:"customModels,omitempty"`
+
+	// Per-model generation parameter overrides. Keys are "provider/model" strings
+	// (e.g. "anthropic/claude-sonnet-4-5-20250929", "openai/gpt-4o"). These
+	// settings act as model-level defaults — CLI flags and global config values
+	// take precedence when explicitly set.
+	ModelSettings map[string]GenerationParams `json:"modelSettings,omitempty" yaml:"modelSettings,omitempty"`
 }
 
 // GetTransportType returns the transport type for the server config, mapping
@@ -234,11 +314,18 @@ func (s *MCPServerConfig) GetTransportType() string {
 			return "stdio"
 		case "remote":
 			return "streamable"
+		case "inprocess":
+			return "inprocess"
 		default:
 			return s.Type
 		}
 	}
 
+	// Programmatic in-process server detection.
+	if s.InProcessServer != nil {
+		return "inprocess"
+	}
+
 	// Backward compatibility: infer transport type
 	if len(s.Command) > 0 {
 		return "stdio"
@@ -258,6 +345,17 @@ func (c *Config) Validate() error {
 			return fmt.Errorf("server %s: allowedTools and excludedTools are mutually exclusive", serverName)
 		}
 
+		// Reject unknown tasksMode values up front so a typo (e.g. "alwasy")
+		// fails loud here instead of being silently downgraded to "auto" by
+		// the runtime parser. Comparison is case-insensitive to match
+		// tools.ParseTaskMode.
+		switch strings.ToLower(strings.TrimSpace(serverConfig.TasksMode)) {
+		case "", "auto", "never", "always":
+			// ok
+		default:
+			return fmt.Errorf("server %s: invalid tasksMode %q (expected one of: auto, never, always)", serverName, serverConfig.TasksMode)
+		}
+
 		transport := serverConfig.GetTransportType()
 		switch transport {
 		case "stdio":
@@ -269,8 +367,12 @@ func (c *Config) Validate() error {
 			if serverConfig.URL == "" {
 				return fmt.Errorf("server %s: url is required for %s transport", serverName, transport)
 			}
+		case "inprocess":
+			if serverConfig.InProcessServer == nil {
+				return fmt.Errorf("server %s: InProcessServer is required for inprocess transport", serverName)
+			}
 		default:
-			return fmt.Errorf("server %s: unsupported transport type '%s'. Supported types: stdio, sse, streamable", serverName, transport)
+			return fmt.Errorf("server %s: unsupported transport type '%s'. Supported types: stdio, sse, streamable, inprocess", serverName, transport)
 		}
 	}
 	return nil
@@ -364,16 +466,55 @@ mcpServers:
 # debug: false                                 # Enable debug logging
 # system-prompt: "/path/to/system-prompt.txt" # System prompt text file
 
-# Model generation parameters (all optional)
+# Model generation parameters (all optional, apply globally to all models)
 # max-tokens: 4096                             # Maximum tokens in response
 # temperature: 0.7                             # Randomness (0.0-1.0)
 # top-p: 0.95                                  # Nucleus sampling (0.0-1.0)
 # top-k: 40                                    # Top K sampling
+# frequency-penalty: 0.0                        # Penalize frequent tokens (0.0-2.0)
+# presence-penalty: 0.0                         # Penalize present tokens (0.0-2.0)
 # stop-sequences: ["Human:", "Assistant:"]     # Custom stop sequences
 
+# Per-model generation parameter overrides (apply to specific models)
+# These act as model-level defaults — CLI flags and global settings above take precedence.
+# Keys are "provider/model" strings matching the model you use.
+# modelSettings:
+#   anthropic/claude-sonnet-4-5-20250929:
+#     temperature: 0.3
+#     maxTokens: 8192
+#   openai/gpt-4o:
+#     temperature: 0.7
+#     topP: 0.95
+#     topK: 40
+#     frequencyPenalty: 0.1
+#     presencePenalty: 0.1
+#   anthropic/claude-opus-4-6:
+#     thinkingLevel: "high"
+#     maxTokens: 16384
+#     systemPrompt: "You are a deep reasoning assistant."  # or a file path
+
 # API Configuration (can also use environment variables)
 # provider-api-key: "your-api-key"         # API key for OpenAI, Anthropic, or Google
 # provider-url: "https://api.openai.com/v1" # Base URL for OpenAI, Anthropic, or Ollama
+
+# Custom model definitions (under custom/ provider)
+# customModels:
+#   my-local-llama:
+#     name: "Local Llama 3"
+#     baseUrl: "http://localhost:8080/v1"
+#     family: "llama"
+#     temperature: true
+#     cost:
+#       input: 0.0
+#       output: 0.0
+#     limit:
+#       context: 131072
+#       output: 8192
+#     params:                              # Generation parameter defaults for this model
+#       temperature: 0.8
+#       topP: 0.95
+#       topK: 40
+#       systemPrompt: "You are a helpful local assistant."
 `
 
 	_, err = file.WriteString(content)

internal/config/config_test.go

@@ -6,6 +6,8 @@ import (
 	"path/filepath"
 	"strings"
 	"testing"
+
+	"gopkg.in/yaml.v3"
 )
 
 func TestMCPServerConfig_NewFormat(t *testing.T) {
@@ -542,3 +544,175 @@ func TestEnsureConfigExistsWhenFileExists(t *testing.T) {
 		t.Error("Existing config file was modified when it shouldn't have been")
 	}
 }
+
+func TestMCPServerConfig_OAuthFields_JSON(t *testing.T) {
+	jsonData := `{
+		"type": "remote",
+		"url": "https://api.githubcopilot.com/mcp/",
+		"oauthClientId": "Ov23liXXXXXXXXXXXXXX",
+		"oauthClientSecret": "secret123",
+		"oauthScopes": ["read:user", "repo"]
+	}`
+
+	var cfg MCPServerConfig
+	err := json.Unmarshal([]byte(jsonData), &cfg)
+	if err != nil {
+		t.Fatalf("Failed to unmarshal: %v", err)
+	}
+
+	if cfg.Type != "remote" {
+		t.Errorf("Expected type 'remote', got %q", cfg.Type)
+	}
+	if cfg.URL != "https://api.githubcopilot.com/mcp/" {
+		t.Errorf("Expected URL, got %q", cfg.URL)
+	}
+	if cfg.OAuthClientID != "Ov23liXXXXXXXXXXXXXX" {
+		t.Errorf("Expected OAuthClientID 'Ov23liXXXXXXXXXXXXXX', got %q", cfg.OAuthClientID)
+	}
+	if cfg.OAuthClientSecret != "secret123" {
+		t.Errorf("Expected OAuthClientSecret 'secret123', got %q", cfg.OAuthClientSecret)
+	}
+	if len(cfg.OAuthScopes) != 2 || cfg.OAuthScopes[0] != "read:user" || cfg.OAuthScopes[1] != "repo" {
+		t.Errorf("Expected OAuthScopes [read:user, repo], got %v", cfg.OAuthScopes)
+	}
+}
+
+func TestMCPServerConfig_OAuthFields_YAML(t *testing.T) {
+	yamlData := `
+type: remote
+url: https://api.githubcopilot.com/mcp/
+oauthClientId: "Ov23liXXXXXXXXXXXXXX"
+oauthScopes:
+  - read:user
+  - repo
+`
+
+	var cfg MCPServerConfig
+	err := yaml.Unmarshal([]byte(yamlData), &cfg)
+	if err != nil {
+		t.Fatalf("Failed to unmarshal YAML: %v", err)
+	}
+
+	if cfg.Type != "remote" {
+		t.Errorf("Expected type 'remote', got %q", cfg.Type)
+	}
+	if cfg.OAuthClientID != "Ov23liXXXXXXXXXXXXXX" {
+		t.Errorf("Expected OAuthClientID 'Ov23liXXXXXXXXXXXXXX', got %q", cfg.OAuthClientID)
+	}
+	if len(cfg.OAuthScopes) != 2 || cfg.OAuthScopes[0] != "read:user" || cfg.OAuthScopes[1] != "repo" {
+		t.Errorf("Expected OAuthScopes [read:user, repo], got %v", cfg.OAuthScopes)
+	}
+}
+
+func TestMCPServerConfig_OAuthFields_Omitted(t *testing.T) {
+	// Verify that omitting OAuth fields still works (backward compat).
+	jsonData := `{
+		"type": "remote",
+		"url": "https://example.com/mcp"
+	}`
+
+	var cfg MCPServerConfig
+	err := json.Unmarshal([]byte(jsonData), &cfg)
+	if err != nil {
+		t.Fatalf("Failed to unmarshal: %v", err)
+	}
+
+	if cfg.OAuthClientID != "" {
+		t.Errorf("Expected empty OAuthClientID, got %q", cfg.OAuthClientID)
+	}
+	if cfg.OAuthClientSecret != "" {
+		t.Errorf("Expected empty OAuthClientSecret, got %q", cfg.OAuthClientSecret)
+	}
+	if len(cfg.OAuthScopes) != 0 {
+		t.Errorf("Expected empty OAuthScopes, got %v", cfg.OAuthScopes)
+	}
+}
+
+func TestMCPServerConfig_TasksMode_NewFormat(t *testing.T) {
+	jsonData := `{
+		"type": "remote",
+		"url": "https://my-mcp-server.com",
+		"tasksMode": "always"
+	}`
+	var cfg MCPServerConfig
+	if err := json.Unmarshal([]byte(jsonData), &cfg); err != nil {
+		t.Fatalf("Failed to unmarshal: %v", err)
+	}
+	if cfg.TasksMode != "always" {
+		t.Errorf("expected TasksMode 'always', got %q", cfg.TasksMode)
+	}
+}
+
+func TestMCPServerConfig_TasksMode_LegacyFormat(t *testing.T) {
+	// tasksMode also recognised in the legacy unmarshal path so users on
+	// the older command/args shape can opt in without migrating.
+	jsonData := `{
+		"command": "npx",
+		"args": ["@modelcontextprotocol/server-filesystem", "/path"],
+		"tasksMode": "never"
+	}`
+	var cfg MCPServerConfig
+	if err := json.Unmarshal([]byte(jsonData), &cfg); err != nil {
+		t.Fatalf("Failed to unmarshal: %v", err)
+	}
+	if cfg.TasksMode != "never" {
+		t.Errorf("expected TasksMode 'never', got %q", cfg.TasksMode)
+	}
+}
+
+func TestMCPServerConfig_TasksMode_DefaultEmpty(t *testing.T) {
+	// When tasksMode is not set the field stays empty, which downstream
+	// resolves to "auto" via tools.ParseTaskMode.
+	jsonData := `{"type":"remote","url":"https://x.example"}`
+	var cfg MCPServerConfig
+	if err := json.Unmarshal([]byte(jsonData), &cfg); err != nil {
+		t.Fatalf("Failed to unmarshal: %v", err)
+	}
+	if cfg.TasksMode != "" {
+		t.Errorf("expected default TasksMode to be empty, got %q", cfg.TasksMode)
+	}
+}
+
+func TestConfig_Validate_TasksMode(t *testing.T) {
+	t.Run("empty is valid", func(t *testing.T) {
+		cfg := &Config{
+			MCPServers: map[string]MCPServerConfig{
+				"a": {Type: "remote", URL: "https://x.example"},
+			},
+		}
+		if err := cfg.Validate(); err != nil {
+			t.Errorf("empty TasksMode should validate, got %v", err)
+		}
+	})
+
+	t.Run("known values are valid", func(t *testing.T) {
+		for _, mode := range []string{"auto", "never", "always", "AUTO", " always "} {
+			cfg := &Config{
+				MCPServers: map[string]MCPServerConfig{
+					"a": {Type: "remote", URL: "https://x.example", TasksMode: mode},
+				},
+			}
+			if err := cfg.Validate(); err != nil {
+				t.Errorf("TasksMode=%q should validate, got %v", mode, err)
+			}
+		}
+	})
+
+	t.Run("typo is rejected with a clear error", func(t *testing.T) {
+		cfg := &Config{
+			MCPServers: map[string]MCPServerConfig{
+				"buildbot": {Type: "remote", URL: "https://x.example", TasksMode: "alwasy"},
+			},
+		}
+		err := cfg.Validate()
+		if err == nil {
+			t.Fatal("expected validation error for invalid TasksMode")
+		}
+		// Error must mention the server name AND the bad value so the
+		// user knows where to look.
+		msg := err.Error()
+		if !strings.Contains(msg, "buildbot") || !strings.Contains(msg, `"alwasy"`) {
+			t.Errorf("error %q should mention both server name and bad value", msg)
+		}
+	})
+}

internal/config/merger.go

@@ -7,32 +7,48 @@ import (
 	"github.com/spf13/viper"
 )
 
-// LoadAndValidateConfig loads configuration from viper, fixes environment variable
-// casing issues, and validates the configuration. Returns an error if loading or
-// validation fails.
+// LoadAndValidateConfig loads configuration from the process-global viper
+// store, fixes environment variable casing issues, and validates the
+// configuration. Returns an error if loading or validation fails.
+//
+// This is a convenience wrapper around [LoadAndValidateConfigFrom] using the
+// shared global store; it is retained for the CLI and other callers that rely
+// on viper's process-global state.
 func LoadAndValidateConfig() (*Config, error) {
+	return LoadAndValidateConfigFrom(viper.GetViper())
+}
+
+// LoadAndValidateConfigFrom loads configuration from the supplied per-instance
+// store, fixes environment variable casing issues, and validates the
+// configuration. When v is nil, the process-global store is used. Threading an
+// explicit store lets each Kit instance own an isolated configuration without
+// clobbering other instances in the same process.
+func LoadAndValidateConfigFrom(v *viper.Viper) (*Config, error) {
+	if v == nil {
+		v = viper.GetViper()
+	}
 	config := &Config{
 		MCPServers: make(map[string]MCPServerConfig),
 	}
-	if err := viper.Unmarshal(config); err != nil {
-		return nil, fmt.Errorf("failed to unmarshal config: %v", err)
+	if err := v.Unmarshal(config); err != nil {
+		return nil, fmt.Errorf("failed to unmarshal config: %w", err)
 	}
 
 	// Fix environment variable case sensitivity issue
 	// Viper lowercases all keys, but we need to preserve the original case for environment variables
-	fixEnvironmentCase(config)
+	fixEnvironmentCase(v, config)
 
 	if err := config.Validate(); err != nil {
-		return nil, fmt.Errorf("invalid config: %v", err)
+		return nil, fmt.Errorf("invalid config: %w", err)
 	}
 
 	return config, nil
 }
 
 // fixEnvironmentCase fixes the case of environment variable keys that were lowercased by Viper
-func fixEnvironmentCase(config *Config) {
+func fixEnvironmentCase(v *viper.Viper, config *Config) {
 	// Get the raw config data from viper
-	rawConfig := viper.AllSettings()
+	rawConfig := v.AllSettings()
 
 	// Check if we have mcpServers in the raw config
 	if mcpServersRaw, ok := rawConfig["mcpservers"]; ok {

internal/core/bash.go

@@ -19,10 +19,18 @@ import (
 // It receives tool call ID, tool name, output chunk, and whether it's stderr.
 type ToolOutputCallback func(toolCallID, toolName, chunk string, isStderr bool)
 
+// PasswordPromptCallback is the signature for password prompts.
+// It receives a prompt message and returns the password and whether it was cancelled.
+type PasswordPromptCallback func(prompt string) (password string, cancelled bool)
+
 // contextKey is a custom type for context keys to avoid collisions.
 type contextKey string
 
-const toolOutputCallbackKey contextKey = "toolOutputCallback"
+const (
+	toolOutputCallbackKey contextKey = "toolOutputCallback"
+	sudoPasswordKey       contextKey = "sudoPassword"
+	passwordPromptKey     contextKey = "passwordPrompt"
+)
 
 // ContextWithToolOutputCallback returns a new context with the tool output callback set.
 func ContextWithToolOutputCallback(ctx context.Context, callback ToolOutputCallback) context.Context {
@@ -37,6 +45,34 @@ func toolOutputCallbackFromContext(ctx context.Context) ToolOutputCallback {
 	return nil
 }
 
+// ContextWithPasswordPrompt returns a new context with the password prompt callback set.
+// This allows the TUI to show a modal password prompt when sudo needs a password.
+func ContextWithPasswordPrompt(ctx context.Context, callback PasswordPromptCallback) context.Context {
+	return context.WithValue(ctx, passwordPromptKey, callback)
+}
+
+// passwordPromptFromContext retrieves the password prompt callback from context.
+func passwordPromptFromContext(ctx context.Context) PasswordPromptCallback {
+	if cb, ok := ctx.Value(passwordPromptKey).(PasswordPromptCallback); ok {
+		return cb
+	}
+	return nil
+}
+
+// ContextWithSudoPassword returns a new context with the sudo password set.
+// When present, the bash tool will use sudo -S to pipe this password to sudo commands.
+func ContextWithSudoPassword(ctx context.Context, password string) context.Context {
+	return context.WithValue(ctx, sudoPasswordKey, password)
+}
+
+// sudoPasswordFromContext retrieves the sudo password from context.
+func sudoPasswordFromContext(ctx context.Context) string {
+	if pw, ok := ctx.Value(sudoPasswordKey).(string); ok {
+		return pw
+	}
+	return ""
+}
+
 const defaultBashTimeout = 120 * time.Second
 const maxBashTimeout = 600 * time.Second
 
@@ -73,6 +109,57 @@ func NewBashTool(opts ...ToolOption) fantasy.AgentTool {
 	}
 }
 
+// sudoCommandRe matches sudo commands that need to be rewritten for -S mode.
+// It matches "sudo" as a word boundary, optionally preceded by environment variables.
+var sudoCommandRe = regexp.MustCompile(`(?i)(^|[&|;|]|\|\||&&)\s*(\w+=\S+\s+)?\bsudo\b`)
+
+// truncateCommand truncates a long command for display.
+func truncateCommand(cmd string, maxLen int) string {
+	if len(cmd) <= maxLen {
+		return cmd
+	}
+	return cmd[:maxLen-3] + "..."
+}
+
+// rewriteSudoForStdin rewrites sudo commands to use -S -p ” for stdin password input.
+// It transforms: sudo cmd → sudo -S -p ” cmd
+func rewriteSudoForStdin(command string) string {
+	// Find all matches and their positions
+	matches := sudoCommandRe.FindAllStringIndex(command, -1)
+	if matches == nil {
+		return command
+	}
+
+	// Build result from end to start to preserve indices
+	result := command
+	for i := len(matches) - 1; i >= 0; i-- {
+		match := matches[i]
+		start, end := match[0], match[1]
+		matchedText := result[start:end]
+
+		// Extract just the "sudo" part (after any prefix)
+		sudoIdx := strings.Index(strings.ToLower(matchedText), "sudo")
+		if sudoIdx == -1 {
+			continue
+		}
+		prefix := matchedText[:sudoIdx]
+		sudoPart := matchedText[sudoIdx:]
+
+		// Check if the text immediately after "sudo" in the result contains -S
+		afterSudo := result[end:]
+		if strings.HasPrefix(strings.TrimLeft(afterSudo, " \t"), "-S") {
+			// Already has -S flag, skip
+			continue
+		}
+
+		// Insert -S -p '' after "sudo"
+		newSudo := strings.Replace(sudoPart, "sudo", "sudo -S -p ''", 1)
+		result = result[:start] + prefix + newSudo + result[end:]
+	}
+
+	return result
+}
+
 func executeBash(ctx context.Context, call fantasy.ToolCall, workDir string) (fantasy.ToolResponse, error) {
 	var args bashArgs
 	if err := parseArgs(call.Input, &args); err != nil {
@@ -97,7 +184,47 @@ func executeBash(ctx context.Context, call fantasy.ToolCall, workDir string) (fa
 	cmdCtx, cancel := context.WithTimeout(ctx, timeout)
 	defer cancel()
 
-	cmd := exec.CommandContext(cmdCtx, "bash", "-c", args.Command)
+	// Check for sudo password in context or environment
+	sudoPassword := sudoPasswordFromContext(ctx)
+	if sudoPassword == "" {
+		sudoPassword = os.Getenv("SUDO_PASSWORD")
+	}
+	command := args.Command
+
+	// If command contains sudo and we don't have a password, check if sudo needs one
+	if sudoPassword == "" && sudoCommandRe.MatchString(command) {
+		// Check if sudo credentials are cached using sudo -n (non-interactive)
+		testCmd := exec.CommandContext(cmdCtx, "sudo", "-n", "true")
+		testCmd.Dir = workDir
+		if err := testCmd.Run(); err != nil {
+			// Sudo needs a password - try to prompt via callback
+			if promptCallback := passwordPromptFromContext(ctx); promptCallback != nil {
+				pw, cancelled := promptCallback("Sudo password required for: " + truncateCommand(args.Command, 60))
+				if cancelled {
+					return fantasy.NewTextErrorResponse("sudo password prompt cancelled"), nil
+				}
+				if pw == "" {
+					return fantasy.NewTextErrorResponse("no sudo password provided"), nil
+				}
+				sudoPassword = pw
+				command = rewriteSudoForStdin(command)
+			} else {
+				// No callback available - return error with helpful message
+				return fantasy.NewTextErrorResponse(
+					"This command requires sudo access. " +
+						"Please run 'sudo -v' in your terminal first to cache credentials, " +
+						"or set the SUDO_PASSWORD environment variable."), nil
+			}
+		}
+		// Credentials are cached or password was provided, proceed
+	}
+
+	// If we have a sudo password, rewrite the command to use sudo -S
+	if sudoPassword != "" && sudoCommandRe.MatchString(command) {
+		command = rewriteSudoForStdin(command)
+	}
+
+	cmd := exec.CommandContext(cmdCtx, "bash", "-c", command)
 	if workDir != "" {
 		cmd.Dir = workDir
 	}
@@ -115,18 +242,18 @@ func executeBash(ctx context.Context, call fantasy.ToolCall, workDir string) (fa
 
 	if outputCallback != nil {
 		// Streaming mode: use pipes to capture output as it arrives
-		return executeBashStreaming(cmdCtx, call, cmd, outputCallback)
+		return executeBashStreaming(cmdCtx, call, cmd, outputCallback, sudoPassword)
 	}
 
 	// Non-streaming mode: collect all output at once (original behavior)
-	return executeBashBuffered(cmdCtx, call, cmd)
+	return executeBashBuffered(cmdCtx, call, cmd, sudoPassword)
 }
 
 // executeBashBuffered collects all output before returning (original behavior).
 // It uses explicit pipes (not cmd.Stdout) so that cmd.WaitDelay can forcibly
 // close them when grandchild processes hold pipe handles open after the
 // direct child exits.
-func executeBashBuffered(cmdCtx context.Context, call fantasy.ToolCall, cmd *exec.Cmd) (fantasy.ToolResponse, error) {
+func executeBashBuffered(cmdCtx context.Context, call fantasy.ToolCall, cmd *exec.Cmd, sudoPassword string) (fantasy.ToolResponse, error) {
 	stdoutPipe, err := cmd.StdoutPipe()
 	if err != nil {
 		return fantasy.NewTextErrorResponse("failed to create stdout pipe"), nil
@@ -136,10 +263,27 @@ func executeBashBuffered(cmdCtx context.Context, call fantasy.ToolCall, cmd *exe
 		return fantasy.NewTextErrorResponse("failed to create stderr pipe"), nil
 	}
 
+	// If we have a sudo password, create a stdin pipe and write the password
+	var stdinPipe io.WriteCloser
+	if sudoPassword != "" {
+		stdinPipe, err = cmd.StdinPipe()
+		if err != nil {
+			return fantasy.NewTextErrorResponse("failed to create stdin pipe"), nil
+		}
+	}
+
 	if err := cmd.Start(); err != nil {
 		return fantasy.NewTextErrorResponse(fmt.Sprintf("failed to start command: %v", err)), nil
 	}
 
+	// Write password to stdin if needed, then close stdin
+	if sudoPassword != "" && stdinPipe != nil {
+		go func() {
+			defer func() { _ = stdinPipe.Close() }()
+			_, _ = io.WriteString(stdinPipe, sudoPassword+"\n")
+		}()
+	}
+
 	// Read pipes concurrently
 	var wg sync.WaitGroup
 	var stdout, stderr strings.Builder
@@ -181,7 +325,7 @@ func executeBashBuffered(cmdCtx context.Context, call fantasy.ToolCall, cmd *exe
 }
 
 // executeBashStreaming streams output as it arrives via the callback.
-func executeBashStreaming(cmdCtx context.Context, call fantasy.ToolCall, cmd *exec.Cmd, outputCallback ToolOutputCallback) (fantasy.ToolResponse, error) {
+func executeBashStreaming(cmdCtx context.Context, call fantasy.ToolCall, cmd *exec.Cmd, outputCallback ToolOutputCallback, sudoPassword string) (fantasy.ToolResponse, error) {
 	stdoutPipe, err := cmd.StdoutPipe()
 	if err != nil {
 		return fantasy.NewTextErrorResponse("failed to create stdout pipe"), nil
@@ -191,11 +335,28 @@ func executeBashStreaming(cmdCtx context.Context, call fantasy.ToolCall, cmd *ex
 		return fantasy.NewTextErrorResponse("failed to create stderr pipe"), nil
 	}
 
+	// If we have a sudo password, create a stdin pipe
+	var stdinPipe io.WriteCloser
+	if sudoPassword != "" {
+		stdinPipe, err = cmd.StdinPipe()
+		if err != nil {
+			return fantasy.NewTextErrorResponse("failed to create stdin pipe"), nil
+		}
+	}
+
 	// Start command execution
 	if err := cmd.Start(); err != nil {
 		return fantasy.NewTextErrorResponse(fmt.Sprintf("failed to start command: %v", err)), nil
 	}
 
+	// Write password to stdin if needed, then close stdin
+	if sudoPassword != "" && stdinPipe != nil {
+		go func() {
+			defer func() { _ = stdinPipe.Close() }()
+			_, _ = io.WriteString(stdinPipe, sudoPassword+"\n")
+		}()
+	}
+
 	// Stream stdout and stderr concurrently
 	var wg sync.WaitGroup
 	var mu sync.Mutex

internal/core/bash_test.go

@@ -127,3 +127,72 @@ func TestBash_EmptyCommand(t *testing.T) {
 		t.Fatal("expected error for empty command")
 	}
 }
+
+func TestRewriteSudoForStdin(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		expected string
+	}{
+		{
+			name:     "simple sudo",
+			input:    "sudo apt update",
+			expected: "sudo -S -p '' apt update",
+		},
+		{
+			name:     "sudo with env var",
+			input:    "DEBIAN_FRONTEND=noninteractive sudo apt update",
+			expected: "DEBIAN_FRONTEND=noninteractive sudo -S -p '' apt update",
+		},
+		{
+			name:     "sudo in pipeline",
+			input:    "echo test | sudo tee /etc/test.conf",
+			expected: "echo test | sudo -S -p '' tee /etc/test.conf",
+		},
+		{
+			name:     "sudo after &&",
+			input:    "apt update && sudo apt upgrade",
+			expected: "apt update && sudo -S -p '' apt upgrade",
+		},
+		{
+			name:     "already has -S flag",
+			input:    "sudo -S apt update",
+			expected: "sudo -S apt update",
+		},
+		{
+			name:     "no sudo",
+			input:    "apt update && apt upgrade",
+			expected: "apt update && apt upgrade",
+		},
+		{
+			name:     "sudo in string (should not match)",
+			input:    "echo 'use sudo carefully'",
+			expected: "echo 'use sudo carefully'",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := rewriteSudoForStdin(tt.input)
+			if result != tt.expected {
+				t.Errorf("rewriteSudoForStdin(%q) = %q, want %q", tt.input, result, tt.expected)
+			}
+		})
+	}
+}
+
+func TestSudoPasswordFromContext(t *testing.T) {
+	// Test with password in context
+	ctx := ContextWithSudoPassword(context.Background(), "secret123")
+	pw := sudoPasswordFromContext(ctx)
+	if pw != "secret123" {
+		t.Errorf("expected password 'secret123', got %q", pw)
+	}
+
+	// Test without password
+	ctx = context.Background()
+	pw = sudoPasswordFromContext(ctx)
+	if pw != "" {
+		t.Errorf("expected empty password, got %q", pw)
+	}
+}

internal/core/edit.go

@@ -21,12 +21,9 @@ type Edit struct {
 }
 
 // editArgs holds the arguments for the edit tool.
-// Supports both single-edit mode (old_text/new_text) and multi-edit mode (edits array).
 type editArgs struct {
-	Path    string `json:"path"`
-	OldText string `json:"old_text"` // Single-edit mode
-	NewText string `json:"new_text"` // Single-edit mode
-	Edits   []Edit `json:"edits"`    // Multi-edit mode
+	Path  string `json:"path"`
+	Edits []Edit `json:"edits"`
 }
 
 // replacement represents a normalized edit ready for processing.
@@ -52,20 +49,12 @@ func NewEditTool(opts ...ToolOption) fantasy.AgentTool {
 	return &coreTool{
 		info: fantasy.ToolInfo{
 			Name:        "edit",
-			Description: "Edit a file by replacing exact text. Supports single edit via old_text/new_text, or multiple edits via the edits array. All edits in the array are matched against the original file content (non-incremental) and must be non-overlapping.",
+			Description: "Edit a file by replacing exact text. All edits in the array are matched against the original file content (non-incremental) and must be non-overlapping.",
 			Parameters: map[string]any{
 				"path": map[string]any{
 					"type":        "string",
 					"description": "Path to the file to edit (relative or absolute)",
 				},
-				"old_text": map[string]any{
-					"type":        "string",
-					"description": "Exact text to find and replace (single-edit mode). Must not be used with 'edits' array.",
-				},
-				"new_text": map[string]any{
-					"type":        "string",
-					"description": "New text to replace the old text with (single-edit mode). Must not be used with 'edits' array.",
-				},
 				"edits": map[string]any{
 					"type":        "array",
 					"description": "Array of edits for multi-region replacement. Each edit must have unique, non-overlapping old_text. All matches are against the original file content.",
@@ -85,7 +74,7 @@ func NewEditTool(opts ...ToolOption) fantasy.AgentTool {
 					},
 				},
 			},
-			Required: []string{"path"},
+			Required: []string{"path", "edits"},
 		},
 		handler: func(ctx context.Context, call fantasy.ToolCall) (fantasy.ToolResponse, error) {
 			return executeEdit(ctx, call, cfg.WorkDir)
@@ -163,36 +152,11 @@ func executeEdit(ctx context.Context, call fantasy.ToolCall, workDir string) (fa
 }
 
 // normalizeEditInput validates and normalizes the edit input.
-// Returns error if both single-edit and multi-edit modes are used.
 func normalizeEditInput(args editArgs) ([]replacement, error) {
-	singleMode := args.OldText != "" || args.NewText != ""
-	multiMode := len(args.Edits) > 0
-
-	if singleMode && multiMode {
-		return nil, fmt.Errorf("cannot use old_text/new_text together with edits array")
+	if len(args.Edits) == 0 {
+		return nil, fmt.Errorf("edits array is required and must not be empty")
 	}
 
-	if !singleMode && !multiMode {
-		return nil, fmt.Errorf("must provide either old_text/new_text or edits array")
-	}
-
-	if singleMode {
-		if args.OldText == "" {
-			return nil, fmt.Errorf("old_text is required when using single-edit mode")
-		}
-		if args.NewText == "" {
-			return nil, fmt.Errorf("new_text is required when using single-edit mode")
-		}
-		return []replacement{{
-			oldText:     strings.ReplaceAll(args.OldText, "\r\n", "\n"),
-			newText:     strings.ReplaceAll(args.NewText, "\r\n", "\n"),
-			originalOld: args.OldText,
-			originalNew: args.NewText,
-			index:       0,
-		}}, nil
-	}
-
-	// Multi-edit mode
 	var reps []replacement
 	for i, edit := range args.Edits {
 		if edit.OldText == "" {

internal/core/edit_test.go

@@ -389,9 +389,11 @@ func TestExecuteEdit_ExactMatch(t *testing.T) {
 	writeFileOrFail(t, path, original)
 
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: "fmt.Println(\"hello\")",
-		NewText: "fmt.Println(\"world\")",
+		Path: path,
+		Edits: []Edit{{
+			OldText: "fmt.Println(\"hello\")",
+			NewText: "fmt.Println(\"world\")",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -426,9 +428,11 @@ func TestExecuteEdit_ExactMatch_DoesNotCorruptRest(t *testing.T) {
 	target := lines[49]
 	replacement := "REPLACED_LINE_50"
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: target,
-		NewText: replacement,
+		Path: path,
+		Edits: []Edit{{
+			OldText: target,
+			NewText: replacement,
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -470,9 +474,11 @@ func TestExecuteEdit_FuzzyMatch_TrailingWhitespace(t *testing.T) {
 
 	// Search without trailing whitespace (common LLM behavior)
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: "func foo() {\n\treturn 1\n}",
-		NewText: "func foo() {\n\treturn 2\n}",
+		Path: path,
+		Edits: []Edit{{
+			OldText: "func foo() {\n\treturn 1\n}",
+			NewText: "func foo() {\n\treturn 2\n}",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -519,9 +525,11 @@ func TestExecuteEdit_FuzzyMatch_DoesNotCorruptRest(t *testing.T) {
 	search := strings.Repeat("x", 10) + "\n" + strings.Repeat("x", 10)
 	// But this matches lines 1-2, 2-3, etc. — should fail due to ambiguity.
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: search,
-		NewText: "REPLACED",
+		Path: path,
+		Edits: []Edit{{
+			OldText: search,
+			NewText: "REPLACED",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -546,9 +554,11 @@ func TestExecuteEdit_MultipleMatches_Fails(t *testing.T) {
 	writeFileOrFail(t, path, "hello\nworld\nhello\n")
 
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: "hello",
-		NewText: "goodbye",
+		Path: path,
+		Edits: []Edit{{
+			OldText: "hello",
+			NewText: "goodbye",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -575,9 +585,11 @@ func TestExecuteEdit_NoMatch_Fails(t *testing.T) {
 	writeFileOrFail(t, path, "hello world\n")
 
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: "nonexistent text",
-		NewText: "replacement",
+		Path: path,
+		Edits: []Edit{{
+			OldText: "nonexistent text",
+			NewText: "replacement",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -601,9 +613,11 @@ func TestExecuteEdit_CRLFNormalization(t *testing.T) {
 	writeFileOrFail(t, path, "line1\r\nline2\r\nline3\r\n")
 
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: "line2",
-		NewText: "LINE2",
+		Path: path,
+		Edits: []Edit{{
+			OldText: "line2",
+			NewText: "LINE2",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -622,8 +636,10 @@ func TestExecuteEdit_CRLFNormalization(t *testing.T) {
 
 func TestExecuteEdit_MissingPath(t *testing.T) {
 	input, _ := json.Marshal(editArgs{
-		OldText: "x",
-		NewText: "y",
+		Edits: []Edit{{
+			OldText: "x",
+			NewText: "y",
+		}},
 	})
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, "")
 	if err != nil {
@@ -636,9 +652,11 @@ func TestExecuteEdit_MissingPath(t *testing.T) {
 
 func TestExecuteEdit_NonexistentFile(t *testing.T) {
 	input, _ := json.Marshal(editArgs{
-		Path:    "/tmp/nonexistent_edit_test_file_12345.go",
-		OldText: "x",
-		NewText: "y",
+		Path: "/tmp/nonexistent_edit_test_file_12345.go",
+		Edits: []Edit{{
+			OldText: "x",
+			NewText: "y",
+		}},
 	})
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, "")
 	if err != nil {
@@ -661,9 +679,11 @@ func TestExecuteEdit_DiffContainsHunkHeader(t *testing.T) {
 	writeFileOrFail(t, path, strings.Join(lines, "\n")+"\n")
 
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: "line_10_content",
-		NewText: "REPLACED",
+		Path: path,
+		Edits: []Edit{{
+			OldText: "line_10_content",
+			NewText: "REPLACED",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -684,9 +704,11 @@ func TestExecuteEdit_MetadataContainsFileDiffs(t *testing.T) {
 	writeFileOrFail(t, path, "old content\n")
 
 	input, _ := json.Marshal(editArgs{
-		Path:    path,
-		OldText: "old content",
-		NewText: "new content",
+		Path: path,
+		Edits: []Edit{{
+			OldText: "old content",
+			NewText: "new content",
+		}},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -905,18 +927,14 @@ func TestExecuteEdit_MultiEdit_EmptyArray(t *testing.T) {
 	}
 }
 
-func TestExecuteEdit_MultiEdit_MixedWithSingleMode(t *testing.T) {
+func TestExecuteEdit_EmptyEditsArray_Fails(t *testing.T) {
 	dir := t.TempDir()
-	path := filepath.Join(dir, "mixed.txt")
+	path := filepath.Join(dir, "empty.txt")
 	writeFileOrFail(t, path, "hello\n")
 
-	input, _ := json.Marshal(map[string]any{
-		"path":     path,
-		"old_text": "hello",
-		"new_text": "HELLO",
-		"edits": []Edit{
-			{OldText: "hello", NewText: "HI"},
-		},
+	input, _ := json.Marshal(editArgs{
+		Path:  path,
+		Edits: []Edit{},
 	})
 
 	resp, err := executeEdit(t.Context(), fantasy.ToolCall{Input: string(input)}, dir)
@@ -924,10 +942,10 @@ func TestExecuteEdit_MultiEdit_MixedWithSingleMode(t *testing.T) {
 		t.Fatalf("executeEdit error: %v", err)
 	}
 	if !resp.IsError {
-		t.Error("expected error when mixing single and multi-edit modes")
+		t.Error("expected error for empty edits array")
 	}
-	if !strings.Contains(resp.Content, "cannot use") {
-		t.Errorf("expected 'cannot use' in error, got: %s", resp.Content)
+	if !strings.Contains(resp.Content, "required") {
+		t.Errorf("expected 'required' in error, got: %s", resp.Content)
 	}
 }
 

internal/core/read.go

@@ -67,7 +67,7 @@ func executeRead(ctx context.Context, call fantasy.ToolCall, workDir string) (fa
 	}
 
 	if info.IsDir() {
-		return readDirectory(absPath)
+		return fantasy.NewTextErrorResponse(fmt.Sprintf("'%s' is a directory, not a file. Use the ls tool to list directory contents.", args.Path)), nil
 	}
 
 	content, err := os.ReadFile(absPath)
@@ -116,25 +116,6 @@ func executeRead(ctx context.Context, call fantasy.ToolCall, workDir string) (fa
 	return fantasy.NewTextResponse(tr.Content), nil
 }
 
-func readDirectory(absPath string) (fantasy.ToolResponse, error) {
-	entries, err := os.ReadDir(absPath)
-	if err != nil {
-		return fantasy.NewTextErrorResponse(fmt.Sprintf("failed to read directory: %v", err)), nil
-	}
-
-	var result strings.Builder
-	for _, entry := range entries {
-		name := entry.Name()
-		if entry.IsDir() {
-			name += "/"
-		}
-		result.WriteString(name + "\n")
-	}
-
-	tr := truncateHead(result.String(), 500, defaultMaxBytes)
-	return fantasy.NewTextResponse(tr.Content), nil
-}
-
 // resolvePathWithWorkDir resolves a path to an absolute path relative to the
 // given workDir. If workDir is empty, os.Getwd() is used.
 func resolvePathWithWorkDir(path, workDir string) (string, error) {

internal/core/subagent.go

@@ -86,7 +86,7 @@ Example use cases:
 				},
 				"model": map[string]any{
 					"type":        "string",
-					"description": "Optional model override (e.g. 'anthropic/claude-haiku-3-5-20241022' for faster/cheaper tasks)",
+					"description": "Optional model override. Empty string uses the current model.",
 				},
 				"system_prompt": map[string]any{
 					"type":        "string",
@@ -94,7 +94,7 @@ Example use cases:
 				},
 				"timeout_seconds": map[string]any{
 					"type":        "number",
-					"description": "Maximum execution time in seconds (default: 300, max: 1800)",
+					"description": "Maximum execution time in seconds (default: 300, max: 1800, minimum recommended: 240)",
 				},
 			},
 			Required: []string{"task"},
@@ -130,13 +130,22 @@ func executeSubagent(ctx context.Context, call fantasy.ToolCall) (fantasy.ToolRe
 		), fmt.Errorf("no subagent spawner in context")
 	}
 
-	// Detach from the parent's deadline so the subagent gets its own
-	// independent timeout (applied downstream in Kit.Subagent). The parent
-	// context may carry a tight deadline from the LLM generation loop or
-	// other tool timeouts that would prematurely kill the subagent.
-	// We preserve context values (spawner, etc.) and propagate parent
-	// cancellation (e.g. user hits Ctrl-C) without inheriting the deadline.
-	spawnCtx := detachedWithCancel(ctx)
+	// Build a clean context for the subagent that inherits values (e.g. the
+	// spawner callback) but is completely detached from the parent's
+	// deadline AND cancellation. The subagent gets its own independent
+	// timeout (applied downstream in Kit.Subagent).
+	//
+	// Why full detachment instead of propagating parent cancellation?
+	// The parent context may already be done (deadline exceeded or
+	// cancelled) by the time this tool handler executes — for example when
+	// the generation loop context carries a deadline, when the user
+	// double-ESC cancels mid-turn, or when parallel tool execution
+	// encounters a race between stream completion and tool dispatch. Using
+	// context.WithoutCancel (Go 1.21+) ensures the subagent always starts
+	// cleanly with a fresh timeout, following the pattern used by crush for
+	// shutdown-resilient child work. The subagent's own timeout
+	// (defaultSubagentTimeout / user-specified) provides the safety net.
+	spawnCtx := context.WithoutCancel(valuesContext{parent: ctx})
 
 	// Spawn in-process subagent.
 	result, err := spawner(spawnCtx, call.ID, args.Task, args.Model, args.SystemPrompt, timeout)
@@ -173,37 +182,21 @@ func executeSubagent(ctx context.Context, call fantasy.ToolCall) (fantasy.ToolRe
 }
 
 // ---------------------------------------------------------------------------
-// Context detachment
+// Context helpers
 // ---------------------------------------------------------------------------
 
-// detachedContext wraps a parent context, preserving its values but removing
-// its deadline and cancellation. This allows the subagent to have its own
-// independent timeout while still accessing context-stored values (e.g. the
-// subagent spawner function).
-type detachedContext struct {
+// valuesContext preserves a parent context's values (e.g. the subagent
+// spawner callback) while stripping its deadline and cancellation. Combined
+// with context.WithoutCancel() this gives the subagent a completely clean
+// context that only inherits value-based dependencies.
+type valuesContext struct {
 	parent context.Context
 }
 
-func (d detachedContext) Deadline() (time.Time, bool) { return time.Time{}, false }
-func (d detachedContext) Done() <-chan struct{}       { return nil }
-func (d detachedContext) Err() error                  { return nil }
-func (d detachedContext) Value(key any) any           { return d.parent.Value(key) }
-
-// detachedWithCancel creates a new context that inherits values from the
-// parent but has no deadline. Cancellation of the parent is propagated: when
-// the parent is cancelled the returned context is also cancelled, but the
-// parent's deadline does not apply to the child.
-func detachedWithCancel(parent context.Context) context.Context {
-	child, cancel := context.WithCancel(detachedContext{parent: parent})
-	go func() {
-		select {
-		case <-parent.Done():
-			cancel()
-		case <-child.Done():
-		}
-	}()
-	return child
-}
+func (v valuesContext) Deadline() (time.Time, bool) { return time.Time{}, false }
+func (v valuesContext) Done() <-chan struct{}       { return nil }
+func (v valuesContext) Err() error                  { return nil }
+func (v valuesContext) Value(key any) any           { return v.parent.Value(key) }
 
 // truncateResponse limits the response length to avoid overwhelming context windows.
 func truncateResponse(s string, maxLen int) string {

internal/core/subagent_test.go

@@ -0,0 +1,115 @@
+package core
+
+import (
+	"context"
+	"testing"
+	"time"
+)
+
+func TestValuesContext_StripsDeadlineAndCancellation(t *testing.T) {
+	// Parent with a tight deadline.
+	parent, cancel := context.WithTimeout(context.Background(), 1*time.Millisecond)
+	defer cancel()
+	time.Sleep(5 * time.Millisecond) // Let deadline expire.
+
+	if parent.Err() == nil {
+		t.Fatal("expected parent to be expired")
+	}
+
+	vc := valuesContext{parent: parent}
+
+	if _, ok := vc.Deadline(); ok {
+		t.Error("valuesContext should report no deadline")
+	}
+	if vc.Done() != nil {
+		t.Error("valuesContext.Done() should return nil")
+	}
+	if vc.Err() != nil {
+		t.Errorf("valuesContext.Err() should be nil, got %v", vc.Err())
+	}
+}
+
+func TestValuesContext_PreservesValues(t *testing.T) {
+	type testKey struct{}
+	parent := context.WithValue(context.Background(), testKey{}, "hello")
+
+	vc := valuesContext{parent: parent}
+
+	got, ok := vc.Value(testKey{}).(string)
+	if !ok || got != "hello" {
+		t.Errorf("expected value 'hello', got %q (ok=%v)", got, ok)
+	}
+}
+
+func TestSpawnContext_SurvivesCancelledParent(t *testing.T) {
+	// Simulate the exact scenario from the bug: the parent generation
+	// context is already cancelled when the subagent tool handler runs.
+	parent, cancel := context.WithCancel(context.Background())
+	cancel() // Cancelled before detach.
+
+	// This is what executeSubagent now does:
+	spawnCtx := context.WithoutCancel(valuesContext{parent: parent})
+
+	// The spawn context must be alive.
+	if spawnCtx.Err() != nil {
+		t.Fatalf("spawnCtx should be alive, got err: %v", spawnCtx.Err())
+	}
+
+	// Adding a timeout should produce a working context.
+	tCtx, tCancel := context.WithTimeout(spawnCtx, 5*time.Second)
+	defer tCancel()
+
+	if tCtx.Err() != nil {
+		t.Fatalf("timeout context should be alive, got err: %v", tCtx.Err())
+	}
+}
+
+func TestSpawnContext_SurvivesDeadlineExceededParent(t *testing.T) {
+	// Simulate: parent had a deadline that already expired.
+	parent, pCancel := context.WithTimeout(context.Background(), 1*time.Millisecond)
+	defer pCancel()
+	time.Sleep(5 * time.Millisecond)
+
+	if parent.Err() != context.DeadlineExceeded {
+		t.Fatalf("expected parent deadline exceeded, got: %v", parent.Err())
+	}
+
+	spawnCtx := context.WithoutCancel(valuesContext{parent: parent})
+
+	if spawnCtx.Err() != nil {
+		t.Fatalf("spawnCtx should be alive after deadline-exceeded parent, got: %v", spawnCtx.Err())
+	}
+}
+
+func TestSpawnContext_PreservesSpawnerValue(t *testing.T) {
+	// Verify the subagent spawner callback survives context detachment.
+	called := false
+	spawner := SubagentSpawnFunc(func(ctx context.Context, toolCallID, prompt, model, systemPrompt string, timeout time.Duration) (*SubagentSpawnResult, error) {
+		called = true
+		return &SubagentSpawnResult{Response: "ok"}, nil
+	})
+
+	parent := WithSubagentSpawner(context.Background(), spawner)
+	// Cancel the parent.
+	parentCtx, cancel := context.WithCancel(parent)
+	cancel()
+
+	spawnCtx := context.WithoutCancel(valuesContext{parent: parentCtx})
+
+	// Should be able to retrieve the spawner from the detached context.
+	recovered := getSubagentSpawner(spawnCtx)
+	if recovered == nil {
+		t.Fatal("spawner should be recoverable from detached context")
+	}
+
+	result, err := recovered(spawnCtx, "tc1", "test task", "", "", time.Minute)
+	if err != nil {
+		t.Fatalf("spawner call failed: %v", err)
+	}
+	if !called {
+		t.Error("spawner was not called")
+	}
+	if result.Response != "ok" {
+		t.Errorf("expected 'ok', got %q", result.Response)
+	}
+}

internal/extbridge/extbridge.go

@@ -0,0 +1,97 @@
+// Package extbridge wires the public Kit SDK to the internal extensions
+// package. It exists so that cmd/ and internal/acpserver/ don't both
+// reimplement the same SDK→extension event/subagent conversions.
+package extbridge
+
+import (
+	"context"
+
+	"github.com/mark3labs/kit/internal/extensions"
+	kit "github.com/mark3labs/kit/pkg/kit"
+)
+
+// SDKEventToSubagentEvent converts an SDK [kit.Event] into the
+// extension-facing [extensions.SubagentEvent]. Returns a zero-value event
+// (Type=="") for events that don't map to anything useful — callers should
+// drop those.
+func SDKEventToSubagentEvent(e kit.Event) extensions.SubagentEvent {
+	switch ev := e.(type) {
+	case kit.MessageUpdateEvent:
+		return extensions.SubagentEvent{Type: "text", Content: ev.Chunk}
+	case kit.ReasoningDeltaEvent:
+		return extensions.SubagentEvent{Type: "reasoning", Content: ev.Delta}
+	case kit.ToolCallEvent:
+		return extensions.SubagentEvent{
+			Type: "tool_call", ToolCallID: ev.ToolCallID,
+			ToolName: ev.ToolName, ToolKind: ev.ToolKind, ToolArgs: ev.ToolArgs,
+		}
+	case kit.ToolExecutionStartEvent:
+		return extensions.SubagentEvent{
+			Type: "tool_execution_start", ToolCallID: ev.ToolCallID,
+			ToolName: ev.ToolName, ToolKind: ev.ToolKind,
+		}
+	case kit.ToolExecutionEndEvent:
+		return extensions.SubagentEvent{
+			Type: "tool_execution_end", ToolCallID: ev.ToolCallID,
+			ToolName: ev.ToolName, ToolKind: ev.ToolKind,
+		}
+	case kit.ToolResultEvent:
+		return extensions.SubagentEvent{
+			Type: "tool_result", ToolCallID: ev.ToolCallID,
+			ToolName: ev.ToolName, ToolKind: ev.ToolKind,
+			ToolResult: ev.Result, IsError: ev.IsError,
+		}
+	case kit.TurnStartEvent:
+		return extensions.SubagentEvent{Type: "turn_start"}
+	case kit.TurnEndEvent:
+		return extensions.SubagentEvent{Type: "turn_end"}
+	default:
+		return extensions.SubagentEvent{}
+	}
+}
+
+// SpawnSubagent runs a subagent in-process via the Kit SDK and translates
+// the result/events back into the extension-facing types. The returned
+// handle is always nil — the SDK path runs synchronously and does not
+// expose a separate process handle. Callers that need non-blocking
+// behaviour should run this in their own goroutine.
+//
+// This function consolidates the previously-duplicated wiring in
+// cmd/root.go (interactive + runtime contexts) and
+// internal/acpserver/session.go.
+func SpawnSubagent(ctx context.Context, k *kit.Kit, cfg extensions.SubagentConfig) (*extensions.SubagentHandle, *extensions.SubagentResult, error) {
+	sdkCfg := kit.SubagentConfig{
+		Prompt:       cfg.Prompt,
+		Model:        cfg.Model,
+		SystemPrompt: cfg.SystemPrompt,
+		Timeout:      cfg.Timeout,
+		NoSession:    cfg.NoSession,
+	}
+	if cfg.OnEvent != nil {
+		sdkCfg.OnEvent = func(e kit.Event) {
+			se := SDKEventToSubagentEvent(e)
+			if se.Type != "" {
+				cfg.OnEvent(se)
+			}
+		}
+	}
+
+	result, err := k.Subagent(ctx, sdkCfg)
+	if result == nil {
+		return nil, &extensions.SubagentResult{Error: err}, err
+	}
+
+	extResult := &extensions.SubagentResult{
+		Response:  result.Response,
+		Error:     err,
+		SessionID: result.SessionID,
+		Elapsed:   result.Elapsed,
+	}
+	if result.Usage != nil {
+		extResult.Usage = &extensions.SubagentUsage{
+			InputTokens:  result.Usage.InputTokens,
+			OutputTokens: result.Usage.OutputTokens,
+		}
+	}
+	return nil, extResult, err
+}

internal/extensions/api.go

@@ -77,6 +77,64 @@ type Context struct {
 	//   ctx.CancelAndSend("Stop what you're doing and focus on the tests")
 	CancelAndSend func(string)
 
+	// Abort cancels the current agent turn (if running) and clears the
+	// message queue. Unlike CancelAndSend, no new message is injected —
+	// the agent simply stops. Safe to call when idle (no-op).
+	//
+	// Example:
+	//
+	//   ctx.Abort()  // stop whatever the agent is doing
+	Abort func()
+
+	// IsIdle returns true when the agent is not processing a turn.
+	// Extensions can use this to decide whether to dispatch immediately
+	// or queue work for later.
+	//
+	// Example:
+	//
+	//   if ctx.IsIdle() {
+	//       ctx.SendMessage("start new task")
+	//   }
+	IsIdle func() bool
+
+	// Compact triggers context compaction, summarising older messages to
+	// free context window space. Returns an error if compaction cannot
+	// start (e.g. agent is busy or app is closed). The actual compaction
+	// runs asynchronously; use OnComplete/OnError callbacks in
+	// CompactConfig to observe the result.
+	//
+	// Example:
+	//
+	//   err := ctx.Compact(ext.CompactConfig{
+	//       OnComplete: func() { ctx.PrintInfo("Compaction done") },
+	//       OnError:    func(errMsg string) { ctx.PrintError("Compact failed: " + errMsg) },
+	//   })
+	Compact func(CompactConfig) error
+
+	// SendMultimodalMessage injects a message with file attachments (images,
+	// documents) into the conversation and triggers a new agent turn. Files
+	// are described by FilePart structs containing the raw bytes, filename,
+	// and MIME type. If the agent is busy the message is queued.
+	//
+	// Example:
+	//
+	//   data, _ := os.ReadFile("photo.jpg")
+	//   ctx.SendMultimodalMessage("Describe this image", []ext.FilePart{
+	//       {Filename: "photo.jpg", Data: data, MediaType: "image/jpeg"},
+	//   })
+	SendMultimodalMessage func(text string, files []FilePart)
+
+	// GetSessionUsage returns aggregated token usage and cost statistics
+	// for the current session. This includes total input/output tokens,
+	// cache read/write tokens, total cost, and request count.
+	//
+	// Example:
+	//
+	//   usage := ctx.GetSessionUsage()
+	//   fmt.Sprintf("Tokens: ↑%d ↓%d Cost: $%.3f",
+	//       usage.TotalInputTokens, usage.TotalOutputTokens, usage.TotalCost)
+	GetSessionUsage func() SessionUsage
+
 	// SetWidget places or updates a persistent widget in the TUI. Widgets
 	// remain visible across agent turns until explicitly removed. The
 	// widget is identified by WidgetConfig.ID; calling SetWidget with the
@@ -860,7 +918,7 @@ type ExtensionEntry struct {
 type ContextMessage struct {
 	// Index is the position of this message in the original context array
 	// (0-based). When returning messages from a ContextPrepareResult,
-	// messages with Index >= 0 reuse the original fantasy.Message at that
+	// messages with Index >= 0 reuse the original LLM message at that
 	// position (preserving tool calls, reasoning, and other complex parts).
 	// Set Index to -1 for newly injected messages (created from Role + Content).
 	Index int
@@ -937,6 +995,48 @@ type StatusBarEntry struct {
 	Priority int
 }
 
+// CompactConfig configures a programmatic context compaction request.
+type CompactConfig struct {
+	// CustomInstructions is optional text appended to the summary prompt
+	// (e.g. "Focus on the API design decisions"). Empty uses the default.
+	CustomInstructions string
+	// OnComplete is called when compaction finishes successfully.
+	// May be nil if the caller doesn't need notification.
+	OnComplete func()
+	// OnError is called when compaction fails. The argument is the error message.
+	// May be nil if the caller doesn't need notification.
+	OnError func(errMsg string)
+}
+
+// FilePart describes a file attachment for multimodal messages. Extensions
+// use this with SendMultimodalMessage to attach images or documents.
+type FilePart struct {
+	// Filename is the name of the file (e.g. "photo.jpg").
+	Filename string
+	// Data is the raw file content.
+	Data []byte
+	// MediaType is the MIME type (e.g. "image/jpeg", "application/pdf").
+	MediaType string
+}
+
+// SessionUsage contains aggregated token usage and cost statistics for
+// the current session. Extensions use this with GetSessionUsage() to
+// report usage information.
+type SessionUsage struct {
+	// TotalInputTokens is the sum of input tokens across all requests.
+	TotalInputTokens int
+	// TotalOutputTokens is the sum of output tokens across all requests.
+	TotalOutputTokens int
+	// TotalCacheReadTokens is the sum of cache read tokens.
+	TotalCacheReadTokens int
+	// TotalCacheWriteTokens is the sum of cache write tokens.
+	TotalCacheWriteTokens int
+	// TotalCost is the total cost in USD across all requests.
+	TotalCost float64
+	// RequestCount is the number of LLM requests made in this session.
+	RequestCount int
+}
+
 // PrintBlockOpts configures a custom styled block for PrintBlock.
 type PrintBlockOpts struct {
 	// Text is the main content to display.
@@ -963,6 +1063,9 @@ type PrintBlockOpts struct {
 type API struct {
 	// Event-specific registration functions (wired by the loader).
 	onToolCall                func(func(ToolCallEvent, Context) *ToolCallResult)
+	onToolCallInputStart      func(func(ToolCallInputStartEvent, Context))
+	onToolCallInputDelta      func(func(ToolCallInputDeltaEvent, Context))
+	onToolCallInputEnd        func(func(ToolCallInputEndEvent, Context))
 	onToolExecStart           func(func(ToolExecutionStartEvent, Context))
 	onToolExecEnd             func(func(ToolExecutionEndEvent, Context))
 	onToolOutput              func(func(ToolOutputEvent, Context))
@@ -991,6 +1094,14 @@ type API struct {
 	onSubagentStart           func(func(SubagentStartEvent, Context))
 	onSubagentChunk           func(func(SubagentChunkEvent, Context))
 	onSubagentEnd             func(func(SubagentEndEvent, Context))
+	onStepStart               func(func(StepStartEvent, Context))
+	onStepFinish              func(func(StepFinishEvent, Context))
+	onReasoningStart          func(func(ReasoningStartEvent, Context))
+	onWarnings                func(func(WarningsEvent, Context))
+	onSource                  func(func(SourceEvent, Context))
+	onError                   func(func(ErrorEvent, Context))
+	onRetry                   func(func(RetryEvent, Context))
+	onPrepareStep             func(func(PrepareStepEvent, Context) *PrepareStepResult)
 }
 
 // OnToolCall registers a handler that fires before a tool executes.
@@ -999,6 +1110,26 @@ func (a *API) OnToolCall(handler func(ToolCallEvent, Context) *ToolCallResult) {
 	a.onToolCall(handler)
 }
 
+// OnToolCallInputStart registers a handler that fires when the LLM begins
+// generating tool call arguments. The tool name is known but the full
+// argument JSON is still being streamed. Useful for showing a "running"
+// indicator immediately without waiting for the full arguments.
+func (a *API) OnToolCallInputStart(handler func(ToolCallInputStartEvent, Context)) {
+	a.onToolCallInputStart(handler)
+}
+
+// OnToolCallInputDelta registers a handler that fires for each streamed
+// fragment of tool call arguments as they arrive from the LLM.
+func (a *API) OnToolCallInputDelta(handler func(ToolCallInputDeltaEvent, Context)) {
+	a.onToolCallInputDelta(handler)
+}
+
+// OnToolCallInputEnd registers a handler that fires when tool argument
+// streaming is complete, before the tool call is parsed and execution begins.
+func (a *API) OnToolCallInputEnd(handler func(ToolCallInputEndEvent, Context)) {
+	a.onToolCallInputEnd(handler)
+}
+
 // OnToolExecutionStart registers a handler for tool execution start.
 func (a *API) OnToolExecutionStart(handler func(ToolExecutionStartEvent, Context)) {
 	a.onToolExecStart(handler)
@@ -1178,6 +1309,56 @@ func (a *API) OnBeforeCompact(handler func(BeforeCompactEvent, Context) *BeforeC
 	a.onBeforeCompact(handler)
 }
 
+// OnStepStart registers a handler that fires when a new LLM call begins
+// within a multi-step agent turn.
+func (a *API) OnStepStart(handler func(StepStartEvent, Context)) {
+	a.onStepStart(handler)
+}
+
+// OnStepFinish registers a handler that fires when a step completes,
+// providing step number, finish reason, and decomposed token usage.
+func (a *API) OnStepFinish(handler func(StepFinishEvent, Context)) {
+	a.onStepFinish(handler)
+}
+
+// OnReasoningStart registers a handler that fires when the LLM begins
+// reasoning/thinking.
+func (a *API) OnReasoningStart(handler func(ReasoningStartEvent, Context)) {
+	a.onReasoningStart(handler)
+}
+
+// OnWarnings registers a handler that fires when the LLM provider returns
+// warnings about the request.
+func (a *API) OnWarnings(handler func(WarningsEvent, Context)) {
+	a.onWarnings(handler)
+}
+
+// OnSource registers a handler that fires when the LLM references a source
+// (e.g. from web search tools).
+func (a *API) OnSource(handler func(SourceEvent, Context)) {
+	a.onSource(handler)
+}
+
+// OnError registers a handler that fires when an agent-level error occurs
+// during streaming.
+func (a *API) OnError(handler func(ErrorEvent, Context)) {
+	a.onError(handler)
+}
+
+// OnRetry registers a handler that fires when the LLM provider request is
+// retried after a transient error.
+func (a *API) OnRetry(handler func(RetryEvent, Context)) {
+	a.onRetry(handler)
+}
+
+// OnPrepareStep registers a handler that fires between steps within a
+// multi-step agent turn, after steering messages are injected and before
+// messages are sent to the LLM. Return a non-nil PrepareStepResult with
+// Messages to replace the context window for this step.
+func (a *API) OnPrepareStep(handler func(PrepareStepEvent, Context) *PrepareStepResult) {
+	a.onPrepareStep(handler)
+}
+
 // RegisterToolRenderer registers a custom renderer for a specific tool's
 // display in the TUI. The renderer controls the header (parameter summary)
 // and/or body (result display) of the tool's output block. If multiple
@@ -1790,6 +1971,34 @@ type ToolCallResult struct {
 
 func (ToolCallResult) isResult() {}
 
+// ToolCallInputStartEvent fires when the LLM begins generating tool call
+// arguments. The tool name is known but the full argument JSON is still
+// being streamed.
+type ToolCallInputStartEvent struct {
+	ToolCallID string
+	ToolName   string
+	ToolKind   string // Tool classification: "execute", "edit", "read", "search", "agent"
+}
+
+func (e ToolCallInputStartEvent) Type() EventType { return ToolCallInputStart }
+
+// ToolCallInputDeltaEvent fires for each streamed fragment of tool call
+// arguments as they arrive from the LLM.
+type ToolCallInputDeltaEvent struct {
+	ToolCallID string
+	Delta      string // JSON fragment of tool arguments
+}
+
+func (e ToolCallInputDeltaEvent) Type() EventType { return ToolCallInputDelta }
+
+// ToolCallInputEndEvent fires when tool argument streaming is complete,
+// before the tool call is parsed and execution begins.
+type ToolCallInputEndEvent struct {
+	ToolCallID string
+}
+
+func (e ToolCallInputEndEvent) Type() EventType { return ToolCallInputEnd }
+
 // ToolExecutionStartEvent fires when a tool begins executing.
 type ToolExecutionStartEvent struct {
 	ToolCallID string
@@ -2102,6 +2311,98 @@ type SubagentEndEvent struct {
 
 func (e SubagentEndEvent) Type() EventType { return SubagentEnd }
 
+// ---------------------------------------------------------------------------
+// Step lifecycle events (exposed to Yaegi — concrete structs)
+// ---------------------------------------------------------------------------
+
+// StepStartEvent fires when a new LLM call begins within a multi-step agent turn.
+type StepStartEvent struct {
+	StepNumber int
+}
+
+func (e StepStartEvent) Type() EventType { return StepStart }
+
+// StepFinishEvent fires when a step completes, providing step metadata and
+// token usage. Usage fields are plain int64 (not LLMUsage) because Yaegi
+// cannot handle fantasy types across the interpreter boundary.
+type StepFinishEvent struct {
+	StepNumber       int
+	HasToolCalls     bool
+	FinishReason     string
+	InputTokens      int64
+	OutputTokens     int64
+	CacheReadTokens  int64
+	CacheWriteTokens int64
+}
+
+func (e StepFinishEvent) Type() EventType { return StepFinish }
+
+// ReasoningStartEvent fires when the LLM begins reasoning/thinking.
+type ReasoningStartEvent struct {
+	ID string
+}
+
+func (e ReasoningStartEvent) Type() EventType { return ReasoningStart }
+
+// WarningsEvent fires when the LLM provider returns warnings about the request.
+type WarningsEvent struct {
+	Warnings []string
+}
+
+func (e WarningsEvent) Type() EventType { return Warnings }
+
+// SourceEvent fires when the LLM references a source (e.g. from web search).
+type SourceEvent struct {
+	SourceType string
+	ID         string
+	URL        string
+	Title      string
+}
+
+func (e SourceEvent) Type() EventType { return Source }
+
+// ErrorEvent fires when an agent-level error occurs during streaming.
+// Uses string instead of error because Yaegi cannot handle the error
+// interface reliably across the interpreter boundary.
+type ErrorEvent struct {
+	Error string
+}
+
+func (e ErrorEvent) Type() EventType { return Error }
+
+// RetryEvent fires when the LLM provider request is retried after a
+// transient error.
+type RetryEvent struct {
+	Attempt int
+	Error   string
+}
+
+func (e RetryEvent) Type() EventType { return Retry }
+
+// PrepareStepEvent fires between steps within a multi-step agent turn,
+// after steering messages are injected and before messages are sent to
+// the LLM. Handlers can inspect and replace the context window.
+type PrepareStepEvent struct {
+	// StepNumber is the zero-based step index within the current turn.
+	StepNumber int
+	// Messages is the current context window that will be sent to the LLM.
+	Messages []ContextMessage
+}
+
+func (e PrepareStepEvent) Type() EventType { return PrepareStep }
+
+// PrepareStepResult allows extensions to replace the context window between
+// steps. Return nil Messages to leave the context unchanged.
+type PrepareStepResult struct {
+	// Messages replaces the entire context window for this step. If nil,
+	// the original messages are used unchanged. Messages with a non-negative
+	// Index reuse the original message at that position; messages with
+	// Index < 0 are created fresh from Role + Content.
+	Messages []ContextMessage
+}
+
+func (PrepareStepResult) isResult() {}
+
 // ThemeColor is an adaptive color pair with light and dark hex values.
 // Either field may be empty to inherit from the default theme.
 type ThemeColor struct {

internal/extensions/events.go

@@ -13,6 +13,19 @@ const (
 	// ToolCall fires before a tool executes. Handlers can block execution.
 	ToolCall EventType = "tool_call"
 
+	// ToolCallInputStart fires when the LLM begins generating tool call
+	// arguments. The tool name is known but the full argument JSON is still
+	// being streamed.
+	ToolCallInputStart EventType = "tool_call_input_start"
+
+	// ToolCallInputDelta fires for each streamed fragment of tool call
+	// arguments as they arrive from the LLM.
+	ToolCallInputDelta EventType = "tool_call_input_delta"
+
+	// ToolCallInputEnd fires when tool argument streaming is complete,
+	// before the tool call is parsed and execution begins.
+	ToolCallInputEnd EventType = "tool_call_input_end"
+
 	// ToolExecutionStart fires when a tool begins executing.
 	ToolExecutionStart EventType = "tool_execution_start"
 
@@ -83,18 +96,50 @@ const (
 	// SubagentEnd fires when a subagent tool call completes (success
 	// or error). Carries the final response and any error message.
 	SubagentEnd EventType = "subagent_end"
+
+	// StepStart fires when a new LLM call begins within a multi-step
+	// agent turn.
+	StepStart EventType = "step_start"
+
+	// StepFinish fires when a step completes, providing step number,
+	// finish reason, and token usage.
+	StepFinish EventType = "step_finish"
+
+	// ReasoningStart fires when the LLM begins reasoning/thinking.
+	ReasoningStart EventType = "reasoning_start"
+
+	// Warnings fires when the LLM provider returns warnings.
+	Warnings EventType = "warnings"
+
+	// Source fires when the LLM references a source (e.g. web search).
+	Source EventType = "source"
+
+	// Error fires when an agent-level error occurs during streaming.
+	Error EventType = "error"
+
+	// Retry fires when the LLM provider request is retried after a
+	// transient error.
+	Retry EventType = "retry"
+
+	// PrepareStep fires between steps within a multi-step agent turn,
+	// after steering messages are injected and before messages are sent
+	// to the LLM. Handlers can replace the context window for this step.
+	PrepareStep EventType = "prepare_step"
 )
 
 // AllEventTypes returns every supported event type.
 func AllEventTypes() []EventType {
 	return []EventType{
-		ToolCall, ToolExecutionStart, ToolExecutionEnd, ToolResult,
+		ToolCall, ToolCallInputStart, ToolCallInputDelta, ToolCallInputEnd,
+		ToolExecutionStart, ToolExecutionEnd, ToolResult,
 		Input, BeforeAgentStart, AgentStart, AgentEnd,
 		MessageStart, MessageUpdate, MessageEnd,
 		SessionStart, SessionShutdown,
 		ModelChange, ContextPrepare,
 		BeforeFork, BeforeSessionSwitch, BeforeCompact,
 		SubagentStart, SubagentChunk, SubagentEnd,
+		StepStart, StepFinish, ReasoningStart, Warnings, Source, Error, Retry,
+		PrepareStep,
 	}
 }
 

internal/extensions/events_test.go

@@ -4,8 +4,8 @@ import "testing"
 
 func TestAllEventTypes_Count(t *testing.T) {
 	all := AllEventTypes()
-	if len(all) != 21 {
-		t.Fatalf("expected 21 event types, got %d", len(all))
+	if len(all) != 32 {
+		t.Fatalf("expected 32 event types, got %d", len(all))
 	}
 }
 
@@ -38,6 +38,9 @@ func TestEventType_TypeMethod(t *testing.T) {
 		want  EventType
 	}{
 		{ToolCallEvent{ToolName: "test"}, ToolCall},
+		{ToolCallInputStartEvent{ToolCallID: "x", ToolName: "test"}, ToolCallInputStart},
+		{ToolCallInputDeltaEvent{ToolCallID: "x", Delta: "{"}, ToolCallInputDelta},
+		{ToolCallInputEndEvent{ToolCallID: "x"}, ToolCallInputEnd},
 		{ToolExecutionStartEvent{ToolName: "test"}, ToolExecutionStart},
 		{ToolExecutionEndEvent{ToolName: "test"}, ToolExecutionEnd},
 		{ToolResultEvent{ToolName: "test"}, ToolResult},

internal/extensions/installer.go

@@ -154,6 +154,11 @@ func NewInstaller(projectDir string) *Installer {
 
 // Install clones a git repository to the appropriate scope.
 func (i *Installer) Install(source *GitSource, scope InstallScope) error {
+	return i.install(source, scope, nil)
+}
+
+// install is the internal implementation that supports optional include paths.
+func (i *Installer) install(source *GitSource, scope InstallScope, includePaths []string) error {
 	targetDir := i.getInstallPath(source, scope)
 
 	// Check if already installed
@@ -199,6 +204,7 @@ func (i *Installer) Install(source *GitSource, scope InstallScope) error {
 		Pinned:    source.Pinned,
 		Scope:     scope,
 		Installed: time.Now(),
+		Include:   includePaths,
 	}
 	if err := i.addToManifest(entry, scope); err != nil {
 		// Don't fail the install, just log the error
@@ -268,7 +274,22 @@ func (i *Installer) Update(source *GitSource, scope InstallScope) error {
 	cleanCmd.Dir = targetDir
 	_ = cleanCmd.Run() // Ignore errors - clean is best effort
 
-	// Update manifest timestamp
+	// Update manifest timestamp, preserving existing fields like Include
+	existing, _ := i.loadManifest(scope)
+	var include []string
+	var installed time.Time
+	if existing != nil {
+		for _, p := range existing.Packages {
+			if p.Host+"/"+p.Path == source.Identity() {
+				include = p.Include
+				installed = p.Installed
+				break
+			}
+		}
+	}
+	if installed.IsZero() {
+		installed = time.Now()
+	}
 	entry := ManifestEntry{
 		Source:    source.String(),
 		Repo:      source.Repo,
@@ -277,8 +298,9 @@ func (i *Installer) Update(source *GitSource, scope InstallScope) error {
 		Ref:       "",
 		Pinned:    false,
 		Scope:     scope,
-		Installed: time.Now(),
+		Installed: installed,
 		Updated:   time.Now(),
+		Include:   include,
 	}
 	_ = i.addToManifest(entry, scope) // Best effort - don't fail update if manifest fails
 
@@ -428,25 +450,6 @@ func globalGitInstallRoot() string {
 	return filepath.Join(base, "kit", "git")
 }
 
-// GetInstalledPackages returns all installed packages from both scopes.
-func (i *Installer) GetInstalledPackages() ([]ManifestEntry, error) {
-	var all []ManifestEntry
-
-	global, err := i.loadManifest(ScopeGlobal)
-	if err != nil {
-		return nil, fmt.Errorf("loading global manifest: %w", err)
-	}
-	all = append(all, global.Packages...)
-
-	project, err := i.loadManifest(ScopeProject)
-	if err != nil {
-		return nil, fmt.Errorf("loading project manifest: %w", err)
-	}
-	all = append(all, project.Packages...)
-
-	return all, nil
-}
-
 // IsInstalled checks if a package is installed in either scope.
 // Returns (scope, true) if installed, ("", false) otherwise.
 func (i *Installer) IsInstalled(source *GitSource) (InstallScope, bool) {
@@ -503,30 +506,7 @@ func (i *Installer) PreviewExtensions(source *GitSource) ([]ExtensionPreview, st
 // InstallWithInclude clones a repo and installs only the specified extensions.
 // includePaths are relative paths like "./git/main.go" - if empty, installs all.
 func (i *Installer) InstallWithInclude(source *GitSource, scope InstallScope, includePaths []string) error {
-	// First, do a regular install
-	if err := i.Install(source, scope); err != nil {
-		return err
-	}
-
-	// If specific includes were requested, update the manifest
-	if len(includePaths) > 0 {
-		entry := ManifestEntry{
-			Source:  source.String(),
-			Repo:    source.Repo,
-			Host:    source.Host,
-			Path:    source.Path,
-			Ref:     source.Ref,
-			Pinned:  source.Pinned,
-			Scope:   scope,
-			Include: includePaths,
-		}
-
-		if err := addEntryToManifest(entry, scope); err != nil {
-			return fmt.Errorf("updating manifest with includes: %w", err)
-		}
-	}
-
-	return nil
+	return i.install(source, scope, includePaths)
 }
 
 // CleanupTempDir removes a temporary directory used for preview.

internal/extensions/installer_test.go

@@ -245,14 +245,21 @@ func TestManifestEntryIdentity(t *testing.T) {
 	}
 }
 
+// TestLoadAndSaveManifest exercises the live *Installer.loadManifest /
+// saveManifest round-trip against a temp directory, ensuring an absent
+// manifest loads as empty and a saved manifest reads back identically.
 func TestLoadAndSaveManifest(t *testing.T) {
 	tempDir := t.TempDir()
+	installer := &Installer{
+		projectGitRoot: tempDir,
+		globalGitRoot:  tempDir,
+	}
 	manifestPath := filepath.Join(tempDir, "packages.json")
 
 	// Test loading non-existent manifest
-	manifest, err := loadManifestFromPath(manifestPath)
+	manifest, err := installer.loadManifest(ScopeGlobal)
 	if err != nil {
-		t.Fatalf("loadManifestFromPath() error = %v", err)
+		t.Fatalf("loadManifest() error = %v", err)
 	}
 	if len(manifest.Packages) != 0 {
 		t.Errorf("Expected empty packages, got %d", len(manifest.Packages))
@@ -273,15 +280,20 @@ func TestLoadAndSaveManifest(t *testing.T) {
 	}
 
 	// Save it
-	err = saveManifestToPath(manifest, manifestPath)
+	err = installer.saveManifest(manifest, ScopeGlobal)
 	if err != nil {
-		t.Fatalf("saveManifestToPath() error = %v", err)
+		t.Fatalf("saveManifest() error = %v", err)
+	}
+
+	// Verify it was written to expected path
+	if _, err := os.Stat(manifestPath); err != nil {
+		t.Fatalf("manifest file not created: %v", err)
 	}
 
 	// Load it back
-	loaded, err := loadManifestFromPath(manifestPath)
+	loaded, err := installer.loadManifest(ScopeGlobal)
 	if err != nil {
-		t.Fatalf("loadManifestFromPath() error = %v", err)
+		t.Fatalf("loadManifest() error = %v", err)
 	}
 	if len(loaded.Packages) != 1 {
 		t.Errorf("Expected 1 package, got %d", len(loaded.Packages))
@@ -291,21 +303,15 @@ func TestLoadAndSaveManifest(t *testing.T) {
 	}
 }
 
+// TestAddAndRemoveFromManifest verifies that *Installer.addToManifest
+// followed by removeFromManifest leaves the manifest in its original
+// (empty) state, using a temp-directory installer scope.
 func TestAddAndRemoveFromManifest(t *testing.T) {
 	tempDir := t.TempDir()
-
-	// Set up environment for manifest path
-	if err := os.Setenv("XDG_DATA_HOME", tempDir); err != nil {
-		t.Fatalf("Setenv() error = %v", err)
+	installer := &Installer{
+		projectGitRoot: tempDir,
+		globalGitRoot:  tempDir,
 	}
-	defer func() {
-		if err := os.Unsetenv("XDG_DATA_HOME"); err != nil {
-			t.Logf("Unsetenv() error = %v", err)
-		}
-	}()
-
-	// The manifest path when XDG_DATA_HOME is set
-	manifestPath := filepath.Join(tempDir, "kit", "git", "packages.json")
 
 	// Add an entry
 	entry := ManifestEntry{
@@ -315,58 +321,51 @@ func TestAddAndRemoveFromManifest(t *testing.T) {
 		Scope:  ScopeGlobal,
 	}
 
-	err := addEntryToManifest(entry, ScopeGlobal)
-	if err != nil {
-		t.Fatalf("addEntryToManifest() error = %v", err)
+	if err := installer.addToManifest(entry, ScopeGlobal); err != nil {
+		t.Fatalf("addToManifest() error = %v", err)
 	}
 
 	// Verify it was added
-	manifest, err := loadManifestFromPath(manifestPath)
+	manifest, err := installer.loadManifest(ScopeGlobal)
 	if err != nil {
-		t.Fatalf("loadManifestFromPath() error = %v", err)
+		t.Fatalf("loadManifest() error = %v", err)
 	}
 	if len(manifest.Packages) != 1 {
 		t.Errorf("Expected 1 package, got %d", len(manifest.Packages))
 	}
 
 	// Remove it
-	err = removeEntryFromManifest("github.com/user/repo", ScopeGlobal)
-	if err != nil {
-		t.Fatalf("removeEntryFromManifest() error = %v", err)
+	if err := installer.removeFromManifest("github.com/user/repo", ScopeGlobal); err != nil {
+		t.Fatalf("removeFromManifest() error = %v", err)
 	}
 
 	// Verify it was removed
-	manifest, err = loadManifestFromPath(manifestPath)
+	manifest, err = installer.loadManifest(ScopeGlobal)
 	if err != nil {
-		t.Fatalf("loadManifestFromPath() error = %v", err)
+		t.Fatalf("loadManifest() error = %v", err)
 	}
 	if len(manifest.Packages) != 0 {
 		t.Errorf("Expected 0 packages, got %d", len(manifest.Packages))
 	}
 }
 
+// TestFindInManifest writes a manifest file directly to the path
+// resolved by the package-level manifestPathForScope helper and then
+// confirms FindInManifest locates the entry by identity (and returns
+// nil for a non-existent identity).
 func TestFindInManifest(t *testing.T) {
 	tempDir := t.TempDir()
-	if err := os.Setenv("XDG_DATA_HOME", tempDir); err != nil {
-		t.Fatalf("Setenv() error = %v", err)
-	}
-	defer func() {
-		if err := os.Unsetenv("XDG_DATA_HOME"); err != nil {
-			t.Logf("Unsetenv() error = %v", err)
-		}
-	}()
+	t.Setenv("XDG_DATA_HOME", tempDir)
 
-	// Add an entry to global manifest
-	entry := ManifestEntry{
-		Source: "git:github.com/user/repo",
-		Host:   "github.com",
-		Path:   "user/repo",
-		Scope:  ScopeGlobal,
+	// Write a manifest entry directly via the package-level path resolver
+	// so FindInManifest (which uses manifestPathForScope) can read it back.
+	manifestPath := manifestPathForScope(ScopeGlobal)
+	if err := os.MkdirAll(filepath.Dir(manifestPath), 0755); err != nil {
+		t.Fatalf("MkdirAll() error = %v", err)
 	}
-
-	err := addEntryToManifest(entry, ScopeGlobal)
-	if err != nil {
-		t.Fatalf("addEntryToManifest() error = %v", err)
+	data := []byte(`{"packages":[{"source":"git:github.com/user/repo","repo":"","host":"github.com","path":"user/repo","pinned":false,"scope":"global","installed":"0001-01-01T00:00:00Z"}]}`)
+	if err := os.WriteFile(manifestPath, data, 0644); err != nil {
+		t.Fatalf("WriteFile() error = %v", err)
 	}
 
 	// Find it

internal/extensions/loader.go

@@ -34,15 +34,10 @@ func LoadExtensions(extraPaths []string) ([]LoadedExtension, error) {
 	for _, p := range paths {
 		ext, err := loadSingleExtension(p)
 		if err != nil {
-			log.Warn("skipping extension", "path", p, "err", err)
 			continue
 		}
 		loaded = append(loaded, *ext)
-		log.Debug("loaded extension", "path", p,
-			"handlers", countHandlers(ext),
-			"tools", len(ext.Tools),
-			"commands", len(ext.Commands),
-			"tool_renderers", len(ext.ToolRenderers))
+		log.Debug("loaded extension", "path", p, "handlers", countHandlers(ext), "tools", len(ext.Tools), "commands", len(ext.Commands), "tool_renderers", len(ext.ToolRenderers))
 	}
 	return loaded, nil
 }
@@ -133,7 +128,7 @@ func findExtensionsInDir(dir string) []string {
 
 	for _, entry := range entries {
 		full := filepath.Join(dir, entry.Name())
-		if !entry.IsDir() && strings.HasSuffix(entry.Name(), ".go") {
+		if !entry.IsDir() && strings.HasSuffix(entry.Name(), ".go") && !strings.HasSuffix(entry.Name(), "_test.go") {
 			results = append(results, full)
 		} else if entry.IsDir() {
 			main := filepath.Join(full, "main.go")
@@ -190,9 +185,13 @@ func findExtensionsInRepo(repoPath string) []string {
 			isExtDir := base == "extensions" || base == "ext" ||
 				strings.HasSuffix(base, "-extensions") || strings.HasSuffix(base, "-ext")
 
-			isExamplesSubdir := relPath == "examples" || strings.HasPrefix(relPath, "examples/")
+			// Allow walking into examples/ so we can reach examples/extensions/ etc,
+			// but don't treat examples/ itself or non-extension subdirs as extension locations.
+			if relPath == "examples" {
+				return nil
+			}
 
-			if !isExtDir && !isExamplesSubdir {
+			if !isExtDir {
 				mainPath := filepath.Join(path, "main.go")
 				if _, err := os.Stat(mainPath); err == nil {
 					if relPath == base { // Top-level directory
@@ -202,13 +201,6 @@ func findExtensionsInRepo(repoPath string) []string {
 						}
 						return filepath.SkipDir
 					}
-					if isExamplesSubdir || isExtDir {
-						if !multiFileDirs[relPath] {
-							multiFileDirs[relPath] = true
-							results = append(results, mainPath)
-						}
-						return filepath.SkipDir
-					}
 				}
 				return filepath.SkipDir
 			}
@@ -227,7 +219,7 @@ func findExtensionsInRepo(repoPath string) []string {
 		}
 
 		// It's a file
-		if !strings.HasSuffix(info.Name(), ".go") {
+		if !strings.HasSuffix(info.Name(), ".go") || strings.HasSuffix(info.Name(), "_test.go") {
 			return nil
 		}
 
@@ -437,6 +429,24 @@ func loadSingleExtension(path string) (*LoadedExtension, error) {
 				return *r
 			})
 		},
+		onToolCallInputStart: func(h func(ToolCallInputStartEvent, Context)) {
+			reg(ToolCallInputStart, func(e Event, c Context) Result {
+				h(e.(ToolCallInputStartEvent), c)
+				return nil
+			})
+		},
+		onToolCallInputDelta: func(h func(ToolCallInputDeltaEvent, Context)) {
+			reg(ToolCallInputDelta, func(e Event, c Context) Result {
+				h(e.(ToolCallInputDeltaEvent), c)
+				return nil
+			})
+		},
+		onToolCallInputEnd: func(h func(ToolCallInputEndEvent, Context)) {
+			reg(ToolCallInputEnd, func(e Event, c Context) Result {
+				h(e.(ToolCallInputEndEvent), c)
+				return nil
+			})
+		},
 		onToolExecStart: func(h func(ToolExecutionStartEvent, Context)) {
 			reg(ToolExecutionStart, func(e Event, c Context) Result {
 				h(e.(ToolExecutionStartEvent), c)
@@ -608,6 +618,57 @@ func loadSingleExtension(path string) (*LoadedExtension, error) {
 				return nil
 			})
 		},
+		onStepStart: func(h func(StepStartEvent, Context)) {
+			reg(StepStart, func(e Event, c Context) Result {
+				h(e.(StepStartEvent), c)
+				return nil
+			})
+		},
+		onStepFinish: func(h func(StepFinishEvent, Context)) {
+			reg(StepFinish, func(e Event, c Context) Result {
+				h(e.(StepFinishEvent), c)
+				return nil
+			})
+		},
+		onReasoningStart: func(h func(ReasoningStartEvent, Context)) {
+			reg(ReasoningStart, func(e Event, c Context) Result {
+				h(e.(ReasoningStartEvent), c)
+				return nil
+			})
+		},
+		onWarnings: func(h func(WarningsEvent, Context)) {
+			reg(Warnings, func(e Event, c Context) Result {
+				h(e.(WarningsEvent), c)
+				return nil
+			})
+		},
+		onSource: func(h func(SourceEvent, Context)) {
+			reg(Source, func(e Event, c Context) Result {
+				h(e.(SourceEvent), c)
+				return nil
+			})
+		},
+		onError: func(h func(ErrorEvent, Context)) {
+			reg(Error, func(e Event, c Context) Result {
+				h(e.(ErrorEvent), c)
+				return nil
+			})
+		},
+		onRetry: func(h func(RetryEvent, Context)) {
+			reg(Retry, func(e Event, c Context) Result {
+				h(e.(RetryEvent), c)
+				return nil
+			})
+		},
+		onPrepareStep: func(h func(PrepareStepEvent, Context) *PrepareStepResult) {
+			reg(PrepareStep, func(e Event, c Context) Result {
+				r := h(e.(PrepareStepEvent), c)
+				if r == nil {
+					return nil
+				}
+				return *r
+			})
+		},
 	}
 
 	// Call Init — the extension registers its handlers, tools, commands.

internal/extensions/manifest.go

@@ -72,30 +72,6 @@ func loadManifestFromPath(path string) (*Manifest, error) {
 	return &manifest, nil
 }
 
-// saveManifestToScope saves the manifest to the given scope.
-func saveManifestToScope(manifest *Manifest, scope InstallScope) error {
-	path := manifestPathForScope(scope)
-	return saveManifestToPath(manifest, path)
-}
-
-// saveManifestToPath saves a manifest to a specific file path.
-func saveManifestToPath(manifest *Manifest, path string) error {
-	if err := os.MkdirAll(filepath.Dir(path), 0755); err != nil {
-		return fmt.Errorf("creating manifest directory: %w", err)
-	}
-
-	data, err := json.MarshalIndent(manifest, "", "  ")
-	if err != nil {
-		return fmt.Errorf("encoding manifest: %w", err)
-	}
-
-	if err := os.WriteFile(path, data, 0644); err != nil {
-		return fmt.Errorf("writing manifest: %w", err)
-	}
-
-	return nil
-}
-
 // manifestPathForScope returns the manifest file path for a scope.
 func manifestPathForScope(scope InstallScope) string {
 	if scope == ScopeProject {
@@ -113,55 +89,6 @@ func manifestPathForScope(scope InstallScope) string {
 	return filepath.Join(base, "kit", "git", "packages.json")
 }
 
-// GetGlobalManifest returns the global manifest.
-func GetGlobalManifest() (*Manifest, error) {
-	return loadManifestFromScope(ScopeGlobal)
-}
-
-// GetProjectManifest returns the project manifest.
-func GetProjectManifest() (*Manifest, error) {
-	return loadManifestFromScope(ScopeProject)
-}
-
-// addEntryToManifest adds or replaces an entry in the manifest for a scope.
-func addEntryToManifest(entry ManifestEntry, scope InstallScope) error {
-	manifest, err := loadManifestFromScope(scope)
-	if err != nil {
-		return err
-	}
-
-	// Remove any existing entry with same identity
-	identity := entry.Identity()
-	filtered := make([]ManifestEntry, 0, len(manifest.Packages))
-	for _, p := range manifest.Packages {
-		if p.Identity() != identity {
-			filtered = append(filtered, p)
-		}
-	}
-	filtered = append(filtered, entry)
-	manifest.Packages = filtered
-
-	return saveManifestToScope(manifest, scope)
-}
-
-// removeEntryFromManifest removes an entry by identity from the manifest for a scope.
-func removeEntryFromManifest(identity string, scope InstallScope) error {
-	manifest, err := loadManifestFromScope(scope)
-	if err != nil {
-		return err
-	}
-
-	filtered := make([]ManifestEntry, 0, len(manifest.Packages))
-	for _, p := range manifest.Packages {
-		if p.Identity() != identity {
-			filtered = append(filtered, p)
-		}
-	}
-	manifest.Packages = filtered
-
-	return saveManifestToScope(manifest, scope)
-}
-
 // FindInManifest finds an entry by identity in either global or project manifest.
 // Returns the entry and its scope, or nil if not found.
 func FindInManifest(identity string) (*ManifestEntry, InstallScope, error) {
@@ -253,10 +180,13 @@ func ScanForExtensions(dir string) ([]ExtensionPreview, error) {
 			isExtDir := base == "extensions" || base == "ext" ||
 				strings.HasSuffix(base, "-extensions") || strings.HasSuffix(base, "-ext")
 
-			// Or check if it's a subdirectory of examples/ that might contain extensions
-			isExamplesSubdir := relPath == "examples" || strings.HasPrefix(relPath, "examples/")
+			// Allow walking into examples/ so we can reach examples/extensions/ etc,
+			// but don't treat examples/ itself or non-extension subdirs as extension locations.
+			if relPath == "examples" {
+				return nil
+			}
 
-			if !isExtDir && !isExamplesSubdir {
+			if !isExtDir {
 				// Check for main.go before skipping
 				mainPath := filepath.Join(path, "main.go")
 				if _, err := os.Stat(mainPath); err == nil {
@@ -272,18 +202,6 @@ func ScanForExtensions(dir string) ([]ExtensionPreview, error) {
 						}
 						return filepath.SkipDir
 					}
-					// Inside a valid extensions directory
-					if isExamplesSubdir || isExtDir {
-						if !multiFileDirs[relPath] {
-							multiFileDirs[relPath] = true
-							previews = append(previews, ExtensionPreview{
-								Path:   "./" + relPath + "/main.go",
-								Name:   deriveExtensionName(relPath+"/main.go", true),
-								IsMain: true,
-							})
-						}
-						return filepath.SkipDir
-					}
 				}
 
 				// Not an extension location
@@ -309,7 +227,7 @@ func ScanForExtensions(dir string) ([]ExtensionPreview, error) {
 		}
 
 		// It's a file - check if it's a valid extension
-		if !strings.HasSuffix(info.Name(), ".go") {
+		if !strings.HasSuffix(info.Name(), ".go") || strings.HasSuffix(info.Name(), "_test.go") {
 			return nil
 		}
 

internal/extensions/runner.go

@@ -1,21 +1,93 @@
 package extensions
 
 import (
+	"bytes"
 	"fmt"
+	"log"
 	"os"
+	"runtime"
 	"sort"
+	"strconv"
 	"strings"
 	"sync"
 
-	"github.com/charmbracelet/log"
 	"github.com/spf13/viper"
 )
 
+// ---------------------------------------------------------------------------
+// reentrantMu — a per-extension mutex that allows the same goroutine to
+// re-enter (e.g. handler → ctx.EmitCustomEvent → handler in same extension).
+// Different goroutines are serialized, preventing concurrent state mutation.
+// ---------------------------------------------------------------------------
+
+type reentrantMu struct {
+	mu    sync.Mutex
+	cond  *sync.Cond
+	owner int64 // goroutine ID that holds the lock, or 0
+	depth int   // re-entrancy depth
+}
+
+// initReentrantMu initializes the reentrant mutex in-place. Must be called
+// after the struct is at its final memory location (not before copying).
+func (r *reentrantMu) init() {
+	r.cond = sync.NewCond(&r.mu)
+}
+
+// lock acquires the mutex. If the calling goroutine already holds it, the
+// call succeeds immediately (re-entrant). Every call to lock must be paired
+// with a call to unlock.
+func (r *reentrantMu) lock() {
+	gid := goroutineID()
+	r.mu.Lock()
+	if r.owner == gid {
+		// Re-entrant: same goroutine already holds the lock.
+		r.depth++
+		r.mu.Unlock()
+		return
+	}
+	// Wait for the current owner to release.
+	for r.owner != 0 {
+		r.cond.Wait() // releases mu, blocks, re-acquires mu on wake
+	}
+	r.owner = gid
+	r.depth = 1
+	r.mu.Unlock()
+}
+
+// unlock releases the mutex (or decrements re-entrancy depth).
+func (r *reentrantMu) unlock() {
+	r.mu.Lock()
+	r.depth--
+	if r.depth == 0 {
+		r.owner = 0
+		r.cond.Signal()
+	}
+	r.mu.Unlock()
+}
+
+// goroutineID extracts the current goroutine's ID from runtime.Stack output.
+// This is a well-known technique used by Go testing infrastructure.
+func goroutineID() int64 {
+	var buf [64]byte
+	n := runtime.Stack(buf[:], false)
+	// Stack output starts with "goroutine NNN ["
+	s := buf[:n]
+	s = s[len("goroutine "):]
+	s = s[:bytes.IndexByte(s, ' ')]
+	id, _ := strconv.ParseInt(string(s), 10, 64)
+	return id
+}
+
 // Runner manages loaded extensions and dispatches events to their handlers
 // sequentially. Handlers execute in extension
 // load order; for cancellable events the first blocking result wins.
+//
+// Each extension has a dedicated reentrant mutex so that handlers for the
+// same extension are serialized (preventing data races on shared package-level
+// state), while handlers for different extensions may execute concurrently.
 type Runner struct {
 	extensions      []LoadedExtension
+	extMu           []reentrantMu // per-extension reentrant mutex, indexed by extension position
 	ctx             Context
 	widgets         map[string]WidgetConfig   // keyed by widget ID
 	statusEntries   map[string]StatusBarEntry // keyed by status key
@@ -26,9 +98,20 @@ type Runner struct {
 	disabledTools   map[string]bool           // nil = all tools enabled
 	customEventSubs map[string][]func(string) // inter-extension event bus
 	optionOverrides map[string]string         // runtime option overrides
+	configStore     *viper.Viper              // per-instance config store (nil = global)
 	mu              sync.RWMutex
 }
 
+// SetConfigStore sets the per-instance configuration store used by GetOption
+// to resolve "options.<name>" config values. When unset (nil), GetOption falls
+// back to the process-global viper store. Threading a per-Kit store keeps
+// extension option resolution isolated between Kit instances.
+func (r *Runner) SetConfigStore(v *viper.Viper) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	r.configStore = v
+}
+
 // ShortcutEntry pairs a shortcut definition with its handler.
 type ShortcutEntry struct {
 	Def     ShortcutDef
@@ -52,7 +135,11 @@ type LoadedExtension struct {
 
 // NewRunner creates a Runner from a set of loaded extensions.
 func NewRunner(exts []LoadedExtension) *Runner {
-	return &Runner{extensions: exts}
+	mus := make([]reentrantMu, len(exts))
+	for i := range mus {
+		mus[i].init()
+	}
+	return &Runner{extensions: exts, extMu: mus}
 }
 
 // SetContext updates the runtime context (session ID, model, etc.) that is
@@ -86,6 +173,21 @@ func normalizeContext(ctx Context) Context {
 	if ctx.CancelAndSend == nil {
 		ctx.CancelAndSend = func(string) {}
 	}
+	if ctx.Abort == nil {
+		ctx.Abort = func() {}
+	}
+	if ctx.IsIdle == nil {
+		ctx.IsIdle = func() bool { return true }
+	}
+	if ctx.Compact == nil {
+		ctx.Compact = func(CompactConfig) error { return fmt.Errorf("compact not available") }
+	}
+	if ctx.SendMultimodalMessage == nil {
+		ctx.SendMultimodalMessage = func(string, []FilePart) {}
+	}
+	if ctx.GetSessionUsage == nil {
+		ctx.GetSessionUsage = func() SessionUsage { return SessionUsage{} }
+	}
 	if ctx.SetWidget == nil {
 		ctx.SetWidget = func(WidgetConfig) {}
 	}
@@ -352,13 +454,15 @@ func (r *Runner) Emit(event Event) (Result, error) {
 	for i := range r.extensions {
 		ext := &r.extensions[i]
 		handlers := ext.Handlers[event.Type()]
+		if len(handlers) == 0 {
+			continue
+		}
+
+		r.extMu[i].lock()
 		for _, handler := range handlers {
 			result, err := safeCall(handler, event, ctx)
 			if err != nil {
-				log.Warn("extension handler error",
-					"path", ext.Path,
-					"event", event.Type(),
-					"err", err)
+				log.Printf("WARN extension handler error: path=%s event=%s err=%v", ext.Path, event.Type(), err)
 				continue
 			}
 			if result == nil {
@@ -367,6 +471,7 @@ func (r *Runner) Emit(event Event) (Result, error) {
 
 			// Check for blocking/short-circuit results.
 			if isBlocking(result) {
+				r.extMu[i].unlock()
 				return result, nil
 			}
 
@@ -374,6 +479,7 @@ func (r *Runner) Emit(event Event) (Result, error) {
 			// the caller is responsible for applying the modifications.
 			accumulated = result
 		}
+		r.extMu[i].unlock()
 	}
 	return accumulated, nil
 }
@@ -692,9 +798,7 @@ func (r *Runner) EmitCustomEvent(name, data string) {
 	safeInvoke := func(h func(string)) {
 		defer func() {
 			if rec := recover(); rec != nil {
-				log.Warn("custom event handler panicked",
-					"event", name,
-					"err", fmt.Sprintf("%v", rec))
+				log.Printf("WARN custom event handler panicked: event=%s err=%v", name, rec)
 			}
 		}()
 		h(data)
@@ -702,11 +806,17 @@ func (r *Runner) EmitCustomEvent(name, data string) {
 
 	// Extension-registered handlers first (in load order).
 	for i := range r.extensions {
-		for _, h := range r.extensions[i].CustomEventHandlers[name] {
+		extHandlers := r.extensions[i].CustomEventHandlers[name]
+		if len(extHandlers) == 0 {
+			continue
+		}
+		r.extMu[i].lock()
+		for _, h := range extHandlers {
 			safeInvoke(h)
 		}
+		r.extMu[i].unlock()
 	}
-	// Then dynamic subscriptions.
+	// Then dynamic subscriptions (not extension-scoped, no per-ext lock).
 	for _, h := range dynamicHandlers {
 		safeInvoke(h)
 	}
@@ -773,7 +883,13 @@ func (r *Runner) GetOption(name string) string {
 
 	// 3. Viper config: options.<name>
 	configKey := "options." + name
-	if v := viper.GetString(configKey); v != "" {
+	r.mu.RLock()
+	store := r.configStore
+	r.mu.RUnlock()
+	if store == nil {
+		store = viper.GetViper()
+	}
+	if v := store.GetString(configKey); v != "" {
 		return v
 	}
 

internal/extensions/runner_test.go

@@ -1,6 +1,7 @@
 package extensions
 
 import (
+	"sync"
 	"testing"
 )
 
@@ -571,3 +572,142 @@ func TestRunner_ContextPrintNilSafe(t *testing.T) {
 		t.Fatalf("unexpected error: %v", err)
 	}
 }
+
+func TestRunner_ConcurrentEmitSameExtension(t *testing.T) {
+	// Verify that concurrent Emit calls for the same extension are serialized
+	// and don't cause data races on shared handler state.
+	var counter int
+	ext := makeHandlerExt("shared-state.go", map[EventType][]HandlerFunc{
+		SubagentStart: {
+			func(e Event, c Context) Result {
+				// Read-modify-write: racy without serialization.
+				v := counter
+				counter = v + 1
+				return nil
+			},
+		},
+		SubagentChunk: {
+			func(e Event, c Context) Result {
+				v := counter
+				counter = v + 1
+				return nil
+			},
+		},
+	})
+
+	r := makeRunner(ext)
+	var wg sync.WaitGroup
+	const goroutines = 20
+	const iterations = 50
+	wg.Add(goroutines)
+	for range goroutines {
+		go func() {
+			defer wg.Done()
+			for range iterations {
+				_, _ = r.Emit(SubagentStartEvent{ToolCallID: "x"})
+				_, _ = r.Emit(SubagentChunkEvent{ToolCallID: "x"})
+			}
+		}()
+	}
+	wg.Wait()
+	if counter != goroutines*iterations*2 {
+		t.Errorf("expected counter=%d, got %d (race detected)", goroutines*iterations*2, counter)
+	}
+}
+
+func TestRunner_ConcurrentEmitDifferentExtensions(t *testing.T) {
+	// Two extensions with independent state should not block each other
+	// and should both run correctly under concurrent Emit calls.
+	var counter1, counter2 int
+	ext1 := makeHandlerExt("ext1.go", map[EventType][]HandlerFunc{
+		SubagentStart: {
+			func(e Event, c Context) Result {
+				v := counter1
+				counter1 = v + 1
+				return nil
+			},
+		},
+	})
+	ext2 := makeHandlerExt("ext2.go", map[EventType][]HandlerFunc{
+		SubagentStart: {
+			func(e Event, c Context) Result {
+				v := counter2
+				counter2 = v + 1
+				return nil
+			},
+		},
+	})
+
+	r := makeRunner(ext1, ext2)
+	var wg sync.WaitGroup
+	const goroutines = 20
+	const iterations = 50
+	wg.Add(goroutines)
+	for range goroutines {
+		go func() {
+			defer wg.Done()
+			for range iterations {
+				_, _ = r.Emit(SubagentStartEvent{ToolCallID: "x"})
+			}
+		}()
+	}
+	wg.Wait()
+	expected := goroutines * iterations
+	if counter1 != expected {
+		t.Errorf("ext1 counter: expected %d, got %d", expected, counter1)
+	}
+	if counter2 != expected {
+		t.Errorf("ext2 counter: expected %d, got %d", expected, counter2)
+	}
+}
+
+func TestRunner_ReentrantEmitCustomEvent(t *testing.T) {
+	// Verify that a handler can call EmitCustomEvent (which dispatches to
+	// the same extension's custom event handlers) without deadlocking.
+	var order []string
+	ext := LoadedExtension{
+		Path: "reentrant.go",
+		Handlers: map[EventType][]HandlerFunc{
+			SessionStart: {
+				func(e Event, c Context) Result {
+					order = append(order, "session_start")
+					// This triggers EmitCustomEvent for the same extension
+					// via a direct runner call (simulating ctx.EmitCustomEvent).
+					return nil
+				},
+			},
+		},
+		CustomEventHandlers: map[string][]func(string){
+			"test-event": {
+				func(data string) {
+					order = append(order, "custom:"+data)
+				},
+			},
+		},
+	}
+
+	r := makeRunner(ext)
+
+	// Wire up the handler to call EmitCustomEvent re-entrantly.
+	ext.Handlers[SessionStart] = []HandlerFunc{
+		func(e Event, c Context) Result {
+			order = append(order, "session_start")
+			r.EmitCustomEvent("test-event", "hello")
+			return nil
+		},
+	}
+	r.extensions[0] = ext
+	// Rebuild mutexes after modifying extensions slice.
+	r.extMu = make([]reentrantMu, len(r.extensions))
+	for i := range r.extMu {
+		r.extMu[i].init()
+	}
+
+	_, err := r.Emit(SessionStartEvent{})
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if len(order) != 2 || order[0] != "session_start" || order[1] != "custom:hello" {
+		t.Errorf("expected [session_start, custom:hello], got %v", order)
+	}
+}

internal/extensions/subagent.go

@@ -2,22 +2,15 @@
 package extensions
 
 import (
-	"bufio"
-	"context"
-	"encoding/json"
 	"fmt"
 	"os"
-	"os/exec"
-	"strings"
 	"sync"
-	"sync/atomic"
 	"time"
 )
 
 // ---------------------------------------------------------------------------
 // Subagent types
 // ---------------------------------------------------------------------------
-
 // SubagentConfig configures a subagent spawn.
 type SubagentConfig struct {
 	// Prompt is the task/instruction for the subagent (required).
@@ -157,221 +150,3 @@ func (h *SubagentHandle) Wait() SubagentResult {
 func (h *SubagentHandle) Done() <-chan struct{} {
 	return h.done
 }
-
-// ---------------------------------------------------------------------------
-// Internal helpers
-// ---------------------------------------------------------------------------
-
-// subagentJSONOutput matches the JSON envelope produced by `kit --json`.
-type subagentJSONOutput struct {
-	Response   string `json:"response"`
-	StopReason string `json:"stop_reason,omitempty"`
-	SessionID  string `json:"session_id,omitempty"`
-	Usage      *struct {
-		InputTokens  int64 `json:"input_tokens"`
-		OutputTokens int64 `json:"output_tokens"`
-	} `json:"usage,omitempty"`
-}
-
-var subagentCounter atomic.Uint64
-
-func generateSubagentID() string {
-	n := subagentCounter.Add(1)
-	return fmt.Sprintf("sub-%d-%d", time.Now().UnixNano(), n)
-}
-
-func findKitBinary() string {
-	// Try the current process executable first.
-	if exe, err := os.Executable(); err == nil {
-		if _, err := os.Stat(exe); err == nil {
-			return exe
-		}
-	}
-	// Fall back to PATH lookup.
-	if p, err := exec.LookPath("kit"); err == nil {
-		return p
-	}
-	return "kit"
-}
-
-// ---------------------------------------------------------------------------
-// SpawnSubagent implementation
-// ---------------------------------------------------------------------------
-
-// SpawnSubagent spawns a child Kit instance to perform a task.
-//
-// When config.Blocking is true, blocks until completion and returns the result
-// directly (handle is nil). When false, returns immediately with a handle for
-// monitoring/cancellation.
-//
-// The subagent runs with --json --no-session --no-extensions flags by default,
-// ensuring isolation from the parent's extensions and session state.
-func SpawnSubagent(cfg SubagentConfig) (*SubagentHandle, *SubagentResult, error) {
-	if cfg.Prompt == "" {
-		return nil, nil, fmt.Errorf("prompt is required")
-	}
-
-	timeout := cfg.Timeout
-	if timeout == 0 {
-		timeout = 5 * time.Minute
-	}
-
-	kitBinary := findKitBinary()
-
-	// Build subprocess arguments.
-	args := []string{
-		"--json",
-		"--no-extensions",
-	}
-	if cfg.NoSession {
-		args = append(args, "--no-session")
-	}
-	if cfg.Model != "" {
-		args = append(args, "--model", cfg.Model)
-	}
-
-	// Handle system prompt - write to temp file if provided.
-	var tmpFile *os.File
-	if cfg.SystemPrompt != "" {
-		var err error
-		tmpFile, err = os.CreateTemp("", "kit-subagent-*.txt")
-		if err != nil {
-			return nil, nil, fmt.Errorf("create temp file: %w", err)
-		}
-		if _, err := tmpFile.WriteString(cfg.SystemPrompt); err != nil {
-			_ = tmpFile.Close()
-			_ = os.Remove(tmpFile.Name())
-			return nil, nil, fmt.Errorf("write system prompt: %w", err)
-		}
-		_ = tmpFile.Close()
-		args = append(args, "--system-prompt", tmpFile.Name())
-	}
-
-	// Add the prompt as a positional argument.
-	args = append(args, cfg.Prompt)
-
-	// Create command with timeout context.
-	ctx, cancel := context.WithTimeout(context.Background(), timeout)
-
-	cmd := exec.CommandContext(ctx, kitBinary, args...)
-	cmd.Env = os.Environ()
-
-	stdout, err := cmd.StdoutPipe()
-	if err != nil {
-		cancel()
-		if tmpFile != nil {
-			_ = os.Remove(tmpFile.Name())
-		}
-		return nil, nil, fmt.Errorf("stdout pipe: %w", err)
-	}
-	stderr, err := cmd.StderrPipe()
-	if err != nil {
-		cancel()
-		if tmpFile != nil {
-			_ = os.Remove(tmpFile.Name())
-		}
-		return nil, nil, fmt.Errorf("stderr pipe: %w", err)
-	}
-
-	handle := &SubagentHandle{
-		ID:   generateSubagentID(),
-		done: make(chan struct{}),
-	}
-
-	// Start the subprocess.
-	start := time.Now()
-	if err := cmd.Start(); err != nil {
-		cancel()
-		if tmpFile != nil {
-			_ = os.Remove(tmpFile.Name())
-		}
-		return nil, nil, fmt.Errorf("start subprocess: %w", err)
-	}
-
-	handle.mu.Lock()
-	handle.proc = cmd.Process
-	handle.mu.Unlock()
-
-	// Run the subprocess monitoring in a goroutine.
-	go func() {
-		defer close(handle.done)
-		defer cancel()
-		if tmpFile != nil {
-			defer func() { _ = os.Remove(tmpFile.Name()) }()
-		}
-
-		var wg sync.WaitGroup
-		var stdoutBuf strings.Builder
-
-		// Read stderr (live output).
-		wg.Go(func() {
-			scanner := bufio.NewScanner(stderr)
-			scanner.Buffer(make([]byte, 256*1024), 256*1024)
-			for scanner.Scan() {
-				line := scanner.Text()
-				if cfg.OnOutput != nil && strings.TrimSpace(line) != "" {
-					cfg.OnOutput(line + "\n")
-				}
-			}
-		})
-
-		// Read stdout (JSON output).
-		scanner := bufio.NewScanner(stdout)
-		scanner.Buffer(make([]byte, 256*1024), 256*1024)
-		for scanner.Scan() {
-			stdoutBuf.WriteString(scanner.Text() + "\n")
-		}
-
-		wg.Wait()
-		waitErr := cmd.Wait()
-		elapsed := time.Since(start)
-
-		// Build result.
-		result := SubagentResult{Elapsed: elapsed}
-		if waitErr != nil {
-			result.Error = waitErr
-			if exitErr, ok := waitErr.(*exec.ExitError); ok {
-				result.ExitCode = exitErr.ExitCode()
-			} else {
-				result.ExitCode = 1
-			}
-		}
-
-		// Parse JSON output.
-		raw := strings.TrimSpace(stdoutBuf.String())
-		var parsed subagentJSONOutput
-		if raw != "" && json.Unmarshal([]byte(raw), &parsed) == nil {
-			result.Response = parsed.Response
-			result.SessionID = parsed.SessionID
-			if parsed.Usage != nil {
-				result.Usage = &SubagentUsage{
-					InputTokens:  parsed.Usage.InputTokens,
-					OutputTokens: parsed.Usage.OutputTokens,
-				}
-			}
-		} else {
-			// Fallback: use raw stdout.
-			result.Response = raw
-		}
-
-		handle.mu.Lock()
-		handle.result = &result
-		handle.proc = nil
-		handle.mu.Unlock()
-
-		if cfg.OnComplete != nil {
-			cfg.OnComplete(result)
-		}
-	}()
-
-	if cfg.Blocking {
-		// Wait for completion and return result directly.
-		<-handle.done
-		handle.mu.Lock()
-		r := handle.result
-		handle.mu.Unlock()
-		return nil, r, nil
-	}
-
-	return handle, nil, nil
-}

internal/extensions/symbols.go

@@ -31,6 +31,7 @@ func Symbols() interp.Exports {
 			// Session types
 			"SessionMessage": reflect.ValueOf((*SessionMessage)(nil)),
 			"ExtensionEntry": reflect.ValueOf((*ExtensionEntry)(nil)),
+			"SessionUsage":   reflect.ValueOf((*SessionUsage)(nil)),
 
 			// Option types
 			"OptionDef": reflect.ValueOf((*OptionDef)(nil)),
@@ -44,6 +45,8 @@ func Symbols() interp.Exports {
 			// LLM completion types
 			"CompleteRequest":  reflect.ValueOf((*CompleteRequest)(nil)),
 			"CompleteResponse": reflect.ValueOf((*CompleteResponse)(nil)),
+			"CompactConfig":    reflect.ValueOf((*CompactConfig)(nil)),
+			"FilePart":         reflect.ValueOf((*FilePart)(nil)),
 
 			// Status bar types
 			"StatusBarEntry": reflect.ValueOf((*StatusBarEntry)(nil)),
@@ -149,6 +152,9 @@ func Symbols() interp.Exports {
 			// Event structs
 			"ToolCallEvent":           reflect.ValueOf((*ToolCallEvent)(nil)),
 			"ToolCallResult":          reflect.ValueOf((*ToolCallResult)(nil)),
+			"ToolCallInputStartEvent": reflect.ValueOf((*ToolCallInputStartEvent)(nil)),
+			"ToolCallInputDeltaEvent": reflect.ValueOf((*ToolCallInputDeltaEvent)(nil)),
+			"ToolCallInputEndEvent":   reflect.ValueOf((*ToolCallInputEndEvent)(nil)),
 			"ToolExecutionStartEvent": reflect.ValueOf((*ToolExecutionStartEvent)(nil)),
 			"ToolExecutionEndEvent":   reflect.ValueOf((*ToolExecutionEndEvent)(nil)),
 			"ToolOutputEvent":         reflect.ValueOf((*ToolOutputEvent)(nil)),
@@ -166,6 +172,17 @@ func Symbols() interp.Exports {
 			"SessionStartEvent":       reflect.ValueOf((*SessionStartEvent)(nil)),
 			"SessionShutdownEvent":    reflect.ValueOf((*SessionShutdownEvent)(nil)),
 			"ModelChangeEvent":        reflect.ValueOf((*ModelChangeEvent)(nil)),
+
+			// Step lifecycle events
+			"StepStartEvent":      reflect.ValueOf((*StepStartEvent)(nil)),
+			"StepFinishEvent":     reflect.ValueOf((*StepFinishEvent)(nil)),
+			"ReasoningStartEvent": reflect.ValueOf((*ReasoningStartEvent)(nil)),
+			"WarningsEvent":       reflect.ValueOf((*WarningsEvent)(nil)),
+			"SourceEvent":         reflect.ValueOf((*SourceEvent)(nil)),
+			"ErrorEvent":          reflect.ValueOf((*ErrorEvent)(nil)),
+			"RetryEvent":          reflect.ValueOf((*RetryEvent)(nil)),
+			"PrepareStepEvent":    reflect.ValueOf((*PrepareStepEvent)(nil)),
+			"PrepareStepResult":   reflect.ValueOf((*PrepareStepResult)(nil)),
 		},
 	}
 }

internal/extensions/watcher.go

@@ -0,0 +1,192 @@
+package extensions
+
+import (
+	"context"
+	"fmt"
+	"log"
+	"os"
+	"path/filepath"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/fsnotify/fsnotify"
+)
+
+// Watcher monitors extension directories for file changes and triggers
+// a reload callback when .go files are created, modified, or removed.
+// It uses fsnotify for kernel-level file notifications (inotify on Linux,
+// kqueue on macOS) with debouncing to coalesce rapid editor writes.
+type Watcher struct {
+	watcher  *fsnotify.Watcher
+	onReload func()
+	debounce time.Duration
+	cancel   context.CancelFunc
+	done     chan struct{}
+	mu       sync.Mutex
+}
+
+// NewWatcher creates a file watcher that monitors the given directories
+// for .go file changes. When a change is detected (after debouncing),
+// onReload is called. The watcher must be started with Start() and
+// stopped with Close().
+func NewWatcher(dirs []string, onReload func()) (*Watcher, error) {
+	fsw, err := fsnotify.NewWatcher()
+	if err != nil {
+		return nil, fmt.Errorf("creating file watcher: %w", err)
+	}
+
+	for _, dir := range dirs {
+		// Watch the directory itself.
+		if err := fsw.Add(dir); err != nil {
+			log.Printf("DEBUG watcher: skipping directory: dir=%s err=%v", dir, err)
+			continue
+		}
+
+		// Also watch immediate subdirectories (for */main.go pattern).
+		entries, err := os.ReadDir(dir)
+		if err != nil {
+			continue
+		}
+		for _, entry := range entries {
+			if entry.IsDir() {
+				subdir := filepath.Join(dir, entry.Name())
+				if err := fsw.Add(subdir); err != nil {
+					log.Printf("DEBUG watcher: skipping subdirectory: dir=%s err=%v", subdir, err)
+				}
+			}
+		}
+	}
+
+	return &Watcher{
+		watcher:  fsw,
+		onReload: onReload,
+		debounce: 300 * time.Millisecond,
+		done:     make(chan struct{}),
+	}, nil
+}
+
+// Start begins watching for file changes. It blocks until the context
+// is cancelled or Close() is called. Typically called in a goroutine.
+func (w *Watcher) Start(ctx context.Context) {
+	w.mu.Lock()
+	ctx, w.cancel = context.WithCancel(ctx)
+	w.mu.Unlock()
+
+	defer close(w.done)
+
+	var timer *time.Timer
+	var timerC <-chan time.Time
+
+	for {
+		select {
+		case <-ctx.Done():
+			if timer != nil {
+				timer.Stop()
+			}
+			return
+
+		case event, ok := <-w.watcher.Events:
+			if !ok {
+				return
+			}
+
+			// Only care about .go files.
+			if !strings.HasSuffix(event.Name, ".go") {
+				continue
+			}
+
+			// React to write, create, remove, rename events.
+			if event.Op&(fsnotify.Write|fsnotify.Create|fsnotify.Remove|fsnotify.Rename) == 0 {
+				continue
+			}
+
+			log.Printf("DEBUG watcher: file changed: file=%s op=%s", event.Name, event.Op)
+
+			// Debounce: reset timer on each event.
+			if timer != nil {
+				timer.Stop()
+			}
+			timer = time.NewTimer(w.debounce)
+			timerC = timer.C
+
+		case <-timerC:
+			timerC = nil
+			timer = nil
+			log.Printf("DEBUG watcher: reloading extensions")
+			w.onReload()
+
+		case err, ok := <-w.watcher.Errors:
+			if !ok {
+				return
+			}
+			log.Printf("WARN watcher: error: %v", err)
+		}
+	}
+}
+
+// Close stops the watcher and releases resources.
+func (w *Watcher) Close() error {
+	w.mu.Lock()
+	cancel := w.cancel
+	w.mu.Unlock()
+
+	if cancel != nil {
+		cancel()
+	}
+
+	// Wait for the event loop to finish.
+	<-w.done
+	return w.watcher.Close()
+}
+
+// WatchedDirs returns the directories to watch for extension changes.
+// This includes the global extensions directory and the project-local
+// .kit/extensions/ directory (if they exist). Explicit -e paths that
+// point to directories are also included; explicit file paths cause
+// their parent directory to be watched instead.
+func WatchedDirs(extraPaths []string) []string {
+	var dirs []string
+	seen := make(map[string]bool)
+
+	add := func(dir string) {
+		abs, err := filepath.Abs(dir)
+		if err != nil {
+			return
+		}
+		if seen[abs] {
+			return
+		}
+
+		// Verify the directory exists.
+		info, err := os.Stat(abs)
+		if err != nil || !info.IsDir() {
+			return
+		}
+
+		seen[abs] = true
+		dirs = append(dirs, abs)
+	}
+
+	// Global extensions dir.
+	add(globalExtensionsDir())
+
+	// Project-local extensions dir.
+	add(filepath.Join(".kit", "extensions"))
+
+	// Explicit paths that are directories.
+	for _, p := range extraPaths {
+		info, err := os.Stat(p)
+		if err != nil {
+			continue
+		}
+		if info.IsDir() {
+			add(p)
+		} else {
+			// For explicit files, watch the parent directory.
+			add(filepath.Dir(p))
+		}
+	}
+
+	return dirs
+}

internal/extensions/watcher_test.go

@@ -0,0 +1,158 @@
+package extensions
+
+import (
+	"os"
+	"path/filepath"
+	"sync/atomic"
+	"testing"
+	"time"
+)
+
+func TestWatcher_ReloadsOnGoFileChange(t *testing.T) {
+	dir := t.TempDir()
+
+	// Write an initial extension file.
+	extFile := filepath.Join(dir, "test.go")
+	if err := os.WriteFile(extFile, []byte("package main\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	var reloadCount atomic.Int32
+
+	w, err := NewWatcher([]string{dir}, func() {
+		reloadCount.Add(1)
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	go w.Start(t.Context())
+
+	// Modify the file.
+	time.Sleep(50 * time.Millisecond) // let watcher settle
+	if err := os.WriteFile(extFile, []byte("package main\n// changed\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	// Wait for debounce (300ms) + margin.
+	time.Sleep(600 * time.Millisecond)
+
+	if got := reloadCount.Load(); got != 1 {
+		t.Errorf("expected 1 reload, got %d", got)
+	}
+
+	if err := w.Close(); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestWatcher_IgnoresNonGoFiles(t *testing.T) {
+	dir := t.TempDir()
+
+	var reloadCount atomic.Int32
+
+	w, err := NewWatcher([]string{dir}, func() {
+		reloadCount.Add(1)
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	go w.Start(t.Context())
+
+	// Write a non-.go file.
+	time.Sleep(50 * time.Millisecond)
+	txtFile := filepath.Join(dir, "notes.txt")
+	if err := os.WriteFile(txtFile, []byte("hello"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	// Wait past the debounce window.
+	time.Sleep(600 * time.Millisecond)
+
+	if got := reloadCount.Load(); got != 0 {
+		t.Errorf("expected 0 reloads for .txt file, got %d", got)
+	}
+
+	if err := w.Close(); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestWatcher_Debounces(t *testing.T) {
+	dir := t.TempDir()
+
+	extFile := filepath.Join(dir, "ext.go")
+	if err := os.WriteFile(extFile, []byte("package main\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	var reloadCount atomic.Int32
+
+	w, err := NewWatcher([]string{dir}, func() {
+		reloadCount.Add(1)
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	go w.Start(t.Context())
+
+	time.Sleep(50 * time.Millisecond)
+
+	// Rapid-fire writes (simulating editor save: write temp, rename, etc.).
+	for range 5 {
+		if err := os.WriteFile(extFile, []byte("package main\n// changed\n"), 0o644); err != nil {
+			t.Fatal(err)
+		}
+		time.Sleep(50 * time.Millisecond)
+	}
+
+	// Wait for debounce to fire.
+	time.Sleep(600 * time.Millisecond)
+
+	if got := reloadCount.Load(); got != 1 {
+		t.Errorf("expected 1 debounced reload, got %d", got)
+	}
+
+	if err := w.Close(); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestWatchedDirs_Deduplicates(t *testing.T) {
+	dir := t.TempDir()
+	dirs := WatchedDirs([]string{dir, dir})
+
+	count := 0
+	for _, d := range dirs {
+		abs, _ := filepath.Abs(dir)
+		if d == abs {
+			count++
+		}
+	}
+	if count != 1 {
+		t.Errorf("expected directory to appear once, got %d", count)
+	}
+}
+
+func TestWatchedDirs_FileParent(t *testing.T) {
+	dir := t.TempDir()
+	file := filepath.Join(dir, "ext.go")
+	if err := os.WriteFile(file, []byte("package main\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	dirs := WatchedDirs([]string{file})
+
+	abs, _ := filepath.Abs(dir)
+	found := false
+	for _, d := range dirs {
+		if d == abs {
+			found = true
+		}
+	}
+	if !found {
+		t.Errorf("expected parent dir %s in watched dirs %v", abs, dirs)
+	}
+}

internal/extensions/wrapper.go

@@ -28,11 +28,11 @@ func WrapToolsWithExtensions(tools []fantasy.AgentTool, runner *Runner) []fantas
 	return wrapped
 }
 
-// ExtensionToolsAsFantasy converts ToolDef values registered by extensions
-// into fantasy.AgentTool implementations so the LLM can invoke them.
+// ExtensionToolsAsLLMTools converts ToolDef values registered by extensions
+// into LLM agent tool implementations so the LLM can invoke them.
 // The runner is optional; if provided, ToolContext.OnProgress routes
 // progress messages through the runner's Print function.
-func ExtensionToolsAsFantasy(defs []ToolDef, runner *Runner) []fantasy.AgentTool {
+func ExtensionToolsAsLLMTools(defs []ToolDef, runner *Runner) []fantasy.AgentTool {
 	tools := make([]fantasy.AgentTool, 0, len(defs))
 	for _, def := range defs {
 		tools = append(tools, &extensionTool{def: def, runner: runner})
@@ -90,8 +90,7 @@ func (w *wrappedTool) Run(ctx context.Context, call fantasy.ToolCall) (fantasy.T
 	// 0. Check if tool is disabled via SetActiveTools.
 	if w.runner.IsToolDisabled(toolName) {
 		return fantasy.NewTextErrorResponse(
-				fmt.Sprintf("Error: tool %q is currently disabled", toolName)),
-			fmt.Errorf("tool %q disabled by extension", toolName)
+			fmt.Sprintf("Error: tool %q is currently disabled", toolName)), nil
 	}
 
 	kind := toolKindFor(toolName)
@@ -111,8 +110,7 @@ func (w *wrappedTool) Run(ctx context.Context, call fantasy.ToolCall) (fantasy.T
 			if reason == "" {
 				reason = "blocked by extension"
 			}
-			return fantasy.NewTextErrorResponse(fmt.Sprintf("Error: %s", reason)),
-				fmt.Errorf("tool blocked by extension: %s", reason)
+			return fantasy.NewTextErrorResponse(fmt.Sprintf("Error: %s", reason)), nil
 		}
 	}
 
@@ -154,7 +152,7 @@ func (w *wrappedTool) Run(ctx context.Context, call fantasy.ToolCall) (fantasy.T
 }
 
 // ---------------------------------------------------------------------------
-// extensionTool — wraps a ToolDef into a fantasy.AgentTool
+// extensionTool — wraps a ToolDef into an LLM agent tool
 // ---------------------------------------------------------------------------
 
 type extensionTool struct {
@@ -182,7 +180,7 @@ func (t *extensionTool) Info() fantasy.ToolInfo {
 				info.Parameters = props
 			} else {
 				// Schema doesn't have "properties" — use as-is (may be
-				// a flat property map already matching fantasy's format).
+				// a flat property map already matching the expected format).
 				info.Parameters = schema
 			}
 			// Extract required fields if present.
@@ -238,7 +236,7 @@ func (t *extensionTool) Run(ctx context.Context, call fantasy.ToolCall) (fantasy
 	}
 
 	if err != nil {
-		return fantasy.NewTextErrorResponse(err.Error()), err
+		return fantasy.NewTextErrorResponse(err.Error()), nil
 	}
 	return fantasy.NewTextResponse(result), nil
 }

internal/extensions/wrapper_test.go

@@ -142,8 +142,8 @@ func TestWrappedTool_BlockExecution(t *testing.T) {
 	if toolRan {
 		t.Error("tool should not have run after block")
 	}
-	if err == nil {
-		t.Error("expected error from blocked tool")
+	if err != nil {
+		t.Error("expected nil error for blocked tool (error is conveyed via IsError response)")
 	}
 	if resp.IsError != true {
 		t.Error("expected IsError=true from blocked response")
@@ -192,7 +192,7 @@ func TestWrappedTool_ExecutionStartEnd(t *testing.T) {
 	}
 }
 
-func TestExtensionToolsAsFantasy(t *testing.T) {
+func TestExtensionToolsAsLLMTools(t *testing.T) {
 	defs := []ToolDef{
 		{
 			Name:        "greet",
@@ -202,7 +202,7 @@ func TestExtensionToolsAsFantasy(t *testing.T) {
 		},
 	}
 
-	tools := ExtensionToolsAsFantasy(defs, nil)
+	tools := ExtensionToolsAsLLMTools(defs, nil)
 	if len(tools) != 1 {
 		t.Fatalf("expected 1 tool, got %d", len(tools))
 	}
@@ -232,10 +232,10 @@ func TestExtensionTool_Error(t *testing.T) {
 		},
 	}
 
-	tools := ExtensionToolsAsFantasy(defs, nil)
+	tools := ExtensionToolsAsLLMTools(defs, nil)
 	resp, err := tools[0].Run(context.Background(), fantasy.ToolCall{Input: "x"})
-	if err == nil {
-		t.Error("expected error")
+	if err != nil {
+		t.Error("expected nil error (error is conveyed via IsError response)")
 	}
 	if !resp.IsError {
 		t.Error("expected IsError=true")
@@ -259,7 +259,7 @@ func TestExtensionTool_ExecuteWithContext(t *testing.T) {
 	}
 
 	// Without runner, OnProgress is a no-op.
-	tools := ExtensionToolsAsFantasy(defs, nil)
+	tools := ExtensionToolsAsLLMTools(defs, nil)
 	resp, err := tools[0].Run(context.Background(), fantasy.ToolCall{Input: "test"})
 	if err != nil {
 		t.Fatalf("unexpected error: %v", err)
@@ -285,7 +285,7 @@ func TestExtensionTool_ExecuteWithContext(t *testing.T) {
 			},
 		},
 	}
-	tools2 := ExtensionToolsAsFantasy(defs2, runner)
+	tools2 := ExtensionToolsAsLLMTools(defs2, runner)
 	_, err = tools2[0].Run(context.Background(), fantasy.ToolCall{Input: ""})
 	if err != nil {
 		t.Fatalf("unexpected error: %v", err)
@@ -306,7 +306,7 @@ func TestExtensionTool_ExecuteWithContextPriority(t *testing.T) {
 			},
 		},
 	}
-	tools := ExtensionToolsAsFantasy(defs, nil)
+	tools := ExtensionToolsAsLLMTools(defs, nil)
 	resp, err := tools[0].Run(context.Background(), fantasy.ToolCall{Input: ""})
 	if err != nil {
 		t.Fatalf("unexpected error: %v", err)
@@ -330,7 +330,7 @@ func TestExtensionTool_CancelledContext(t *testing.T) {
 			},
 		},
 	}
-	tools := ExtensionToolsAsFantasy(defs, nil)
+	tools := ExtensionToolsAsLLMTools(defs, nil)
 	_, _ = tools[0].Run(ctx, fantasy.ToolCall{Input: ""})
 	if !sawCancelled {
 		t.Error("expected IsCancelled=true for cancelled context")
@@ -339,7 +339,7 @@ func TestExtensionTool_CancelledContext(t *testing.T) {
 
 func TestExtensionTool_ProviderOptions(t *testing.T) {
 	defs := []ToolDef{{Name: "test", Execute: func(string) (string, error) { return "", nil }}}
-	tools := ExtensionToolsAsFantasy(defs, nil)
+	tools := ExtensionToolsAsLLMTools(defs, nil)
 
 	// Initially nil.
 	opts := tools[0].ProviderOptions()

internal/fences/fences.go

@@ -0,0 +1,248 @@
+// Package fences provides utilities for detecting markdown code regions
+// (fenced code blocks and inline code spans) and applying transformations
+// only to text outside those regions.
+//
+// This prevents special tokens like $1, $@, or @file from being interpreted
+// when they appear inside ``` fences, ~~~ fences, or `inline` code spans.
+package fences
+
+import "strings"
+
+// Ranges returns byte ranges [start, end) of fenced code blocks in content.
+// Recognises both backtick (```) and tilde (~~~) fences, with optional
+// leading indentation (up to 3 spaces) and optional info strings.
+// An unclosed fence extends to the end of content.
+func Ranges(content string) [][2]int {
+	var result [][2]int
+	var inFence bool
+	var fenceChar byte
+	var fenceCount int
+	var fenceStart int
+
+	pos := 0
+	for pos < len(content) {
+		// Find the end of the current line.
+		lineEnd := strings.IndexByte(content[pos:], '\n')
+		var line string
+		var nextPos int
+		if lineEnd < 0 {
+			line = content[pos:]
+			nextPos = len(content)
+		} else {
+			line = content[pos : pos+lineEnd]
+			nextPos = pos + lineEnd + 1
+		}
+
+		trimmed := strings.TrimLeft(line, " ")
+		indent := len(line) - len(trimmed)
+
+		if !inFence {
+			if indent <= 3 {
+				if ch, n := parseFenceOpen(trimmed); n > 0 {
+					inFence = true
+					fenceChar = ch
+					fenceCount = n
+					fenceStart = pos
+				}
+			}
+		} else {
+			if indent <= 3 && isFenceClose(trimmed, fenceChar, fenceCount) {
+				result = append(result, [2]int{fenceStart, nextPos})
+				inFence = false
+			}
+		}
+
+		pos = nextPos
+	}
+
+	// Unclosed fence extends to end of content.
+	if inFence {
+		result = append(result, [2]int{fenceStart, len(content)})
+	}
+
+	return result
+}
+
+// ReplaceOutside applies fn to each text segment that is outside fenced code
+// blocks and inline code spans, leaving code content unchanged. This is the
+// primary entry point for callers that need to do regex replacement only on
+// non-code text.
+func ReplaceOutside(content string, fn func(string) string) string {
+	ranges := Ranges(content)
+	if len(ranges) == 0 {
+		return replaceOutsideInline(content, fn)
+	}
+
+	var b strings.Builder
+	b.Grow(len(content))
+	pos := 0
+	for _, r := range ranges {
+		if pos < r[0] {
+			// Within non-fenced segments, also skip inline code spans.
+			b.WriteString(replaceOutsideInline(content[pos:r[0]], fn))
+		}
+		// Preserve fenced content verbatim.
+		b.WriteString(content[r[0]:r[1]])
+		pos = r[1]
+	}
+	if pos < len(content) {
+		b.WriteString(replaceOutsideInline(content[pos:], fn))
+	}
+	return b.String()
+}
+
+// StripCode returns content with fenced code blocks and inline code spans
+// removed. Useful for detection/matching where only non-code text matters.
+func StripCode(content string) string {
+	// First strip fenced blocks.
+	stripped := StripFenced(content)
+	// Then strip inline code spans from what remains.
+	return stripInlineCode(stripped)
+}
+
+// StripFenced returns content with fenced code block regions removed.
+// Useful for detection/matching where only non-fenced text matters.
+// NOTE: this does NOT strip inline code spans; use StripCode for both.
+func StripFenced(content string) string {
+	ranges := Ranges(content)
+	if len(ranges) == 0 {
+		return content
+	}
+
+	var b strings.Builder
+	b.Grow(len(content))
+	pos := 0
+	for _, r := range ranges {
+		b.WriteString(content[pos:r[0]])
+		pos = r[1]
+	}
+	b.WriteString(content[pos:])
+	return b.String()
+}
+
+// parseFenceOpen checks whether trimmed (leading spaces already removed)
+// starts a fenced code block. Returns the fence character and count, or
+// (0, 0) if it is not a fence opener.
+func parseFenceOpen(trimmed string) (byte, int) {
+	if len(trimmed) == 0 {
+		return 0, 0
+	}
+	ch := trimmed[0]
+	if ch != '`' && ch != '~' {
+		return 0, 0
+	}
+	count := 0
+	for count < len(trimmed) && trimmed[count] == ch {
+		count++
+	}
+	if count < 3 {
+		return 0, 0
+	}
+	// Per CommonMark: backtick fences cannot have backticks in the info string.
+	if ch == '`' && strings.ContainsRune(trimmed[count:], '`') {
+		return 0, 0
+	}
+	return ch, count
+}
+
+// isFenceClose checks whether trimmed is a closing fence matching fenceChar
+// with at least minCount characters. A closing fence line contains only the
+// fence characters and optional trailing spaces.
+func isFenceClose(trimmed string, fenceChar byte, minCount int) bool {
+	if len(trimmed) == 0 || trimmed[0] != fenceChar {
+		return false
+	}
+	count := 0
+	for count < len(trimmed) && trimmed[count] == fenceChar {
+		count++
+	}
+	if count < minCount {
+		return false
+	}
+	// Closing fence must contain only fence chars (and optional trailing spaces).
+	return strings.TrimRight(trimmed[count:], " ") == ""
+}
+
+// --------------------------------------------------------------------------
+// Inline code span handling
+// --------------------------------------------------------------------------
+
+// inlineCodeRanges returns byte ranges [start, end) of inline code spans
+// in segment. Per CommonMark, a code span opens with N backticks and closes
+// with exactly N backticks.
+func inlineCodeRanges(s string) [][2]int {
+	var result [][2]int
+	i := 0
+	for i < len(s) {
+		if s[i] != '`' {
+			i++
+			continue
+		}
+		// Count opening backticks.
+		start := i
+		n := 0
+		for i < len(s) && s[i] == '`' {
+			n++
+			i++
+		}
+		// Scan for a closing run of exactly n backticks.
+		for j := i; j < len(s); {
+			if s[j] != '`' {
+				j++
+				continue
+			}
+			m := 0
+			for j < len(s) && s[j] == '`' {
+				m++
+				j++
+			}
+			if m == n {
+				result = append(result, [2]int{start, j})
+				i = j
+				break
+			}
+		}
+		// If no closing run was found, i is already past the opening
+		// backticks so the outer loop advances naturally.
+	}
+	return result
+}
+
+// replaceOutsideInline applies fn only to text outside inline code spans.
+func replaceOutsideInline(segment string, fn func(string) string) string {
+	ranges := inlineCodeRanges(segment)
+	if len(ranges) == 0 {
+		return fn(segment)
+	}
+	var b strings.Builder
+	b.Grow(len(segment))
+	pos := 0
+	for _, r := range ranges {
+		if pos < r[0] {
+			b.WriteString(fn(segment[pos:r[0]]))
+		}
+		b.WriteString(segment[r[0]:r[1]])
+		pos = r[1]
+	}
+	if pos < len(segment) {
+		b.WriteString(fn(segment[pos:]))
+	}
+	return b.String()
+}
+
+// stripInlineCode removes inline code spans from s.
+func stripInlineCode(s string) string {
+	ranges := inlineCodeRanges(s)
+	if len(ranges) == 0 {
+		return s
+	}
+	var b strings.Builder
+	b.Grow(len(s))
+	pos := 0
+	for _, r := range ranges {
+		b.WriteString(s[pos:r[0]])
+		pos = r[1]
+	}
+	b.WriteString(s[pos:])
+	return b.String()
+}

internal/fences/fences_test.go

@@ -0,0 +1,313 @@
+package fences
+
+import (
+	"testing"
+)
+
+func TestRanges(t *testing.T) {
+	tests := []struct {
+		name    string
+		content string
+		want    [][2]int
+	}{
+		{
+			name:    "no fences",
+			content: "hello world\nno code here",
+			want:    nil,
+		},
+		{
+			name:    "single backtick fence",
+			content: "before\n```\ncode\n```\nafter",
+			want:    [][2]int{{7, 20}},
+		},
+		{
+			name:    "single tilde fence",
+			content: "before\n~~~\ncode\n~~~\nafter",
+			want:    [][2]int{{7, 20}},
+		},
+		{
+			name:    "fence with info string",
+			content: "before\n```go\ncode\n```\nafter",
+			want:    [][2]int{{7, 22}},
+		},
+		{
+			name:    "multiple fences",
+			content: "a\n```\nx\n```\nb\n~~~\ny\n~~~\nc",
+			want:    [][2]int{{2, 12}, {14, 24}},
+		},
+		{
+			name:    "unclosed fence",
+			content: "before\n```\ncode\nmore code",
+			want:    [][2]int{{7, 25}},
+		},
+		{
+			name:    "longer closing fence",
+			content: "before\n```\ncode\n`````\nafter",
+			want:    [][2]int{{7, 22}},
+		},
+		{
+			name:    "shorter closing fence ignored",
+			content: "before\n`````\ncode\n```\nmore\n`````\nafter",
+			want:    [][2]int{{7, 33}},
+		},
+		{
+			name:    "indented fence up to 3 spaces",
+			content: "before\n   ```\ncode\n   ```\nafter",
+			want:    [][2]int{{7, 26}},
+		},
+		{
+			name:    "4 space indent is not a fence",
+			content: "before\n    ```\ncode\n    ```\nafter",
+			want:    nil,
+		},
+		{
+			name: "backtick in info string rejects open",
+			// The ```foo`bar line is not a valid opener (backtick in info).
+			// The standalone ``` becomes an opener with no close.
+			content: "before\n```foo`bar\ncode\n```\nafter",
+			want:    [][2]int{{23, 32}},
+		},
+		{
+			name:    "empty content",
+			content: "",
+			want:    nil,
+		},
+		{
+			name:    "fence only",
+			content: "```\ncode\n```",
+			want:    [][2]int{{0, 12}},
+		},
+		{
+			name:    "fence at end without trailing newline",
+			content: "```\ncode\n```",
+			want:    [][2]int{{0, 12}},
+		},
+		{
+			name:    "tilde fence does not close with backticks",
+			content: "~~~\ncode\n```\nmore\n~~~\nafter",
+			want:    [][2]int{{0, 22}},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := Ranges(tt.content)
+			if len(got) != len(tt.want) {
+				t.Fatalf("Ranges() = %v, want %v", got, tt.want)
+			}
+			for i := range got {
+				if got[i] != tt.want[i] {
+					t.Errorf("Ranges()[%d] = %v, want %v", i, got[i], tt.want[i])
+				}
+			}
+		})
+	}
+}
+
+func TestReplaceOutside(t *testing.T) {
+	upper := func(s string) string {
+		b := []byte(s)
+		for i, c := range b {
+			if c >= 'a' && c <= 'z' {
+				b[i] = c - 32
+			}
+		}
+		return string(b)
+	}
+
+	tests := []struct {
+		name    string
+		content string
+		want    string
+	}{
+		{
+			name:    "no fences",
+			content: "hello world",
+			want:    "HELLO WORLD",
+		},
+		{
+			name:    "text around fence",
+			content: "before\n```\ncode\n```\nafter",
+			want:    "BEFORE\n```\ncode\n```\nAFTER",
+		},
+		{
+			name:    "multiple fences",
+			content: "aaa\n```\nxxx\n```\nbbb\n~~~\nyyy\n~~~\nccc",
+			want:    "AAA\n```\nxxx\n```\nBBB\n~~~\nyyy\n~~~\nCCC",
+		},
+		{
+			name:    "unclosed fence preserves code",
+			content: "before\n```\ncode",
+			want:    "BEFORE\n```\ncode",
+		},
+		{
+			name:    "only fenced content",
+			content: "```\ncode\n```",
+			want:    "```\ncode\n```",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := ReplaceOutside(tt.content, upper)
+			if got != tt.want {
+				t.Errorf("ReplaceOutside() =\n%s\nwant:\n%s", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestStripFenced(t *testing.T) {
+	tests := []struct {
+		name    string
+		content string
+		want    string
+	}{
+		{
+			name:    "no fences",
+			content: "hello $1 world",
+			want:    "hello $1 world",
+		},
+		{
+			name:    "strips fenced code",
+			content: "before $1\n```\n$2 inside\n```\nafter $3",
+			want:    "before $1\nafter $3",
+		},
+		{
+			name:    "multiple fences",
+			content: "a\n```\nx\n```\nb\n~~~\ny\n~~~\nc",
+			want:    "a\nb\nc",
+		},
+		{
+			name:    "unclosed fence",
+			content: "before\n```\n$1 inside",
+			want:    "before\n",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := StripFenced(tt.content)
+			if got != tt.want {
+				t.Errorf("StripFenced() = %q, want %q", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestInlineCodeRanges(t *testing.T) {
+	tests := []struct {
+		name string
+		s    string
+		want [][2]int
+	}{
+		{"no backticks", "hello world", nil},
+		{"single backtick span", "use `$1` here", [][2]int{{4, 8}}},
+		{"double backtick span", "use ``$1`` here", [][2]int{{4, 10}}},
+		{"multiple spans", "`$1` and `$2`", [][2]int{{0, 4}, {9, 13}}},
+		{"unmatched backtick", "use `$1 here", nil},
+		{"mismatched backtick counts", "use ``$1` here", nil},
+		{"empty inline content", "use `` `` here", [][2]int{{4, 9}}},
+		{"backticks inside double", "use ``foo`bar`` here", [][2]int{{4, 15}}},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := inlineCodeRanges(tt.s)
+			if len(got) != len(tt.want) {
+				t.Fatalf("inlineCodeRanges() = %v, want %v", got, tt.want)
+			}
+			for i := range got {
+				if got[i] != tt.want[i] {
+					t.Errorf("inlineCodeRanges()[%d] = %v, want %v", i, got[i], tt.want[i])
+				}
+			}
+		})
+	}
+}
+
+func TestReplaceOutside_InlineCode(t *testing.T) {
+	upper := func(s string) string {
+		b := []byte(s)
+		for i, c := range b {
+			if c >= 'a' && c <= 'z' {
+				b[i] = c - 32
+			}
+		}
+		return string(b)
+	}
+
+	tests := []struct {
+		name    string
+		content string
+		want    string
+	}{
+		{
+			name:    "inline code preserved",
+			content: "use `code` here",
+			want:    "USE `code` HERE",
+		},
+		{
+			name:    "double backtick inline code",
+			content: "use ``co`de`` here",
+			want:    "USE ``co`de`` HERE",
+		},
+		{
+			name:    "mixed fenced and inline",
+			content: "before `x` mid\n```\nfenced\n```\nafter `y` end",
+			want:    "BEFORE `x` MID\n```\nfenced\n```\nAFTER `y` END",
+		},
+		{
+			name:    "only inline code",
+			content: "`code`",
+			want:    "`code`",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := ReplaceOutside(tt.content, upper)
+			if got != tt.want {
+				t.Errorf("ReplaceOutside() =\n%s\nwant:\n%s", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestStripCode(t *testing.T) {
+	tests := []struct {
+		name    string
+		content string
+		want    string
+	}{
+		{
+			name:    "no code",
+			content: "hello $1 world",
+			want:    "hello $1 world",
+		},
+		{
+			name:    "strips inline code",
+			content: "use `$1` and `$2` for positional args",
+			want:    "use  and  for positional args",
+		},
+		{
+			name:    "strips fenced and inline",
+			content: "before `$1`\n```\n$2 inside\n```\nafter",
+			want:    "before \nafter",
+		},
+		{
+			name:    "real world prompt template",
+			content: "Use $@ for all args.\n`$1`, `$2` for positional.\n```bash\necho $1\n```\n",
+			want:    "Use $@ for all args.\n,  for positional.\n",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := StripCode(tt.content)
+			if got != tt.want {
+				t.Errorf("StripCode() = %q, want %q", got, tt.want)
+			}
+		})
+	}
+}

internal/kitsetup/setup.go

@@ -33,6 +33,10 @@ type AgentSetupOptions struct {
 	// CoreTools overrides the default core tool set. If empty, core.AllTools()
 	// is used. Allows SDK users to pass custom tools (e.g. with WithWorkDir).
 	CoreTools []fantasy.AgentTool
+	// DisableCoreTools, when true, prevents loading any core tools.
+	// If both DisableCoreTools is true and CoreTools is empty, the agent
+	// will have no tools (useful for simple chat completions).
+	DisableCoreTools bool
 	// ExtraTools are additional tools added alongside core, MCP, and extension
 	// tools. They do not replace the defaults — they extend them.
 	ExtraTools []fantasy.AgentTool
@@ -40,6 +44,42 @@ type AgentSetupOptions struct {
 	// wrapping. Used by the SDK hook system. Both wrappers compose:
 	// extension wrapper runs first (inner), then this wrapper (outer).
 	ToolWrapper func([]fantasy.AgentTool) []fantasy.AgentTool
+
+	// ProviderConfig, when non-nil, is used directly instead of calling
+	// BuildProviderConfig(). Callers (e.g. Kit.New) pre-build this from their
+	// per-instance config store and pass it here, so the slow agent/MCP
+	// initialisation can run without further config reads.
+	ProviderConfig *models.ProviderConfig
+	// Debug enables debug logging. When zero-value, viper is consulted.
+	// Only meaningful when ProviderConfig is also set.
+	Debug bool
+	// NoExtensions skips extension loading. When false, viper is consulted.
+	// Only meaningful when ProviderConfig is also set.
+	NoExtensions bool
+	// MaxSteps overrides the agent step limit. 0 means use viper value.
+	// Only meaningful when ProviderConfig is also set.
+	MaxSteps int
+	// StreamingEnabled controls streaming. Only meaningful when ProviderConfig
+	// is also set.
+	StreamingEnabled bool
+	// AuthHandler handles OAuth authorization for remote MCP servers.
+	// When set, remote transports are configured with OAuth support.
+	AuthHandler tools.MCPAuthHandler
+	// TokenStoreFactory, if non-nil, creates a custom token store for each
+	// remote MCP server's OAuth tokens. When nil, the default file-based
+	// token store is used.
+	TokenStoreFactory tools.TokenStoreFactory
+	// OnMCPServerLoaded, if non-nil, is called when each MCP server finishes
+	// loading (successfully or with error). Called from the background goroutine.
+	OnMCPServerLoaded func(serverName string, toolCount int, err error)
+	// MCPTaskConfig configures task-augmented tools/call execution. The
+	// zero value preserves historical synchronous-only behaviour.
+	MCPTaskConfig tools.MCPTaskConfig
+	// Viper is the per-instance configuration store. When set, it is used for
+	// any fallback config reads (debug, no-extensions, max-steps, stream,
+	// extension paths) and is attached to the extension runner. When nil, the
+	// process-global viper store is used.
+	Viper *viper.Viper
 }
 
 // AgentSetupResult bundles the created agent and any debug logger so the caller
@@ -52,34 +92,62 @@ type AgentSetupResult struct {
 	ExtRunner *extensions.Runner
 }
 
-// BuildProviderConfig creates a *models.ProviderConfig from the current viper
-// state. All entry points (root, script, SDK) converge through this function.
-func BuildProviderConfig() (*models.ProviderConfig, string, error) {
-	systemPrompt, err := config.LoadSystemPrompt(viper.GetString("system-prompt"))
+// BuildProviderConfig creates a *models.ProviderConfig from the supplied viper
+// store (or the process-global store when v is nil). All entry points (root,
+// script, SDK) converge through this function.
+//
+// Generation parameter pointers (Temperature, TopP, etc.) are only set when
+// the user has explicitly configured them via CLI flag, environment variable,
+// or global config file. This allows per-model defaults from modelSettings
+// and customModels to fill in unset parameters downstream.
+func BuildProviderConfig(v *viper.Viper) (*models.ProviderConfig, string, error) {
+	if v == nil {
+		v = viper.GetViper()
+	}
+	systemPrompt, err := config.LoadSystemPrompt(v.GetString("system-prompt"))
 	if err != nil {
 		return nil, "", fmt.Errorf("failed to load system prompt: %w", err)
 	}
 
-	temperature := float32(viper.GetFloat64("temperature"))
-	topP := float32(viper.GetFloat64("top-p"))
-	topK := int32(viper.GetInt("top-k"))
-	numGPU := int32(viper.GetInt("num-gpu-layers"))
-	mainGPU := int32(viper.GetInt("main-gpu"))
+	numGPU := int32(v.GetInt("num-gpu-layers"))
+	mainGPU := int32(v.GetInt("main-gpu"))
 
 	cfg := &models.ProviderConfig{
-		ModelString:    viper.GetString("model"),
+		ModelString:    v.GetString("model"),
 		SystemPrompt:   systemPrompt,
-		ProviderAPIKey: viper.GetString("provider-api-key"),
-		ProviderURL:    viper.GetString("provider-url"),
-		MaxTokens:      viper.GetInt("max-tokens"),
-		Temperature:    &temperature,
-		TopP:           &topP,
-		TopK:           &topK,
-		StopSequences:  viper.GetStringSlice("stop-sequences"),
+		ProviderAPIKey: v.GetString("provider-api-key"),
+		ProviderURL:    v.GetString("provider-url"),
+		MaxTokens:      v.GetInt("max-tokens"),
+		StopSequences:  v.GetStringSlice("stop-sequences"),
 		NumGPU:         &numGPU,
 		MainGPU:        &mainGPU,
-		TLSSkipVerify:  viper.GetBool("tls-skip-verify"),
-		ThinkingLevel:  models.ParseThinkingLevel(viper.GetString("thinking-level")),
+		TLSSkipVerify:  v.GetBool("tls-skip-verify"),
+		ThinkingLevel:  models.ParseThinkingLevel(v.GetString("thinking-level")),
+		ConfigStore:    v,
+	}
+
+	// Only set generation parameter pointers when the user has explicitly
+	// provided a value. This leaves nil pointers for unset params, allowing
+	// per-model defaults (modelSettings / customModels params) to apply.
+	if v.IsSet("temperature") {
+		val := float32(v.GetFloat64("temperature"))
+		cfg.Temperature = &val
+	}
+	if v.IsSet("top-p") {
+		val := float32(v.GetFloat64("top-p"))
+		cfg.TopP = &val
+	}
+	if v.IsSet("top-k") {
+		val := int32(v.GetInt("top-k"))
+		cfg.TopK = &val
+	}
+	if v.IsSet("frequency-penalty") {
+		val := float32(v.GetFloat64("frequency-penalty"))
+		cfg.FrequencyPenalty = &val
+	}
+	if v.IsSet("presence-penalty") {
+		val := float32(v.GetFloat64("presence-penalty"))
+		cfg.PresencePenalty = &val
 	}
 
 	return cfg, systemPrompt, nil
@@ -88,15 +156,43 @@ func BuildProviderConfig() (*models.ProviderConfig, string, error) {
 // SetupAgent creates an agent from the current viper state + the provided
 // options. It wraps BuildProviderConfig and agent.CreateAgent.
 func SetupAgent(ctx context.Context, opts AgentSetupOptions) (*AgentSetupResult, error) {
-	modelConfig, systemPrompt, err := BuildProviderConfig()
-	if err != nil {
-		return nil, err
+	var modelConfig *models.ProviderConfig
+	var systemPrompt string
+
+	// Resolve the config store: prefer the per-instance store, falling back to
+	// the process-global store.
+	v := opts.Viper
+	if v == nil {
+		v = viper.GetViper()
 	}
 
+	if opts.ProviderConfig != nil {
+		// Pre-built config supplied by caller (e.g. Kit.New after building the
+		// per-instance store). Use it directly — no viper reads needed here.
+		modelConfig = opts.ProviderConfig
+		systemPrompt = modelConfig.SystemPrompt
+	} else {
+		var err error
+		modelConfig, systemPrompt, err = BuildProviderConfig(v)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	// Resolve debug / no-extensions / max-steps / streaming: prefer explicit
+	// fields (set when ProviderConfig was pre-built) over viper fallback.
+	debugEnabled := opts.Debug || v.GetBool("debug")
+	noExtensions := opts.NoExtensions || v.GetBool("no-extensions")
+	maxSteps := opts.MaxSteps
+	if maxSteps == 0 {
+		maxSteps = v.GetInt("max-steps")
+	}
+	streamingEnabled := opts.StreamingEnabled || v.GetBool("stream")
+
 	// Create the appropriate debug logger.
 	var debugLogger tools.DebugLogger
 	var bufferedLogger *tools.BufferedDebugLogger
-	if viper.GetBool("debug") {
+	if debugEnabled {
 		if opts.UseBufferedLogger {
 			bufferedLogger = tools.NewBufferedDebugLogger(true)
 			debugLogger = bufferedLogger
@@ -108,9 +204,9 @@ func SetupAgent(ctx context.Context, opts AgentSetupOptions) (*AgentSetupResult,
 	// Load extensions unless --no-extensions is set.
 	var extRunner *extensions.Runner
 	var extCreationOpts extensionCreationOpts
-	if !viper.GetBool("no-extensions") {
+	if !noExtensions {
 		var extErr error
-		extRunner, extCreationOpts, extErr = loadExtensions()
+		extRunner, extCreationOpts, extErr = loadExtensions(v)
 		if extErr != nil {
 			fmt.Printf("Warning: Failed to load extensions: %v\n", extErr)
 		}
@@ -137,18 +233,23 @@ func SetupAgent(ctx context.Context, opts AgentSetupOptions) (*AgentSetupResult,
 	}
 
 	a, err := agent.CreateAgent(ctx, &agent.AgentCreationOptions{
-		ModelConfig:      modelConfig,
-		MCPConfig:        opts.MCPConfig,
-		SystemPrompt:     systemPrompt,
-		MaxSteps:         viper.GetInt("max-steps"),
-		StreamingEnabled: viper.GetBool("stream"),
-		ShowSpinner:      opts.ShowSpinner,
-		Quiet:            opts.Quiet,
-		SpinnerFunc:      opts.SpinnerFunc,
-		DebugLogger:      debugLogger,
-		CoreTools:        opts.CoreTools,
-		ToolWrapper:      toolWrapper,
-		ExtraTools:       extraTools,
+		ModelConfig:       modelConfig,
+		MCPConfig:         opts.MCPConfig,
+		SystemPrompt:      systemPrompt,
+		MaxSteps:          maxSteps,
+		StreamingEnabled:  streamingEnabled,
+		ShowSpinner:       opts.ShowSpinner,
+		Quiet:             opts.Quiet,
+		SpinnerFunc:       opts.SpinnerFunc,
+		DebugLogger:       debugLogger,
+		AuthHandler:       opts.AuthHandler,
+		TokenStoreFactory: opts.TokenStoreFactory,
+		CoreTools:         opts.CoreTools,
+		DisableCoreTools:  opts.DisableCoreTools,
+		ToolWrapper:       toolWrapper,
+		ExtraTools:        extraTools,
+		OnMCPServerLoaded: opts.OnMCPServerLoaded,
+		MCPTaskConfig:     opts.MCPTaskConfig,
 	})
 	if err != nil {
 		return nil, fmt.Errorf("failed to create agent: %w", err)
@@ -169,9 +270,14 @@ type extensionCreationOpts struct {
 }
 
 // loadExtensions discovers and loads Yaegi extensions, builds the runner,
-// and returns the tool wrapper/extra tools.
-func loadExtensions() (*extensions.Runner, extensionCreationOpts, error) {
-	extraPaths := viper.GetStringSlice("extension")
+// and returns the tool wrapper/extra tools. The supplied store is used to
+// resolve the "extension" config key and is attached to the runner so
+// extension option lookups stay isolated to this Kit instance.
+func loadExtensions(v *viper.Viper) (*extensions.Runner, extensionCreationOpts, error) {
+	if v == nil {
+		v = viper.GetViper()
+	}
+	extraPaths := v.GetStringSlice("extension")
 	loaded, err := extensions.LoadExtensions(extraPaths)
 	if err != nil {
 		return nil, extensionCreationOpts{}, err
@@ -182,12 +288,13 @@ func loadExtensions() (*extensions.Runner, extensionCreationOpts, error) {
 	}
 
 	runner := extensions.NewRunner(loaded)
+	runner.SetConfigStore(v)
 
 	wrapper := func(tools []fantasy.AgentTool) []fantasy.AgentTool {
 		return extensions.WrapToolsWithExtensions(tools, runner)
 	}
 
-	extTools := extensions.ExtensionToolsAsFantasy(runner.RegisteredTools(), runner)
+	extTools := extensions.ExtensionToolsAsLLMTools(runner.RegisteredTools(), runner)
 
 	return runner, extensionCreationOpts{
 		toolWrapper: wrapper,

internal/message/content.go

@@ -325,12 +325,6 @@ func UnmarshalParts(data []byte) ([]ContentPart, error) {
 // mixed TextPart and ToolCallPart content. Tool-role messages produce
 // ToolResultPart entries.
 func (m *Message) ToLLMMessages() []fantasy.Message {
-	return m.ToFantasyMessages()
-}
-
-// Deprecated: Use ToLLMMessages instead.
-// ToFantasyMessages converts a Message to one or more LLM message values.
-func (m *Message) ToFantasyMessages() []fantasy.Message {
 	switch m.Role {
 	case RoleAssistant:
 		var parts []fantasy.MessagePart
@@ -431,13 +425,6 @@ func (m *Message) ToFantasyMessages() []fantasy.Message {
 // FromLLMMessage converts an LLM message into our Message type,
 // extracting all content parts into the appropriate block types.
 func FromLLMMessage(msg fantasy.Message) Message {
-	return FromFantasyMessage(msg)
-}
-
-// Deprecated: Use FromLLMMessage instead.
-// FromFantasyMessage converts an LLM message into our Message type,
-// extracting all content parts into the appropriate block types.
-func FromFantasyMessage(msg fantasy.Message) Message {
 	m := Message{
 		Role:      MessageRole(msg.Role),
 		Parts:     make([]ContentPart, 0),

internal/models/autoroute_test.go

@@ -0,0 +1,266 @@
+package models
+
+import (
+	"context"
+	"io"
+	"net/http"
+	"reflect"
+	"strings"
+	"testing"
+)
+
+// TestNpmToWireProtocol documents the wire protocols that the auto-router
+// understands. Provider-specific bundles (azure, bedrock, vercel, openrouter,
+// google-vertex*) are intentionally absent — they have native top-level cases
+// in CreateProvider and never reach the auto-router.
+func TestNpmToWireProtocol(t *testing.T) {
+	want := map[string]wireProtocol{
+		"@ai-sdk/openai":            wireOpenAI,
+		"@ai-sdk/openai-compatible": wireOpenAI,
+		"@ai-sdk/anthropic":         wireAnthropic,
+		"@ai-sdk/google":            wireGoogle,
+	}
+	for npm, wire := range want {
+		if got := npmToWireProtocol[npm]; got != wire {
+			t.Errorf("npmToWireProtocol[%q] = %d, want %d", npm, got, wire)
+		}
+	}
+
+	// Bundle packages must NOT be in the table (regression guard against the
+	// old npmToLLMProvider map that listed 10 entries but only handled 3).
+	for _, npm := range []string{
+		"@ai-sdk/google-vertex",
+		"@ai-sdk/google-vertex/anthropic",
+		"@ai-sdk/amazon-bedrock",
+		"@ai-sdk/azure",
+		"@openrouter/ai-sdk-provider",
+		"@ai-sdk/vercel",
+	} {
+		if _, ok := npmToWireProtocol[npm]; ok {
+			t.Errorf("npmToWireProtocol unexpectedly contains bundle package %q", npm)
+		}
+	}
+}
+
+// newTestRegistry builds a registry containing a single proxy-style provider
+// ("testproxy") with the given default npm, plus one model that carries the
+// given per-model npm override.
+func newTestRegistry(api, defaultNPM, modelID, modelNPMOverride string) *ModelsRegistry {
+	return &ModelsRegistry{
+		providers: map[string]ProviderInfo{
+			"testproxy": {
+				ID:   "testproxy",
+				Name: "Test Proxy",
+				Env:  []string{"TESTPROXY_API_KEY"},
+				NPM:  defaultNPM,
+				API:  api,
+				Models: map[string]ModelInfo{
+					modelID: {
+						ID:          modelID,
+						Name:        modelID,
+						ProviderNPM: modelNPMOverride,
+					},
+				},
+			},
+		},
+	}
+}
+
+// TestAutoRouteProvider_WireRouting verifies that autoRouteProvider routes each
+// npm package to the correct fantasy provider implementation. This is the core
+// regression test for issue #41: previously any npm that resolved to a
+// non-openai/anthropic/openaicompat LLM provider (notably @ai-sdk/google) hit a
+// dead `default` branch and failed with "has no LLM provider mapping".
+func TestAutoRouteProvider_WireRouting(t *testing.T) {
+	tests := []struct {
+		name        string
+		modelID     string
+		defaultNPM  string
+		overrideNPM string
+		// wantType is the concrete fantasy LanguageModel type the model should
+		// be routed to, identified by reflect type string.
+		wantType string
+	}{
+		{
+			name:       "openai-compatible default",
+			modelID:    "test-model",
+			defaultNPM: "@ai-sdk/openai-compatible",
+			wantType:   "openai.languageModel",
+		},
+		{
+			name:        "anthropic override",
+			modelID:     "test-model",
+			defaultNPM:  "@ai-sdk/openai-compatible",
+			overrideNPM: "@ai-sdk/anthropic",
+			wantType:    "anthropic.languageModel",
+		},
+		{
+			name:        "openai (responses) override",
+			modelID:     "gpt-4o",
+			defaultNPM:  "@ai-sdk/openai-compatible",
+			overrideNPM: "@ai-sdk/openai",
+			wantType:    "openai.responsesLanguageModel",
+		},
+		{
+			// The bug: opencode's gemini-* models override the default
+			// openai-compatible npm with @ai-sdk/google.
+			name:        "google override (issue #41)",
+			modelID:     "gemini-3.5-flash",
+			defaultNPM:  "@ai-sdk/openai-compatible",
+			overrideNPM: "@ai-sdk/google",
+			wantType:    "*google.languageModel",
+		},
+		{
+			// Unknown npm but provider has an API URL → openai-compatible fallback.
+			name:       "unknown npm with API URL falls back to openai-compat",
+			modelID:    "test-model",
+			defaultNPM: "@ai-sdk/some-future-thing",
+			wantType:   "openai.languageModel",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			reg := newTestRegistry("https://proxy.example/v1", tt.defaultNPM, tt.modelID, tt.overrideNPM)
+			config := &ProviderConfig{ProviderAPIKey: "test-key"}
+
+			result, err := autoRouteProvider(context.Background(), config, "testproxy", tt.modelID, reg)
+			if err != nil {
+				t.Fatalf("autoRouteProvider returned error: %v", err)
+			}
+			if result == nil || result.Model == nil {
+				t.Fatalf("autoRouteProvider returned nil model")
+			}
+
+			gotType := reflect.TypeOf(result.Model).String()
+			if gotType != tt.wantType {
+				t.Errorf("routed to %s, want %s", gotType, tt.wantType)
+			}
+		})
+	}
+}
+
+// TestAutoRouteProvider_UnknownNpmNoAPI verifies the improved error message for
+// a provider whose npm has no known wire protocol and that has no API URL to
+// fall back on.
+func TestAutoRouteProvider_UnknownNpmNoAPI(t *testing.T) {
+	reg := newTestRegistry("", "@ai-sdk/unmapped", "test-model", "")
+	config := &ProviderConfig{ProviderAPIKey: "test-key"}
+
+	_, err := autoRouteProvider(context.Background(), config, "testproxy", "test-model", reg)
+	if err == nil {
+		t.Fatal("expected error for unknown npm with no API URL, got nil")
+	}
+	if !strings.Contains(err.Error(), "cannot auto-route provider testproxy") {
+		t.Errorf("unexpected error message: %v", err)
+	}
+	if !strings.Contains(err.Error(), "--provider-url") {
+		t.Errorf("error should suggest --provider-url, got: %v", err)
+	}
+}
+
+// TestAutoRouteProvider_UnknownProvider verifies the not-in-database error.
+func TestAutoRouteProvider_UnknownProvider(t *testing.T) {
+	reg := newTestRegistry("https://proxy.example/v1", "@ai-sdk/openai-compatible", "test-model", "")
+	config := &ProviderConfig{ProviderAPIKey: "test-key"}
+
+	_, err := autoRouteProvider(context.Background(), config, "does-not-exist", "test-model", reg)
+	if err == nil {
+		t.Fatal("expected error for unknown provider, got nil")
+	}
+	if !strings.Contains(err.Error(), "not found in model database") {
+		t.Errorf("unexpected error message: %v", err)
+	}
+}
+
+// TestIsProviderLLMSupported_Google verifies that a provider whose npm is
+// @ai-sdk/google is reported as supported (it now maps to a wire protocol).
+func TestIsProviderLLMSupported_Google(t *testing.T) {
+	info := &ProviderInfo{ID: "testproxy", NPM: "@ai-sdk/google"}
+	if !isProviderLLMSupported("testproxy", info) {
+		t.Error("expected @ai-sdk/google provider to be LLM-supported")
+	}
+}
+
+// TestVersionedBasePath verifies detection of proxy base URLs that already
+// carry an API version segment (which collides with the genai SDK's injected
+// version).
+func TestVersionedBasePath(t *testing.T) {
+	tests := []struct {
+		rawURL string
+		want   string
+	}{
+		{"https://opencode.ai/zen/v1", "/zen/v1"},
+		{"https://opencode.ai/zen/v1/", "/zen/v1"},
+		{"https://example.com/api/v1beta", "/api/v1beta"},
+		{"https://example.com/api/v2alpha", "/api/v2alpha"},
+		{"https://generativelanguage.googleapis.com", ""},
+		{"https://proxy.example/openai", ""},
+		{"", ""},
+	}
+	for _, tt := range tests {
+		if got := versionedBasePath(tt.rawURL); got != tt.want {
+			t.Errorf("versionedBasePath(%q) = %q, want %q", tt.rawURL, got, tt.want)
+		}
+	}
+}
+
+// recordingRoundTripper captures the path of the request it receives.
+type recordingRoundTripper struct{ gotPath string }
+
+func (r *recordingRoundTripper) RoundTrip(req *http.Request) (*http.Response, error) {
+	r.gotPath = req.URL.Path
+	return &http.Response{
+		StatusCode: 200,
+		Body:       io.NopCloser(strings.NewReader("{}")),
+		Header:     make(http.Header),
+	}, nil
+}
+
+// TestGeminiProxyTransport_StripsInjectedVersion verifies that the transport
+// collapses the genai-injected "/v1beta" segment that follows a proxy base
+// URL which already carries its own version segment. This is the second-order
+// fix that makes opencode/gemini-* actually reach the proxy (issue #41).
+func TestGeminiProxyTransport_StripsInjectedVersion(t *testing.T) {
+	tests := []struct {
+		name     string
+		basePath string
+		reqPath  string
+		wantPath string
+	}{
+		{
+			name:     "strips doubled v1beta after /zen/v1",
+			basePath: "/zen/v1",
+			reqPath:  "/zen/v1/v1beta/models/gemini-3.5-flash:generateContent",
+			wantPath: "/zen/v1/models/gemini-3.5-flash:generateContent",
+		},
+		{
+			name:     "strips doubled v1beta1 after /zen/v1",
+			basePath: "/zen/v1",
+			reqPath:  "/zen/v1/v1beta1/models/gemini-3.5-flash:generateContent",
+			wantPath: "/zen/v1/models/gemini-3.5-flash:generateContent",
+		},
+		{
+			name:     "leaves non-matching path untouched",
+			basePath: "/zen/v1",
+			reqPath:  "/other/v1beta/models/x:generateContent",
+			wantPath: "/other/v1beta/models/x:generateContent",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			rec := &recordingRoundTripper{}
+			tr := &geminiProxyTransport{base: rec, basePath: tt.basePath}
+			req, err := http.NewRequest(http.MethodPost, "https://host"+tt.reqPath, nil)
+			if err != nil {
+				t.Fatalf("NewRequest: %v", err)
+			}
+			if _, err := tr.RoundTrip(req); err != nil {
+				t.Fatalf("RoundTrip: %v", err)
+			}
+			if rec.gotPath != tt.wantPath {
+				t.Errorf("forwarded path = %q, want %q", rec.gotPath, tt.wantPath)
+			}
+		})
+	}
+}

internal/models/cache_options.go

@@ -3,7 +3,6 @@ package models
 import (
 	"crypto/sha256"
 	"encoding/hex"
-	"maps"
 	"os"
 
 	"charm.land/fantasy"
@@ -69,19 +68,3 @@ func generateCacheKey(systemPrompt, modelID string) string {
 	// Prefix with "kit-" to identify KIT-generated cache keys
 	return "kit-" + hex.EncodeToString(h.Sum(nil))[:24]
 }
-
-// mergeProviderOptions merges multiple ProviderOptions maps.
-// Later maps take precedence over earlier ones.
-func mergeProviderOptions(opts ...fantasy.ProviderOptions) fantasy.ProviderOptions {
-	result := make(fantasy.ProviderOptions)
-
-	for _, opt := range opts {
-		maps.Copy(result, opt)
-	}
-
-	if len(result) == 0 {
-		return nil
-	}
-
-	return result
-}

internal/models/cache_options_test.go

@@ -3,8 +3,6 @@ package models
 import (
 	"os"
 	"testing"
-
-	"charm.land/fantasy"
 )
 
 func TestModelInfo_SupportsCaching(t *testing.T) {
@@ -192,57 +190,3 @@ func TestCachingPriorityOverThinking(t *testing.T) {
 		t.Errorf("OpenAI caching should work when thinking is OFF")
 	}
 }
-
-func TestMergeProviderOptions(t *testing.T) {
-	opts1 := fantasy.ProviderOptions{
-		"provider1": &testProviderData{value: "value1"},
-	}
-	opts2 := fantasy.ProviderOptions{
-		"provider2": &testProviderData{value: "value2"},
-	}
-
-	merged := mergeProviderOptions(opts1, opts2)
-
-	if len(merged) != 2 {
-		t.Errorf("mergeProviderOptions should combine options from multiple maps, got %d items", len(merged))
-	}
-
-	if _, ok := merged["provider1"]; !ok {
-		t.Errorf("merged options should contain 'provider1' key")
-	}
-
-	if _, ok := merged["provider2"]; !ok {
-		t.Errorf("merged options should contain 'provider2' key")
-	}
-
-	// Later options should override earlier ones
-	opts3 := fantasy.ProviderOptions{
-		"provider1": &testProviderData{value: "overridden"},
-	}
-	merged2 := mergeProviderOptions(opts1, opts3)
-
-	if data, ok := merged2["provider1"].(*testProviderData); ok {
-		if data.value != "overridden" {
-			t.Errorf("later options should override earlier ones, got %q", data.value)
-		}
-	}
-
-	if mergeProviderOptions() != nil {
-		t.Errorf("mergeProviderOptions with no args should return nil")
-	}
-}
-
-// testProviderData is a simple implementation of ProviderOptionsData for testing
-type testProviderData struct {
-	value string
-}
-
-func (t *testProviderData) Options() {}
-
-func (t *testProviderData) MarshalJSON() ([]byte, error) {
-	return []byte(`"` + t.value + `"`), nil
-}
-
-func (t *testProviderData) UnmarshalJSON(data []byte) error {
-	return nil
-}

internal/models/custom.go

@@ -2,20 +2,32 @@ package models
 
 import (
 	"log"
+	"os"
+	"strings"
 
 	"github.com/spf13/viper"
 )
 
 // loadCustomModelsFromConfig loads custom model definitions from the config file
 // and returns them as a map of model ID -> ModelInfo. Returns nil if no custom
-// models are configured.
+// models are configured. Reads from the process-global viper store (the model
+// registry is a process-global singleton).
 func loadCustomModelsFromConfig() map[string]ModelInfo {
-	if !viper.IsSet("customModels") {
+	return loadCustomModelsFrom(viper.GetViper())
+}
+
+// loadCustomModelsFrom loads custom model definitions from the supplied store.
+// When v is nil the process-global store is used.
+func loadCustomModelsFrom(v *viper.Viper) map[string]ModelInfo {
+	if v == nil {
+		v = viper.GetViper()
+	}
+	if !v.IsSet("customModels") {
 		return nil
 	}
 
 	var customModels map[string]CustomModelConfig
-	if err := viper.UnmarshalKey("customModels", &customModels); err != nil {
+	if err := v.UnmarshalKey("customModels", &customModels); err != nil {
 		log.Printf("Warning: Failed to parse customModels: %v", err)
 		return nil
 	}
@@ -31,12 +43,14 @@ func loadCustomModelsFromConfig() map[string]ModelInfo {
 
 // modelConfigToModelInfo converts a CustomModelConfig to a ModelInfo.
 func modelConfigToModelInfo(modelID string, cfg CustomModelConfig) ModelInfo {
-	return ModelInfo{
+	info := ModelInfo{
 		ID:          modelID,
 		Name:        cfg.Name,
 		Attachment:  cfg.Attachment,
 		Reasoning:   cfg.Reasoning,
 		Temperature: cfg.Temperature,
+		BaseURL:     cfg.BaseURL,
+		APIKey:      cfg.APIKey,
 		Cost: Cost{
 			Input:  cfg.Cost.Input,
 			Output: cfg.Cost.Output,
@@ -46,19 +60,265 @@ func modelConfigToModelInfo(modelID string, cfg CustomModelConfig) ModelInfo {
 			Output:  cfg.Limit.Output,
 		},
 	}
+
+	// Convert custom model generation params if any are set.
+	if p := convertGenerationParams(cfg.Params); p != nil {
+		info.Params = p
+	}
+
+	return info
+}
+
+// LoadModelSettingsFromConfig loads per-model generation parameter overrides
+// from the process-global viper store. Keys are "provider/model" strings.
+// Returns nil if no model settings are configured.
+func LoadModelSettingsFromConfig() map[string]*GenerationParams {
+	return LoadModelSettingsFrom(viper.GetViper())
+}
+
+// LoadModelSettingsFrom loads per-model generation parameter overrides from the
+// supplied per-instance store. When v is nil the process-global store is used.
+// Keys are "provider/model" strings. Returns nil if no model settings are
+// configured.
+func LoadModelSettingsFrom(v *viper.Viper) map[string]*GenerationParams {
+	if v == nil {
+		v = viper.GetViper()
+	}
+	if !v.IsSet("modelSettings") {
+		return nil
+	}
+
+	var settings map[string]GenerationParamsConfig
+	if err := v.UnmarshalKey("modelSettings", &settings); err != nil {
+		log.Printf("Warning: Failed to parse modelSettings: %v", err)
+		return nil
+	}
+
+	result := make(map[string]*GenerationParams, len(settings))
+	for modelKey, cfg := range settings {
+		if p := convertGenerationParams(cfg); p != nil {
+			result[modelKey] = p
+		}
+	}
+
+	return result
+}
+
+// convertGenerationParams converts a GenerationParamsConfig to a GenerationParams.
+// Returns nil if no parameters are set.
+func convertGenerationParams(cfg GenerationParamsConfig) *GenerationParams {
+	p := &GenerationParams{}
+	any := false
+
+	if cfg.MaxTokens != nil {
+		p.MaxTokens = cfg.MaxTokens
+		any = true
+	}
+	if cfg.Temperature != nil {
+		p.Temperature = cfg.Temperature
+		any = true
+	}
+	if cfg.TopP != nil {
+		p.TopP = cfg.TopP
+		any = true
+	}
+	if cfg.TopK != nil {
+		p.TopK = cfg.TopK
+		any = true
+	}
+	if cfg.FrequencyPenalty != nil {
+		p.FrequencyPenalty = cfg.FrequencyPenalty
+		any = true
+	}
+	if cfg.PresencePenalty != nil {
+		p.PresencePenalty = cfg.PresencePenalty
+		any = true
+	}
+	if len(cfg.StopSequences) > 0 {
+		p.StopSequences = cfg.StopSequences
+		any = true
+	}
+	if cfg.ThinkingLevel != "" {
+		p.ThinkingLevel = ParseThinkingLevel(cfg.ThinkingLevel)
+		any = true
+	}
+	if cfg.SystemPrompt != "" {
+		p.SystemPrompt = cfg.SystemPrompt
+		any = true
+	}
+
+	if !any {
+		return nil
+	}
+	return p
+}
+
+// ApplyModelSettings merges per-model generation parameter defaults from the
+// registry into a ProviderConfig. Model-level params are only applied for
+// fields where the user has not explicitly set a value (i.e., the
+// corresponding viper key is not set via CLI flag or global config).
+//
+// The lookup order is:
+//  1. modelSettings["provider/model"] from config (highest model-level priority)
+//  2. ModelInfo.Params from custom model definitions
+//
+// Both are overridden by explicit CLI flags / global config values.
+func ApplyModelSettings(config *ProviderConfig, modelInfo *ModelInfo) {
+	provider, modelName, err := ParseModelString(config.ModelString)
+	if err != nil {
+		return
+	}
+
+	// Resolve the config store: prefer the per-instance store carried on the
+	// ProviderConfig (set by BuildProviderConfig / Kit.New), falling back to
+	// the process-global store for callers that don't thread one through.
+	store := config.ConfigStore
+
+	// Collect model-level params: modelSettings override > custom model params.
+	// modelSettings takes priority because it's the more specific/intentional config.
+	var params *GenerationParams
+
+	// First check modelSettings from config.
+	if settings := LoadModelSettingsFrom(store); settings != nil {
+		modelKey := provider + "/" + modelName
+		if p, ok := settings[modelKey]; ok {
+			params = p
+		}
+	}
+
+	// Fall back to ModelInfo.Params (from custom model definitions).
+	if params == nil && modelInfo != nil && modelInfo.Params != nil {
+		params = modelInfo.Params
+	}
+
+	if params == nil {
+		return
+	}
+
+	// Apply each parameter only when the user hasn't explicitly set it.
+	// We check viper.IsSet() which returns true only when the key was
+	// set via CLI flag, environment variable, or config file global section.
+
+	if params.MaxTokens != nil && !isExplicitlySet(store, "max-tokens") {
+		config.MaxTokens = *params.MaxTokens
+	}
+	if params.Temperature != nil && !isExplicitlySet(store, "temperature") {
+		config.Temperature = params.Temperature
+	}
+	if params.TopP != nil && !isExplicitlySet(store, "top-p") {
+		config.TopP = params.TopP
+	}
+	if params.TopK != nil && !isExplicitlySet(store, "top-k") {
+		config.TopK = params.TopK
+	}
+	if params.FrequencyPenalty != nil && !isExplicitlySet(store, "frequency-penalty") {
+		config.FrequencyPenalty = params.FrequencyPenalty
+	}
+	if params.PresencePenalty != nil && !isExplicitlySet(store, "presence-penalty") {
+		config.PresencePenalty = params.PresencePenalty
+	}
+	if len(params.StopSequences) > 0 && !isExplicitlySet(store, "stop-sequences") {
+		config.StopSequences = params.StopSequences
+	}
+	if params.ThinkingLevel != "" && !isExplicitlySet(store, "thinking-level") {
+		config.ThinkingLevel = params.ThinkingLevel
+	}
+	if params.SystemPrompt != "" && config.SystemPrompt == "" {
+		// Resolve file paths: if the value points to an existing file, read it.
+		// We check config.SystemPrompt == "" rather than isExplicitlySet because
+		// viper.BindPFlag causes IsSet to return true even for unset flags.
+		config.SystemPrompt = LoadSystemPromptValue(params.SystemPrompt)
+	}
+}
+
+// LoadSystemPromptValue resolves a system prompt value that may be either
+// inline text or a file path. If the value is a path to an existing file,
+// its contents are read and returned. Otherwise the string is returned as-is.
+// This mirrors config.LoadSystemPrompt but lives in the models package to
+// avoid circular dependencies.
+func LoadSystemPromptValue(input string) string {
+	if input == "" {
+		return ""
+	}
+	if info, err := os.Stat(input); err == nil && !info.IsDir() {
+		content, err := os.ReadFile(input)
+		if err != nil {
+			log.Printf("Warning: failed to read system prompt file %q: %v", input, err)
+			return input
+		}
+		return strings.TrimSpace(string(content))
+	}
+	return input
+}
+
+// isExplicitlySet returns true when the user has explicitly set a config key
+// via CLI flag, environment variable, or the global section of the config file.
+// Model-level defaults should not override explicitly set values.
+//
+// The check runs against the supplied per-instance store when non-nil,
+// otherwise the process-global store. This keeps the "explicit vs unset"
+// precedence contract per-Kit-instance once a store is threaded through.
+func isExplicitlySet(v *viper.Viper, key string) bool {
+	if v == nil {
+		v = viper.GetViper()
+	}
+	// viper.IsSet returns true if the key has been set in any of the
+	// data stores (flag, env, config file, default). We need to check
+	// whether the value was set at the global config level (not just
+	// as a default). For generation params, the global config keys use
+	// hyphenated names (e.g. "max-tokens", "top-p").
+	//
+	// Since viper merges all sources, IsSet returns true even for config
+	// file values. This means global config file values (e.g.
+	// temperature: 0.7 at the top level) will correctly take precedence
+	// over model-level defaults, which is the desired behavior.
+	return v.IsSet(key)
+}
+
+// GenerationParams holds per-model generation parameter defaults.
+// These are stored on ModelInfo and applied during provider creation.
+// Nil pointer fields mean "no model-level default" — the global config
+// or CLI flag value (if any) will be used instead.
+type GenerationParams struct {
+	MaxTokens        *int
+	Temperature      *float32
+	TopP             *float32
+	TopK             *int32
+	FrequencyPenalty *float32
+	PresencePenalty  *float32
+	StopSequences    []string
+	ThinkingLevel    ThinkingLevel
+	SystemPrompt     string // Per-model system prompt (inline text or file path)
 }
 
 // CustomModelConfig defines a custom model configuration loaded from the config file.
 // This is a duplicate here to avoid circular dependencies with internal/config.
 type CustomModelConfig struct {
-	Name        string      `json:"name" yaml:"name"`
-	Family      string      `json:"family,omitempty" yaml:"family,omitempty"`
-	Attachment  bool        `json:"attachment,omitempty" yaml:"attachment,omitempty"`
-	Reasoning   bool        `json:"reasoning,omitempty" yaml:"reasoning,omitempty"`
-	Temperature bool        `json:"temperature,omitempty" yaml:"temperature,omitempty"`
-	Knowledge   string      `json:"knowledge,omitempty" yaml:"knowledge,omitempty"`
-	Cost        CostConfig  `json:"cost" yaml:"cost"`
-	Limit       LimitConfig `json:"limit" yaml:"limit"`
+	Name        string                 `json:"name" yaml:"name"`
+	BaseURL     string                 `json:"baseUrl,omitempty" yaml:"baseUrl,omitempty"`
+	APIKey      string                 `json:"apiKey,omitempty" yaml:"apiKey,omitempty"`
+	Family      string                 `json:"family,omitempty" yaml:"family,omitempty"`
+	Attachment  bool                   `json:"attachment,omitempty" yaml:"attachment,omitempty"`
+	Reasoning   bool                   `json:"reasoning,omitempty" yaml:"reasoning,omitempty"`
+	Temperature bool                   `json:"temperature,omitempty" yaml:"temperature,omitempty"`
+	Knowledge   string                 `json:"knowledge,omitempty" yaml:"knowledge,omitempty"`
+	Cost        CostConfig             `json:"cost" yaml:"cost"`
+	Limit       LimitConfig            `json:"limit" yaml:"limit"`
+	Params      GenerationParamsConfig `json:"params,omitzero" yaml:"params,omitempty"`
+}
+
+// GenerationParamsConfig is the JSON/YAML-serializable form of generation
+// parameter defaults. Used in both customModels[].params and modelSettings[].
+type GenerationParamsConfig struct {
+	MaxTokens        *int     `json:"maxTokens,omitempty" yaml:"maxTokens,omitempty"`
+	Temperature      *float32 `json:"temperature,omitempty" yaml:"temperature,omitempty"`
+	TopP             *float32 `json:"topP,omitempty" yaml:"topP,omitempty"`
+	TopK             *int32   `json:"topK,omitempty" yaml:"topK,omitempty"`
+	FrequencyPenalty *float32 `json:"frequencyPenalty,omitempty" yaml:"frequencyPenalty,omitempty"`
+	PresencePenalty  *float32 `json:"presencePenalty,omitempty" yaml:"presencePenalty,omitempty"`
+	StopSequences    []string `json:"stopSequences,omitempty" yaml:"stopSequences,omitempty"`
+	ThinkingLevel    string   `json:"thinkingLevel,omitempty" yaml:"thinkingLevel,omitempty"`
+	SystemPrompt     string   `json:"systemPrompt,omitempty" yaml:"systemPrompt,omitempty"`
 }
 
 // CostConfig defines the pricing for a custom model.

internal/models/custom_test.go

@@ -0,0 +1,422 @@
+package models
+
+import (
+	"os"
+	"testing"
+
+	"github.com/spf13/viper"
+)
+
+func TestConvertGenerationParams(t *testing.T) {
+	t.Run("empty config returns nil", func(t *testing.T) {
+		cfg := GenerationParamsConfig{}
+		p := convertGenerationParams(cfg)
+		if p != nil {
+			t.Errorf("expected nil, got %+v", p)
+		}
+	})
+
+	t.Run("temperature only", func(t *testing.T) {
+		temp := float32(0.7)
+		cfg := GenerationParamsConfig{Temperature: &temp}
+		p := convertGenerationParams(cfg)
+		if p == nil {
+			t.Fatal("expected non-nil")
+		}
+		if p.Temperature == nil || *p.Temperature != 0.7 {
+			t.Errorf("expected temperature 0.7, got %v", p.Temperature)
+		}
+		if p.TopP != nil {
+			t.Errorf("expected nil TopP, got %v", p.TopP)
+		}
+	})
+
+	t.Run("all params set", func(t *testing.T) {
+		maxTokens := 8192
+		temp := float32(0.5)
+		topP := float32(0.9)
+		topK := int32(50)
+		freqPenalty := float32(0.1)
+		presPenalty := float32(0.2)
+		cfg := GenerationParamsConfig{
+			MaxTokens:        &maxTokens,
+			Temperature:      &temp,
+			TopP:             &topP,
+			TopK:             &topK,
+			FrequencyPenalty: &freqPenalty,
+			PresencePenalty:  &presPenalty,
+			StopSequences:    []string{"STOP"},
+			ThinkingLevel:    "high",
+		}
+		p := convertGenerationParams(cfg)
+		if p == nil {
+			t.Fatal("expected non-nil")
+		}
+		if p.MaxTokens == nil || *p.MaxTokens != 8192 {
+			t.Errorf("expected maxTokens 8192, got %v", p.MaxTokens)
+		}
+		if p.Temperature == nil || *p.Temperature != 0.5 {
+			t.Errorf("expected temperature 0.5, got %v", p.Temperature)
+		}
+		if p.TopP == nil || *p.TopP != 0.9 {
+			t.Errorf("expected topP 0.9, got %v", p.TopP)
+		}
+		if p.TopK == nil || *p.TopK != 50 {
+			t.Errorf("expected topK 50, got %v", p.TopK)
+		}
+		if p.FrequencyPenalty == nil || *p.FrequencyPenalty != 0.1 {
+			t.Errorf("expected frequencyPenalty 0.1, got %v", p.FrequencyPenalty)
+		}
+		if p.PresencePenalty == nil || *p.PresencePenalty != 0.2 {
+			t.Errorf("expected presencePenalty 0.2, got %v", p.PresencePenalty)
+		}
+		if len(p.StopSequences) != 1 || p.StopSequences[0] != "STOP" {
+			t.Errorf("expected stop sequences [STOP], got %v", p.StopSequences)
+		}
+		if p.ThinkingLevel != ThinkingHigh {
+			t.Errorf("expected thinking level high, got %v", p.ThinkingLevel)
+		}
+	})
+
+	t.Run("thinking level parsing", func(t *testing.T) {
+		cfg := GenerationParamsConfig{ThinkingLevel: "medium"}
+		p := convertGenerationParams(cfg)
+		if p == nil {
+			t.Fatal("expected non-nil")
+		}
+		if p.ThinkingLevel != ThinkingMedium {
+			t.Errorf("expected thinking level medium, got %v", p.ThinkingLevel)
+		}
+	})
+	t.Run("system prompt only", func(t *testing.T) {
+		cfg := GenerationParamsConfig{SystemPrompt: "You are helpful."}
+		p := convertGenerationParams(cfg)
+		if p == nil {
+			t.Fatal("expected non-nil")
+		}
+		if p.SystemPrompt != "You are helpful." {
+			t.Errorf("expected system prompt, got %q", p.SystemPrompt)
+		}
+	})
+}
+
+func TestModelConfigToModelInfoWithParams(t *testing.T) {
+	temp := float32(0.8)
+	topP := float32(0.95)
+	cfg := CustomModelConfig{
+		Name:        "Test Model",
+		BaseURL:     "http://localhost:8080/v1",
+		Temperature: true,
+		Params: GenerationParamsConfig{
+			Temperature: &temp,
+			TopP:        &topP,
+		},
+	}
+
+	info := modelConfigToModelInfo("test-model", cfg)
+
+	if info.Params == nil {
+		t.Fatal("expected non-nil Params")
+	}
+	if info.Params.Temperature == nil || *info.Params.Temperature != 0.8 {
+		t.Errorf("expected temperature 0.8, got %v", info.Params.Temperature)
+	}
+	if info.Params.TopP == nil || *info.Params.TopP != 0.95 {
+		t.Errorf("expected topP 0.95, got %v", info.Params.TopP)
+	}
+}
+
+func TestModelConfigToModelInfoWithoutParams(t *testing.T) {
+	cfg := CustomModelConfig{
+		Name:    "Test Model",
+		BaseURL: "http://localhost:8080/v1",
+	}
+
+	info := modelConfigToModelInfo("test-model", cfg)
+
+	if info.Params != nil {
+		t.Errorf("expected nil Params, got %+v", info.Params)
+	}
+}
+
+func TestApplyModelSettings(t *testing.T) {
+	// Save and restore viper state.
+	originalViper := viper.AllSettings()
+	defer func() {
+		viper.Reset()
+		for k, v := range originalViper {
+			viper.Set(k, v)
+		}
+	}()
+
+	t.Run("applies model params when not explicitly set", func(t *testing.T) {
+		viper.Reset()
+
+		temp := float32(0.8)
+		topK := int32(50)
+		maxTokens := 4096
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				Temperature: &temp,
+				TopK:        &topK,
+				MaxTokens:   &maxTokens,
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		if config.Temperature == nil || *config.Temperature != 0.8 {
+			t.Errorf("expected temperature 0.8, got %v", config.Temperature)
+		}
+		if config.TopK == nil || *config.TopK != 50 {
+			t.Errorf("expected topK 50, got %v", config.TopK)
+		}
+		if config.MaxTokens != 4096 {
+			t.Errorf("expected maxTokens 4096, got %d", config.MaxTokens)
+		}
+	})
+
+	t.Run("explicit viper values take precedence", func(t *testing.T) {
+		viper.Reset()
+		viper.Set("temperature", 0.3)
+
+		temp := float32(0.8)
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				Temperature: &temp,
+			},
+		}
+
+		explicitTemp := float32(0.3)
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+			Temperature: &explicitTemp,
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		// Temperature should NOT be overridden because it's explicitly set in viper
+		if config.Temperature == nil || *config.Temperature != 0.3 {
+			t.Errorf("expected temperature 0.3 (explicit), got %v", config.Temperature)
+		}
+	})
+
+	t.Run("nil model info is safe", func(t *testing.T) {
+		viper.Reset()
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		// Should not panic
+		ApplyModelSettings(config, nil)
+
+		if config.Temperature != nil {
+			t.Errorf("expected nil temperature, got %v", config.Temperature)
+		}
+	})
+
+	t.Run("model info without params is safe", func(t *testing.T) {
+		viper.Reset()
+
+		modelInfo := &ModelInfo{ID: "test-model"}
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		if config.Temperature != nil {
+			t.Errorf("expected nil temperature, got %v", config.Temperature)
+		}
+	})
+
+	t.Run("modelSettings from viper takes priority over ModelInfo.Params", func(t *testing.T) {
+		viper.Reset()
+
+		// Set up modelSettings in viper (simulating config file)
+		viper.Set("modelSettings", map[string]any{
+			"custom/test-model": map[string]any{
+				"temperature": 0.5,
+				"topK":        30,
+			},
+		})
+
+		// ModelInfo has different params
+		temp := float32(0.8)
+		topK := int32(50)
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				Temperature: &temp,
+				TopK:        &topK,
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		// modelSettings should win over ModelInfo.Params
+		if config.Temperature == nil || *config.Temperature != 0.5 {
+			t.Errorf("expected temperature 0.5 (from modelSettings), got %v", config.Temperature)
+		}
+		if config.TopK == nil || *config.TopK != 30 {
+			t.Errorf("expected topK 30 (from modelSettings), got %v", config.TopK)
+		}
+	})
+
+	t.Run("stop sequences applied from model params", func(t *testing.T) {
+		viper.Reset()
+
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				StopSequences: []string{"STOP", "END"},
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		if len(config.StopSequences) != 2 || config.StopSequences[0] != "STOP" {
+			t.Errorf("expected stop sequences [STOP END], got %v", config.StopSequences)
+		}
+	})
+
+	t.Run("thinking level applied from model params", func(t *testing.T) {
+		viper.Reset()
+
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				ThinkingLevel: ThinkingHigh,
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		if config.ThinkingLevel != ThinkingHigh {
+			t.Errorf("expected thinking level high, got %v", config.ThinkingLevel)
+		}
+	})
+
+	t.Run("system prompt applied from model params", func(t *testing.T) {
+		viper.Reset()
+
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				SystemPrompt: "You are a coding assistant.",
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		if config.SystemPrompt != "You are a coding assistant." {
+			t.Errorf("expected system prompt to be set, got %q", config.SystemPrompt)
+		}
+	})
+
+	t.Run("explicit system prompt takes precedence", func(t *testing.T) {
+		viper.Reset()
+
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				SystemPrompt: "Model-specific prompt",
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString:  "custom/test-model",
+			SystemPrompt: "Global prompt",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		// Global system prompt should NOT be overridden because config
+		// already has a non-empty SystemPrompt.
+		if config.SystemPrompt != "Global prompt" {
+			t.Errorf("expected global prompt preserved, got %q", config.SystemPrompt)
+		}
+	})
+
+	t.Run("system prompt from file path", func(t *testing.T) {
+		viper.Reset()
+
+		// Create a temp file with a system prompt
+		tmpFile, err := os.CreateTemp("", "kit-test-prompt-*.txt")
+		if err != nil {
+			t.Fatal(err)
+		}
+		defer func() { _ = os.Remove(tmpFile.Name()) }()
+		if _, err := tmpFile.WriteString("  Prompt from file  "); err != nil {
+			t.Fatal(err)
+		}
+		_ = tmpFile.Close()
+
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				SystemPrompt: tmpFile.Name(),
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		if config.SystemPrompt != "Prompt from file" {
+			t.Errorf("expected trimmed file content, got %q", config.SystemPrompt)
+		}
+	})
+
+	t.Run("modelSettings system prompt overrides custom model params", func(t *testing.T) {
+		viper.Reset()
+
+		viper.Set("modelSettings", map[string]any{
+			"custom/test-model": map[string]any{
+				"systemPrompt": "From modelSettings",
+			},
+		})
+
+		modelInfo := &ModelInfo{
+			ID: "test-model",
+			Params: &GenerationParams{
+				SystemPrompt: "From custom model",
+			},
+		}
+
+		config := &ProviderConfig{
+			ModelString: "custom/test-model",
+		}
+
+		ApplyModelSettings(config, modelInfo)
+
+		if config.SystemPrompt != "From modelSettings" {
+			t.Errorf("expected modelSettings prompt, got %q", config.SystemPrompt)
+		}
+	})
+}

internal/models/modelsdb.go

@@ -48,18 +48,28 @@ type modelsDBLimit struct {
 	Output  int `json:"output"`
 }
 
-// npmToLLMProvider maps npm package names from models.dev to LLM
-// provider identifiers. Providers not in this map but with an api URL
-// can be auto-routed through openaicompat.
-var npmToLLMProvider = map[string]string{
-	"@ai-sdk/anthropic":               "anthropic",
-	"@ai-sdk/openai":                  "openai",
-	"@ai-sdk/google":                  "google",
-	"@ai-sdk/google-vertex":           "google-vertex",
-	"@ai-sdk/google-vertex/anthropic": "google-vertex-anthropic",
-	"@ai-sdk/amazon-bedrock":          "bedrock",
-	"@ai-sdk/azure":                   "azure",
-	"@openrouter/ai-sdk-provider":     "openrouter",
-	"@ai-sdk/vercel":                  "vercel",
-	"@ai-sdk/openai-compatible":       "openaicompat",
+// wireProtocol identifies which LLM API protocol an npm package speaks.
+// Fantasy implements three native protocols (openai, anthropic, google);
+// everything else in its providers/ tree is a thin wrapper around one of
+// them with a pre-baked default URL or auth scheme.
+type wireProtocol int
+
+const (
+	wireUnknown wireProtocol = iota
+	wireOpenAI
+	wireAnthropic
+	wireGoogle
+)
+
+// npmToWireProtocol maps npm package names from models.dev to the wire
+// protocol they speak. Provider-specific bundles (azure, bedrock, vercel,
+// openrouter, google-vertex, google-vertex-anthropic) are intentionally
+// absent — they have native top-level cases in CreateProvider and never
+// reach the auto-router. Providers not in this map but with an api URL
+// are auto-routed through the OpenAI-compatible wire.
+var npmToWireProtocol = map[string]wireProtocol{
+	"@ai-sdk/openai":            wireOpenAI,
+	"@ai-sdk/openai-compatible": wireOpenAI,
+	"@ai-sdk/anthropic":         wireAnthropic,
+	"@ai-sdk/google":            wireGoogle,
 }

internal/models/pool.go

@@ -1,168 +0,0 @@
-package models
-
-import (
-	"context"
-	"sync"
-	"time"
-
-	"charm.land/fantasy"
-)
-
-// ProviderPool manages reusable LLM provider instances to reduce overhead
-// when spawning multiple subagents or making repeated completion calls.
-type ProviderPool struct {
-	mu        sync.RWMutex
-	providers map[string]*pooledProvider
-	ttl       time.Duration
-	closed    bool
-	closeCh   chan struct{}
-}
-
-type pooledProvider struct {
-	model        fantasy.LanguageModel
-	closer       func() error
-	providerOpts fantasy.ProviderOptions
-	created      time.Time
-	lastUsed     time.Time
-	refs         int32
-}
-
-// DefaultPoolTTL is the default time-to-live for idle pooled providers.
-const DefaultPoolTTL = 5 * time.Minute
-
-// globalPool is the singleton provider pool instance.
-var globalPool *ProviderPool
-var poolOnce sync.Once
-
-// GetGlobalPool returns the singleton provider pool instance.
-func GetGlobalPool() *ProviderPool {
-	poolOnce.Do(func() {
-		globalPool = NewProviderPool(DefaultPoolTTL)
-	})
-	return globalPool
-}
-
-// NewProviderPool creates a provider pool with the given TTL for idle providers.
-func NewProviderPool(ttl time.Duration) *ProviderPool {
-	p := &ProviderPool{
-		providers: make(map[string]*pooledProvider),
-		ttl:       ttl,
-		closeCh:   make(chan struct{}),
-	}
-	go p.cleanupLoop()
-	return p
-}
-
-// Get returns a provider for the model string, creating one if needed.
-// The returned release function must be called when the provider is no longer
-// needed. The provider may be reused by subsequent Get calls.
-func (p *ProviderPool) Get(ctx context.Context, modelString string) (fantasy.LanguageModel, fantasy.ProviderOptions, func(), error) {
-	p.mu.Lock()
-
-	// Check if we have an existing provider.
-	if pp, ok := p.providers[modelString]; ok {
-		pp.refs++
-		pp.lastUsed = time.Now()
-		p.mu.Unlock()
-		return pp.model, pp.providerOpts, func() { p.release(modelString) }, nil
-	}
-
-	p.mu.Unlock()
-
-	// Create a new provider outside the lock.
-	config := &ProviderConfig{ModelString: modelString}
-	result, err := CreateProvider(ctx, config)
-	if err != nil {
-		return nil, nil, nil, err
-	}
-
-	p.mu.Lock()
-	defer p.mu.Unlock()
-
-	// Double-check: another goroutine may have created one while we were unlocked.
-	if pp, ok := p.providers[modelString]; ok {
-		// Close the one we just created and use the existing one.
-		if result.Closer != nil {
-			_ = result.Closer.Close()
-		}
-		pp.refs++
-		pp.lastUsed = time.Now()
-		return pp.model, pp.providerOpts, func() { p.release(modelString) }, nil
-	}
-
-	var closerFn func() error
-	if result.Closer != nil {
-		closerFn = result.Closer.Close
-	}
-
-	pp := &pooledProvider{
-		model:        result.Model,
-		closer:       closerFn,
-		providerOpts: result.ProviderOptions,
-		created:      time.Now(),
-		lastUsed:     time.Now(),
-		refs:         1,
-	}
-	p.providers[modelString] = pp
-
-	return pp.model, pp.providerOpts, func() { p.release(modelString) }, nil
-}
-
-func (p *ProviderPool) release(modelString string) {
-	p.mu.Lock()
-	defer p.mu.Unlock()
-
-	if pp, ok := p.providers[modelString]; ok {
-		pp.refs--
-		pp.lastUsed = time.Now()
-	}
-}
-
-func (p *ProviderPool) cleanupLoop() {
-	ticker := time.NewTicker(p.ttl / 2)
-	defer ticker.Stop()
-
-	for {
-		select {
-		case <-p.closeCh:
-			return
-		case <-ticker.C:
-			p.cleanup()
-		}
-	}
-}
-
-func (p *ProviderPool) cleanup() {
-	p.mu.Lock()
-	defer p.mu.Unlock()
-
-	now := time.Now()
-	for key, pp := range p.providers {
-		// Only clean up providers with no active references and past TTL.
-		if pp.refs <= 0 && now.Sub(pp.lastUsed) > p.ttl {
-			if pp.closer != nil {
-				_ = pp.closer()
-			}
-			delete(p.providers, key)
-		}
-	}
-}
-
-// Close shuts down the pool and releases all providers.
-func (p *ProviderPool) Close() {
-	p.mu.Lock()
-	if p.closed {
-		p.mu.Unlock()
-		return
-	}
-	p.closed = true
-	close(p.closeCh)
-
-	for key, pp := range p.providers {
-		if pp.closer != nil {
-			_ = pp.closer()
-		}
-		delete(p.providers, key)
-	}
-	p.mu.Unlock()
-}

internal/models/providers.go

@@ -9,6 +9,7 @@ import (
 	"io"
 	"maps"
 	"net/http"
+	"net/url"
 	"os"
 	"regexp"
 	"strings"
@@ -26,7 +27,7 @@ import (
 	openaisdk "github.com/charmbracelet/openai-go"
 
 	"github.com/mark3labs/kit/internal/auth"
-	"github.com/mark3labs/kit/internal/ui/progress"
+	"github.com/spf13/viper"
 )
 
 const (
@@ -87,6 +88,7 @@ type ThinkingLevel string
 
 const (
 	ThinkingOff     ThinkingLevel = "off"
+	ThinkingNone    ThinkingLevel = "none"
 	ThinkingMinimal ThinkingLevel = "minimal"
 	ThinkingLow     ThinkingLevel = "low"
 	ThinkingMedium  ThinkingLevel = "medium"
@@ -95,12 +97,14 @@ const (
 
 // ThinkingLevels returns the ordered list of available thinking levels for cycling.
 func ThinkingLevels() []ThinkingLevel {
-	return []ThinkingLevel{ThinkingOff, ThinkingMinimal, ThinkingLow, ThinkingMedium, ThinkingHigh}
+	return []ThinkingLevel{ThinkingOff, ThinkingNone, ThinkingMinimal, ThinkingLow, ThinkingMedium, ThinkingHigh}
 }
 
-// thinkingBudgetTokens returns the token budget for a thinking level, or 0 for "off".
+// thinkingBudgetTokens returns the token budget for a thinking level, or 0 for "off" or "none".
 func thinkingBudgetTokens(level ThinkingLevel) int64 {
 	switch level {
+	case ThinkingNone:
+		return 1024
 	case ThinkingMinimal:
 		return 1024
 	case ThinkingLow:
@@ -119,6 +123,8 @@ func ThinkingLevelDescription(level ThinkingLevel) string {
 	switch level {
 	case ThinkingOff:
 		return "No reasoning"
+	case ThinkingNone:
+		return "Minimal reasoning (OpenAI 'none')"
 	case ThinkingMinimal:
 		return "Very brief reasoning (~1k tokens)"
 	case ThinkingLow:
@@ -135,7 +141,7 @@ func ThinkingLevelDescription(level ThinkingLevel) string {
 // ParseThinkingLevel converts a string to a ThinkingLevel, defaulting to ThinkingOff.
 func ParseThinkingLevel(s string) ThinkingLevel {
 	switch ThinkingLevel(s) {
-	case ThinkingMinimal, ThinkingLow, ThinkingMedium, ThinkingHigh:
+	case ThinkingNone, ThinkingMinimal, ThinkingLow, ThinkingMedium, ThinkingHigh:
 		return ThinkingLevel(s)
 	default:
 		return ThinkingOff
@@ -144,20 +150,35 @@ func ParseThinkingLevel(s string) ThinkingLevel {
 
 // ProviderConfig holds configuration for creating LLM providers.
 type ProviderConfig struct {
-	ModelString    string
-	SystemPrompt   string
-	ProviderAPIKey string
-	ProviderURL    string
-	MaxTokens      int
-	Temperature    *float32
-	TopP           *float32
-	TopK           *int32
-	StopSequences  []string
-	NumGPU         *int32
-	MainGPU        *int32
-	TLSSkipVerify  bool
-	ThinkingLevel  ThinkingLevel
-	DisableCaching bool // Opt-out: set to true to disable automatic prompt caching
+	ModelString      string
+	SystemPrompt     string
+	ProviderAPIKey   string
+	ProviderURL      string
+	MaxTokens        int
+	Temperature      *float32
+	TopP             *float32
+	TopK             *int32
+	FrequencyPenalty *float32
+	PresencePenalty  *float32
+	StopSequences    []string
+	NumGPU           *int32
+	MainGPU          *int32
+	TLSSkipVerify    bool
+	ThinkingLevel    ThinkingLevel
+	DisableCaching   bool // Opt-out: set to true to disable automatic prompt caching
+
+	// ConfigStore is the per-instance configuration store used to resolve
+	// "explicitly set" precedence checks (isExplicitlySet), per-model
+	// settings, and right-sizing. When nil, the process-global viper store is
+	// used. Threading a per-Kit store here keeps generation-parameter
+	// precedence isolated between Kit instances in the same process.
+	ConfigStore *viper.Viper
+
+	// ProgressReaderFunc, when set, wraps an io.Reader with progress display
+	// for long operations like Ollama model pulls. The returned io.ReadCloser
+	// must be closed when done. When nil, the raw reader is consumed directly
+	// with no progress UI.
+	ProgressReaderFunc func(io.Reader) io.ReadCloser
 }
 
 // ProviderResult contains the result of provider creation.
@@ -201,8 +222,10 @@ func ParseModelString(modelString string) (provider, model string, err error) {
 //
 // Native providers: anthropic, openai, google, ollama, azure, google-vertex-anthropic,
 // openrouter, bedrock, vercel.
-// Any provider in models.dev with an api URL or openai-compatible npm package
-// is auto-routed through fantasy's openaicompat provider.
+// Any other provider in models.dev is auto-routed by wire protocol: its npm
+// package (or per-model override) selects the OpenAI, Anthropic, or Google
+// transport, using the provider's api URL as the base. Providers with an api
+// URL but an unrecognized npm package fall back to the OpenAI-compatible wire.
 func CreateProvider(ctx context.Context, config *ProviderConfig) (*ProviderResult, error) {
 	provider, modelName, err := ParseModelString(config.ModelString)
 	if err != nil {
@@ -240,6 +263,16 @@ func CreateProvider(ctx context.Context, config *ProviderConfig) (*ProviderResul
 		validateModelConfig(config, modelInfo)
 	}
 
+	// Apply per-model generation parameter defaults. Model-level params are
+	// only applied for fields where the user hasn't explicitly set a value
+	// via CLI flag or global config.
+	ApplyModelSettings(config, modelInfo)
+
+	// Auto-raise MaxTokens toward the model's known output ceiling when the
+	// user hasn't explicitly set --max-tokens and no per-model override
+	// applied. Runs after ApplyModelSettings so explicit modelSettings win.
+	rightSizeMaxTokens(config, modelInfo)
+
 	// Create the base provider
 	var result *ProviderResult
 	var createErr error
@@ -284,9 +317,18 @@ func CreateProvider(ctx context.Context, config *ProviderConfig) (*ProviderResul
 			// Only add cache options for providers that don't already have
 			// options set, to avoid type conflicts (e.g., Anthropic has
 			// different types for regular options vs cache control options).
-			for k, v := range cacheOpts {
-				if _, exists := result.ProviderOptions[k]; !exists {
-					result.ProviderOptions[k] = v
+			//
+			// For OpenAI Responses API models, we skip merging entirely because
+			// ResponsesProviderOptions and ProviderOptions are incompatible types.
+			skipMerge := false
+			if provider == "openai" && openai.IsResponsesModel(modelName) {
+				skipMerge = true
+			}
+			if !skipMerge {
+				for k, v := range cacheOpts {
+					if _, exists := result.ProviderOptions[k]; !exists {
+						result.ProviderOptions[k] = v
+					}
 				}
 			}
 		}
@@ -297,43 +339,62 @@ func CreateProvider(ctx context.Context, config *ProviderConfig) (*ProviderResul
 
 // autoRouteProvider attempts to create a provider by looking up its npm package
 // in the models.dev database and routing through the appropriate fantasy provider.
-// For openai-compatible providers, it uses the api URL from models.dev.
-// Models may have a provider override that specifies a different npm package than
-// the provider's default (e.g., opencode's claude-opus-4-6 uses @ai-sdk/anthropic).
+// It routes on wire protocol (openai, anthropic, google) rather than per-npm
+// provider name: fantasy implements three native wire protocols, and every other
+// entry in its providers/ tree is a thin wrapper around one of them. Using the
+// provider's api URL from models.dev as the base URL, any proxy that re-flavors
+// one of these protocols (e.g. opencode's Gemini routes) Just Works.
+//
+// Models may carry a provider override that specifies a different npm package
+// than the provider's default (e.g. opencode's claude-* uses @ai-sdk/anthropic
+// and its gemini-* uses @ai-sdk/google), which is resolved first.
 func autoRouteProvider(ctx context.Context, config *ProviderConfig, provider, modelName string, registry *ModelsRegistry) (*ProviderResult, error) {
 	providerInfo := registry.GetProviderInfo(provider)
 	if providerInfo == nil {
 		return nil, fmt.Errorf("unsupported provider: %s (not found in model database)", provider)
 	}
 
-	// Check for model-specific provider override
+	// Resolve npm: per-model override > provider default.
 	npmPackage := providerInfo.NPM
 	if modelInfo := registry.LookupModel(provider, modelName); modelInfo != nil && modelInfo.ProviderNPM != "" {
 		npmPackage = modelInfo.ProviderNPM
 	}
 
-	// Determine the LLM provider for this npm package
-	llmProvider := npmToLLMProvider[npmPackage]
-	if llmProvider == "" && providerInfo.API != "" {
-		// Unknown npm but has API URL → route through openaicompat
-		llmProvider = "openaicompat"
+	wire, known := npmToWireProtocol[npmPackage]
+	if !known {
+		// Unknown npm but the provider has an API URL → assume OpenAI-compatible.
+		// (Preserves the long-standing "any provider in models.dev with an api URL
+		// is auto-routed through openaicompat" behaviour.)
+		if providerInfo.API == "" {
+			return nil, fmt.Errorf(
+				"cannot auto-route provider %s: npm package %q has no known wire protocol "+
+					"and the registry has no API URL (use --provider-url to override)",
+				provider, npmPackage,
+			)
+		}
+		wire = wireOpenAI
 	}
 
-	switch llmProvider {
-	case "openaicompat":
+	// All three wires use the provider's API URL from models.dev as the base.
+	if config.ProviderURL == "" && providerInfo.API != "" {
+		config.ProviderURL = providerInfo.API
+	}
+
+	switch wire {
+	case wireOpenAI:
+		// The native OpenAI SDK package (@ai-sdk/openai) speaks the Responses
+		// API; openai-compatible proxies (and unknown-npm fallbacks) use the
+		// chat-completions wire via fantasy's openaicompat provider.
+		if npmPackage == "@ai-sdk/openai" {
+			return createAutoRoutedOpenAIProvider(ctx, config, modelName, providerInfo)
+		}
 		return createAutoRoutedOpenAICompatProvider(ctx, config, modelName, providerInfo)
-	case "anthropic":
-		if config.ProviderURL == "" && providerInfo.API != "" {
-			config.ProviderURL = providerInfo.API
-		}
+	case wireAnthropic:
 		return createAutoRoutedAnthropicProvider(ctx, config, modelName, providerInfo)
-	case "openai":
-		if config.ProviderURL == "" && providerInfo.API != "" {
-			config.ProviderURL = providerInfo.API
-		}
-		return createAutoRoutedOpenAIProvider(ctx, config, modelName, providerInfo)
+	case wireGoogle:
+		return createAutoRoutedGoogleProvider(ctx, config, modelName, providerInfo)
 	default:
-		return nil, fmt.Errorf("unsupported provider: %s (npm: %s has no LLM provider mapping)", provider, npmPackage)
+		return nil, fmt.Errorf("internal error: unknown wire protocol for provider %s (npm: %s)", provider, npmPackage)
 	}
 }
 
@@ -450,6 +511,115 @@ func createAutoRoutedOpenAIProvider(ctx context.Context, config *ProviderConfig,
 	return &ProviderResult{Model: model, ProviderOptions: providerOpts}, nil
 }
 
+// createAutoRoutedGoogleProvider creates a Google (Gemini) provider for
+// third-party providers that expose a Gemini-compatible API (e.g. opencode's
+// Gemini routes, which carry an @ai-sdk/google per-model override).
+//
+// The underlying genai SDK always injects its own API version segment
+// ("v1beta") between the base URL and the resource path. When the proxy's
+// base URL from models.dev already carries a version segment (e.g. opencode's
+// https://opencode.ai/zen/v1), that produces a doubled ".../v1/v1beta/..."
+// path that the proxy rejects. In that case we install a transport that
+// strips the injected segment so the proxy's own version is used.
+func createAutoRoutedGoogleProvider(ctx context.Context, config *ProviderConfig, modelName string, info *ProviderInfo) (*ProviderResult, error) {
+	apiKey := resolveAPIKey(config.ProviderAPIKey, info.Env)
+	if apiKey == "" {
+		return nil, fmt.Errorf("%s API key not provided. Use --provider-api-key or set %s",
+			info.Name, strings.Join(info.Env, " / "))
+	}
+
+	opts := []google.Option{
+		google.WithGeminiAPIKey(apiKey),
+		google.WithName(info.ID),
+	}
+
+	if config.ProviderURL != "" {
+		opts = append(opts, google.WithBaseURL(config.ProviderURL))
+	}
+
+	// Decide whether the genai-injected version segment needs stripping.
+	var httpClient *http.Client
+	if basePath := versionedBasePath(config.ProviderURL); basePath != "" {
+		httpClient = newGeminiProxyHTTPClient(basePath, config.TLSSkipVerify)
+	} else if config.TLSSkipVerify {
+		httpClient = createHTTPClientWithTLSConfig(true)
+	}
+	if httpClient != nil {
+		opts = append(opts, google.WithHTTPClient(httpClient))
+	}
+
+	p, err := google.New(opts...)
+	if err != nil {
+		return nil, fmt.Errorf("failed to create %s provider: %w", info.Name, err)
+	}
+
+	model, err := p.LanguageModel(ctx, modelName)
+	if err != nil {
+		return nil, fmt.Errorf("failed to create %s model: %w", info.Name, err)
+	}
+
+	return &ProviderResult{Model: model}, nil
+}
+
+// versionSegmentRe matches a trailing API version segment in a URL path,
+// e.g. "/v1", "/v1beta", "/v1beta1", "/v2alpha".
+var versionSegmentRe = regexp.MustCompile(`/v\d+(?:beta\d*|alpha\d*)?$`)
+
+// versionedBasePath returns the path component of rawURL when that path ends
+// with an API version segment (e.g. opencode's ".../zen/v1" → "/zen/v1").
+// It returns "" when rawURL is empty, unparseable, or has no version suffix
+// — in which case the genai SDK's default version injection is correct and
+// no rewriting is needed.
+func versionedBasePath(rawURL string) string {
+	if rawURL == "" {
+		return ""
+	}
+	u, err := url.Parse(rawURL)
+	if err != nil {
+		return ""
+	}
+	path := strings.TrimSuffix(u.Path, "/")
+	if versionSegmentRe.MatchString(path) {
+		return path
+	}
+	return ""
+}
+
+// newGeminiProxyHTTPClient builds an HTTP client whose transport strips the
+// genai-injected version segment ("v1beta"/"v1beta1") that directly follows
+// basePath, collapsing "{basePath}/v1beta/..." back to "{basePath}/...".
+func newGeminiProxyHTTPClient(basePath string, skipVerify bool) *http.Client {
+	var base http.RoundTripper
+	if skipVerify {
+		base = &http.Transport{TLSClientConfig: &tls.Config{InsecureSkipVerify: true}}
+	} else {
+		base = http.DefaultTransport
+	}
+	return &http.Client{
+		Transport: &geminiProxyTransport{base: base, basePath: basePath},
+	}
+}
+
+// geminiProxyTransport removes the redundant API version segment that the
+// genai SDK injects after a proxy base URL that already carries its own
+// version segment.
+type geminiProxyTransport struct {
+	base     http.RoundTripper
+	basePath string
+}
+
+func (t *geminiProxyTransport) RoundTrip(req *http.Request) (*http.Response, error) {
+	for _, injected := range []string{"/v1beta1", "/v1beta"} {
+		prefix := t.basePath + injected + "/"
+		if strings.HasPrefix(req.URL.Path, prefix) {
+			newReq := req.Clone(req.Context())
+			newReq.URL.Path = t.basePath + strings.TrimPrefix(req.URL.Path, t.basePath+injected)
+			return t.base.RoundTrip(newReq)
+		}
+	}
+	return t.base.RoundTrip(req)
+}
+
 // resolveAPIKey returns the first non-empty API key from the explicit key
 // or the environment variables.
 func resolveAPIKey(explicitKey string, envVars []string) string {
@@ -478,6 +648,37 @@ func validateModelConfig(config *ProviderConfig, modelInfo *ModelInfo) {
 	}
 }
 
+// defaultRightSizeCap bounds auto-raised MaxTokens so that we don't silently
+// allocate enormous output budgets for models with very high ceilings (e.g.
+// Devstral at 262144, Mistral at 128000). Users who genuinely want more can
+// pass --max-tokens explicitly or set modelSettings[...].maxTokens in config.
+const defaultRightSizeCap = 32768
+
+// rightSizeMaxTokens raises config.MaxTokens toward the model's known output
+// ceiling when:
+//   - the user has not explicitly set --max-tokens (or the KIT_MAX_TOKENS env
+//     var, or the top-level max-tokens key in config.yaml), AND
+//   - no per-model override already bumped MaxTokens (ApplyModelSettings runs
+//     before this function), AND
+//   - modelInfo.Limit.Output is known and larger than the current MaxTokens.
+//
+// The raised value is capped at defaultRightSizeCap to keep accidental
+// allocations reasonable on very-large-output models. This prevents the
+// common "ghost" where the agent's reply is silently truncated at the 8192
+// default even though the selected model supports 64k or 262k output tokens.
+func rightSizeMaxTokens(config *ProviderConfig, modelInfo *ModelInfo) {
+	if modelInfo == nil || modelInfo.Limit.Output <= 0 {
+		return
+	}
+	if isExplicitlySet(config.ConfigStore, "max-tokens") {
+		return
+	}
+	target := min(modelInfo.Limit.Output, defaultRightSizeCap)
+	if config.MaxTokens < target {
+		config.MaxTokens = target
+	}
+}
+
 // clearConflictingAnthropicSamplingParams ensures that temperature and top_p are
 // not both sent to the Anthropic API, which rejects requests containing both.
 // When both are set (typically from defaults), top_p is cleared so that
@@ -524,19 +725,71 @@ func buildOpenAIProviderOptions(config *ProviderConfig, modelName string) fantas
 // Returns nil for ThinkingOff (use the model's default).
 func thinkingLevelToReasoningEffort(level ThinkingLevel) *openai.ReasoningEffort {
 	switch level {
+	case ThinkingNone:
+		return new(openai.ReasoningEffortNone)
 	case ThinkingMinimal:
-		return openai.ReasoningEffortOption(openai.ReasoningEffortMinimal)
+		return new(openai.ReasoningEffortMinimal)
 	case ThinkingLow:
-		return openai.ReasoningEffortOption(openai.ReasoningEffortLow)
+		return new(openai.ReasoningEffortLow)
 	case ThinkingMedium:
-		return openai.ReasoningEffortOption(openai.ReasoningEffortMedium)
+		return new(openai.ReasoningEffortMedium)
 	case ThinkingHigh:
-		return openai.ReasoningEffortOption(openai.ReasoningEffortHigh)
+		return new(openai.ReasoningEffortHigh)
 	default:
 		return nil
 	}
 }
 
+// IsValidThinkingLevelForModel checks if a thinking level is valid for the given
+// model. Some OpenAI models like gpt-5.4 don't support "minimal" and require
+// "none" instead.
+func IsValidThinkingLevelForModel(level ThinkingLevel, modelName string) bool {
+	if level == ThinkingOff {
+		return true
+	}
+
+	// Check if this is an OpenAI model that doesn't support "minimal"
+	// gpt-5.4 and newer gpt-5.x models use "none" instead of "minimal"
+	if level == ThinkingMinimal {
+		if strings.Contains(modelName, "gpt-5.4") ||
+			strings.Contains(modelName, "gpt-5-pro") ||
+			strings.Contains(modelName, "gpt-5-chat") {
+			return false
+		}
+	}
+
+	// Check if this is an OpenAI model that doesn't support "none"
+	// Older gpt-5 models only support "minimal", not "none"
+	if level == ThinkingNone {
+		if strings.Contains(modelName, "gpt-5") &&
+			!strings.Contains(modelName, "gpt-5.4") &&
+			!strings.Contains(modelName, "gpt-5-pro") &&
+			!strings.Contains(modelName, "gpt-5-chat") {
+			// Older gpt-5 models might not support "none"
+			// They only added "none" support in newer versions
+			return false
+		}
+	}
+
+	// All other levels are generally valid for reasoning models
+	return true
+}
+
+// SuggestThinkingLevelFallback returns a recommended fallback level when the
+// requested level is not valid for the model. Returns ThinkingOff if no
+// suitable fallback exists.
+func SuggestThinkingLevelFallback(level ThinkingLevel, modelName string) ThinkingLevel {
+	if level == ThinkingMinimal && !IsValidThinkingLevelForModel(level, modelName) {
+		// For models that don't support "minimal", suggest "none" (~same token budget)
+		return ThinkingNone
+	}
+	if level == ThinkingNone && !IsValidThinkingLevelForModel(level, modelName) {
+		// For models that don't support "none", suggest "minimal" (~same token budget)
+		return ThinkingMinimal
+	}
+	return ThinkingOff
+}
+
 // buildAnthropicProviderOptions returns fantasy.ProviderOptions configured for
 // Anthropic models with extended thinking. When thinking is enabled, it sets
 // SendReasoning to true and configures the thinking budget. For thinking-off
@@ -1000,139 +1253,29 @@ func createVercelProvider(ctx context.Context, config *ProviderConfig, modelName
 	return &ProviderResult{Model: model}, nil
 }
 
-// thinkTagRegex matches <think>...</think> tags for extracting reasoning content
-// from models that wrap thinking in XML-like tags (e.g., Qwen, DeepSeek).
-var thinkTagRegex = regexp.MustCompile(`(?s)<think>(.*?)</think>`)
-
-// customExtraContentFunc extracts reasoning from <think> tags in the content field.
-// This handles models like Qwen and DeepSeek that return reasoning wrapped in XML tags
-// rather than using a separate reasoning_content field.
-func customExtraContentFunc(choice openaisdk.ChatCompletionChoice) []fantasy.Content {
-	var content []fantasy.Content
-	if choice.Message.Content == "" {
-		return content
-	}
-
-	// Check for <think> tags in the content
-	matches := thinkTagRegex.FindStringSubmatch(choice.Message.Content)
-	if len(matches) > 1 {
-		// Found reasoning content in <think> tags
-		reasoning := strings.TrimSpace(matches[1])
-		if reasoning != "" {
-			content = append(content, fantasy.ReasoningContent{
-				Text: reasoning,
-			})
-		}
-	}
-
-	return content
-}
-
-// customStreamExtraFunc handles streaming responses with <think> tags.
-// It extracts reasoning content and emits proper reasoning events.
-func customStreamExtraFunc(
-	chunk openaisdk.ChatCompletionChunk,
-	yield func(fantasy.StreamPart) bool,
-	ctx map[string]any,
-) (map[string]any, bool) {
-	if len(chunk.Choices) == 0 {
-		return ctx, true
-	}
-
-	const reasoningStartedKey = "reasoning_started"
-	const reasoningBufferKey = "reasoning_buffer"
-	const inThinkTagKey = "in_think_tag"
-
-	reasoningStarted, _ := ctx[reasoningStartedKey].(bool)
-	inThinkTag, _ := ctx[inThinkTagKey].(bool)
-	reasoningBuffer, _ := ctx[reasoningBufferKey].(string)
-
-	for i, choice := range chunk.Choices {
-		content := choice.Delta.Content
-		if content == "" {
-			continue
-		}
-
-		// Check for <think> tag start
-		if strings.Contains(content, "<think>") {
-			inThinkTag = true
-			ctx[inThinkTagKey] = true
-
-			// Emit reasoning start event
-			if !reasoningStarted {
-				reasoningStarted = true
-				ctx[reasoningStartedKey] = true
-				if !yield(fantasy.StreamPart{
-					Type: fantasy.StreamPartTypeReasoningStart,
-					ID:   fmt.Sprintf("%d", i),
-				}) {
-					return ctx, false
-				}
-			}
-
-			// Extract content after <think>
-			parts := strings.SplitN(content, "<think>", 2)
-			if len(parts) > 1 && parts[1] != "" {
-				reasoningBuffer += parts[1]
-				ctx[reasoningBufferKey] = reasoningBuffer
-			}
-			continue
-		}
-
-		// Check for </think> tag end
-		if strings.Contains(content, "</think>") {
-			inThinkTag = false
-			ctx[inThinkTagKey] = false
-
-			// Extract content before </think>
-			parts := strings.SplitN(content, "</think>", 2)
-			if len(parts) > 0 {
-				reasoningBuffer += parts[0]
-			}
-
-			// Emit the accumulated reasoning
-			if reasoningBuffer != "" {
-				if !yield(fantasy.StreamPart{
-					Type:  fantasy.StreamPartTypeReasoningDelta,
-					ID:    fmt.Sprintf("%d", i),
-					Delta: reasoningBuffer,
-				}) {
-					return ctx, false
-				}
-				ctx[reasoningBufferKey] = ""
-			}
-
-			// Emit reasoning end
-			if !yield(fantasy.StreamPart{
-				Type: fantasy.StreamPartTypeReasoningEnd,
-				ID:   fmt.Sprintf("%d", i),
-			}) {
-				return ctx, false
-			}
-			continue
-		}
-
-		// Accumulate reasoning content while in think tag
-		if inThinkTag {
-			reasoningBuffer += content
-			ctx[reasoningBufferKey] = reasoningBuffer
-		}
-	}
-
-	return ctx, true
-}
-
 // customToPromptFunc converts prompts to OpenAI format using the default conversion.
 func customToPromptFunc(prompt fantasy.Prompt, systemPrompt, user string) ([]openaisdk.ChatCompletionMessageParamUnion, []fantasy.CallWarning) {
 	return openai.DefaultToPrompt(prompt, systemPrompt, user)
 }
 
 func createCustomProvider(ctx context.Context, config *ProviderConfig, modelName string) (*ProviderResult, error) {
-	if config.ProviderURL == "" {
-		return nil, fmt.Errorf("custom provider requires --provider-url")
+	// Resolve base URL: per-model override > global provider-url flag/config
+	registry := GetGlobalRegistry()
+	modelInfo := registry.LookupModel("custom", modelName)
+
+	baseURL := config.ProviderURL
+	if modelInfo != nil && modelInfo.BaseURL != "" {
+		baseURL = modelInfo.BaseURL
+	}
+
+	if baseURL == "" {
+		return nil, fmt.Errorf("custom provider requires --provider-url or a baseUrl in the model config")
 	}
 
 	apiKey := config.ProviderAPIKey
+	if modelInfo != nil && modelInfo.APIKey != "" {
+		apiKey = modelInfo.APIKey
+	}
 	if apiKey == "" {
 		apiKey = os.Getenv("CUSTOM_API_KEY")
 	}
@@ -1141,15 +1284,13 @@ func createCustomProvider(ctx context.Context, config *ProviderConfig, modelName
 		apiKey = "custom"
 	}
 
-	// Use the openai provider directly with custom hooks to handle <think> tags
-	// from models like Qwen and DeepSeek that wrap reasoning in XML tags.
+	// <think> tag extraction is handled transparently at the agent layer,
+	// so no provider-level hooks are needed here.
 	var opts []openai.Option
-	opts = append(opts, openai.WithBaseURL(config.ProviderURL))
+	opts = append(opts, openai.WithBaseURL(baseURL))
 	opts = append(opts, openai.WithAPIKey(apiKey))
 	opts = append(opts, openai.WithName("custom"))
 	opts = append(opts, openai.WithLanguageModelOptions(
-		openai.WithLanguageModelExtraContentFunc(customExtraContentFunc),
-		openai.WithLanguageModelStreamExtraFunc(customStreamExtraFunc),
 		openai.WithLanguageModelToPromptFunc(customToPromptFunc),
 	))
 
@@ -1234,7 +1375,7 @@ func loadOllamaModelWithFallback(ctx context.Context, baseURL, modelName string,
 	// Phase 1: Check if model exists locally
 	if err := checkOllamaModelExists(client, baseURL, modelName); err != nil {
 		// Phase 2: Pull model if not found
-		if err := pullOllamaModel(ctx, client, baseURL, modelName); err != nil {
+		if err := pullOllamaModel(ctx, client, baseURL, modelName, config.ProgressReaderFunc); err != nil {
 			return nil, fmt.Errorf("failed to pull model %s: %v", modelName, err)
 		}
 	}
@@ -1277,6 +1418,12 @@ func buildOllamaOptions(config *ProviderConfig) map[string]any {
 	if config.TopK != nil {
 		options["top_k"] = int(*config.TopK)
 	}
+	if config.FrequencyPenalty != nil {
+		options["frequency_penalty"] = *config.FrequencyPenalty
+	}
+	if config.PresencePenalty != nil {
+		options["presence_penalty"] = *config.PresencePenalty
+	}
 	if len(config.StopSequences) > 0 {
 		options["stop"] = config.StopSequences
 	}
@@ -1317,11 +1464,7 @@ func checkOllamaModelExists(client *http.Client, baseURL, modelName string) erro
 	return nil
 }
 
-func pullOllamaModel(ctx context.Context, client *http.Client, baseURL, modelName string) error {
-	return pullOllamaModelWithProgress(ctx, client, baseURL, modelName, true)
-}
-
-func pullOllamaModelWithProgress(ctx context.Context, client *http.Client, baseURL, modelName string, showProgress bool) error {
+func pullOllamaModel(ctx context.Context, client *http.Client, baseURL, modelName string, progressFn func(io.Reader) io.ReadCloser) error {
 	reqBody := map[string]string{"name": modelName}
 	jsonBody, _ := json.Marshal(reqBody)
 
@@ -1345,10 +1488,10 @@ func pullOllamaModelWithProgress(ctx context.Context, client *http.Client, baseU
 		return fmt.Errorf("failed to pull model (status %d): %s", resp.StatusCode, string(body))
 	}
 
-	if showProgress {
-		progressReader := progress.NewProgressReader(resp.Body)
-		defer func() { _ = progressReader.Close() }()
-		_, err = io.ReadAll(progressReader)
+	if progressFn != nil {
+		pr := progressFn(resp.Body)
+		defer func() { _ = pr.Close() }()
+		_, err = io.ReadAll(pr)
 	} else {
 		_, err = io.ReadAll(resp.Body)
 	}

internal/models/registry.go

@@ -4,6 +4,7 @@ import (
 	_ "embed"
 	"encoding/json"
 	"fmt"
+	"maps"
 	"os"
 	"strings"
 
@@ -24,6 +25,13 @@ type ModelInfo struct {
 	Cost        Cost
 	Limit       Limit
 	ProviderNPM string // Model-specific provider npm override (e.g. "@ai-sdk/anthropic")
+	BaseURL     string // Per-model base URL override (custom models only)
+	APIKey      string // Per-model API key override (custom models only)
+
+	// Params holds per-model generation parameter defaults. These are applied
+	// when the user hasn't explicitly set the corresponding CLI flag or global
+	// config value. Nil pointer fields mean "no model-level default".
+	Params *GenerationParams
 }
 
 // SupportsCaching returns true if this model family supports prompt caching.
@@ -104,13 +112,30 @@ func NewModelsRegistry() *ModelsRegistry {
 }
 
 // buildFromModelsDB converts models.dev provider data into our internal format.
-// It tries the on-disk cache first and falls back to the embedded database.
+// It starts from the compile-time embedded database and merges on-disk cached
+// data from `kit update-models` on top. Cached provider metadata replaces
+// embedded metadata, and model entries are merged with cached models taking
+// precedence. This means newly synced models are available while embedded
+// models that haven't been synced yet are still reachable.
 func buildFromModelsDB() map[string]ProviderInfo {
-	// Try cached data first (from `kit update-models`)
-	dbProviders, _ := LoadCachedProviders()
-	if len(dbProviders) == 0 {
-		// Fall back to compile-time embedded data
-		dbProviders = loadEmbeddedProviders()
+	// Start with compile-time embedded data as the base.
+	dbProviders := loadEmbeddedProviders()
+	if dbProviders == nil {
+		dbProviders = make(ModelsDBProviders)
+	}
+
+	// Merge on-disk cached data on top (cached takes precedence).
+	if cached, _ := LoadCachedProviders(); len(cached) > 0 {
+		for providerID, cp := range cached {
+			if existing, ok := dbProviders[providerID]; ok {
+				// Merge models: embedded base + cached overrides.
+				mergedModels := make(map[string]modelsDBModel, len(existing.Models)+len(cp.Models))
+				maps.Copy(mergedModels, existing.Models)
+				maps.Copy(mergedModels, cp.Models)
+				cp.Models = mergedModels
+			}
+			dbProviders[providerID] = cp
+		}
 	}
 
 	providers := make(map[string]ProviderInfo, len(dbProviders))
@@ -234,6 +259,18 @@ func (r *ModelsRegistry) LookupModel(provider, modelID string) *ModelInfo {
 	return &modelInfo
 }
 
+// LookupModelForSettings is a convenience function that parses a
+// "provider/model" string and looks up the ModelInfo in the global registry.
+// Returns nil when the model string is invalid or the model is unknown.
+// Used by Kit.SetModel to pre-apply per-model settings before CreateProvider.
+func LookupModelForSettings(modelString string) *ModelInfo {
+	provider, modelName, err := ParseModelString(modelString)
+	if err != nil {
+		return nil
+	}
+	return GetGlobalRegistry().LookupModel(provider, modelName)
+}
+
 // getRequiredEnvVars returns the required environment variables for a provider.
 func (r *ModelsRegistry) getRequiredEnvVars(provider string) ([]string, error) {
 	providerInfo, exists := r.providers[provider]
@@ -360,20 +397,15 @@ func (r *ModelsRegistry) GetLLMProviders() []string {
 	return providers
 }
 
-// Deprecated: Use GetLLMProviders instead.
-func (r *ModelsRegistry) GetFantasyProviders() []string {
-	return r.GetLLMProviders()
-}
-
 // isProviderLLMSupported checks if a provider can be used with the LLM layer.
 func isProviderLLMSupported(providerID string, info *ProviderInfo) bool {
-	// Ollama is always supported (via openaicompat pointed at localhost)
-	if providerID == "ollama" {
+	// Ollama and custom are always supported (model names are user-defined).
+	if providerID == "ollama" || providerID == "custom" {
 		return true
 	}
 
-	// Check if npm maps to an LLM provider
-	if _, ok := npmToLLMProvider[info.NPM]; ok {
+	// Check if npm maps to a known wire protocol
+	if _, ok := npmToWireProtocol[info.NPM]; ok {
 		return true
 	}
 
@@ -400,6 +432,52 @@ func (r *ModelsRegistry) GetProviderInfo(provider string) *ProviderInfo {
 	return &info
 }
 
+// ValidateModelString checks whether a model string is well-formed and refers
+// to a known provider. It returns a user-friendly error with suggestions when
+// the model or provider is unrecognised. Passing validation does not guarantee
+// that API authentication will succeed — it only catches obvious mistakes
+// (typos, missing provider prefix, non-existent provider names) early so that
+// callers such as subagent spawning can return fast feedback.
+//
+// Unknown models under a known provider are allowed (the provider API is the
+// authority), but a completely unknown provider is rejected.
+func (r *ModelsRegistry) ValidateModelString(modelString string) error {
+	provider, modelName, err := ParseModelString(modelString)
+	if err != nil {
+		return err
+	}
+
+	// Ollama and custom are always valid — model names are user-defined.
+	if provider == "ollama" || provider == "custom" {
+		return nil
+	}
+
+	// Check if the provider exists in the registry.
+	providerInfo := r.GetProviderInfo(provider)
+	if providerInfo == nil {
+		known := r.GetSupportedProviders()
+		return fmt.Errorf(
+			"unknown provider %q in model string %q. Known providers: %s",
+			provider, modelString, strings.Join(known, ", "),
+		)
+	}
+
+	// Provider exists — check if the model is known. An unknown model is
+	// only a warning (the provider API decides), but we surface suggestions
+	// so the caller can self-correct.
+	if r.LookupModel(provider, modelName) == nil {
+		if suggestions := r.SuggestModels(provider, modelName); len(suggestions) > 0 {
+			return fmt.Errorf(
+				"model %q not found for provider %s. Did you mean one of: %s",
+				modelName, provider, strings.Join(suggestions, ", "),
+			)
+		}
+		// No suggestions — let it through; the provider API is the authority.
+	}
+
+	return nil
+}
+
 // Global registry instance
 var globalRegistry = NewModelsRegistry()
 

internal/models/registry_test.go

@@ -0,0 +1,92 @@
+package models
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestValidateModelString(t *testing.T) {
+	registry := GetGlobalRegistry()
+
+	tests := []struct {
+		name      string
+		model     string
+		wantErr   bool
+		errSubstr string // expected substring in error message (empty = don't check)
+	}{
+		{
+			name:    "valid anthropic model",
+			model:   "anthropic/claude-sonnet-4-6",
+			wantErr: false,
+		},
+		{
+			name:      "missing provider prefix",
+			model:     "claude-sonnet-4-6",
+			wantErr:   true,
+			errSubstr: "invalid model format",
+		},
+		{
+			name:      "empty string",
+			model:     "",
+			wantErr:   true,
+			errSubstr: "invalid model format",
+		},
+		{
+			name:      "unknown provider",
+			model:     "fakeprovider/some-model",
+			wantErr:   true,
+			errSubstr: "unknown provider",
+		},
+		{
+			name:    "ollama always valid",
+			model:   "ollama/llama3",
+			wantErr: false,
+		},
+		{
+			name:    "custom always valid",
+			model:   "custom/my-fine-tune",
+			wantErr: false,
+		},
+		{
+			name:      "empty provider",
+			model:     "/claude-sonnet-4-6",
+			wantErr:   true,
+			errSubstr: "invalid model format",
+		},
+		{
+			name:      "empty model name",
+			model:     "anthropic/",
+			wantErr:   true,
+			errSubstr: "invalid model format",
+		},
+		{
+			name:    "unknown model under known provider (no suggestions)",
+			model:   "anthropic/totally-unknown-xyz-999",
+			wantErr: false, // no suggestions → passes through
+		},
+		{
+			name:      "typo model under known provider with suggestions",
+			model:     "anthropic/claude-sonet", // misspelled "sonnet"
+			wantErr:   true,
+			errSubstr: "Did you mean",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := registry.ValidateModelString(tt.model)
+			if tt.wantErr && err == nil {
+				t.Errorf("ValidateModelString(%q) = nil, want error", tt.model)
+			}
+			if !tt.wantErr && err != nil {
+				t.Errorf("ValidateModelString(%q) = %v, want nil", tt.model, err)
+			}
+			if tt.errSubstr != "" && err != nil {
+				if !strings.Contains(err.Error(), tt.errSubstr) {
+					t.Errorf("ValidateModelString(%q) error = %q, want substring %q",
+						tt.model, err.Error(), tt.errSubstr)
+				}
+			}
+		})
+	}
+}

internal/models/rightsize_test.go

@@ -0,0 +1,148 @@
+package models
+
+import (
+	"testing"
+
+	"github.com/spf13/pflag"
+	"github.com/spf13/viper"
+)
+
+// bindMaxTokensFlag wires a fresh pflag-backed "max-tokens" key into viper so
+// isExplicitlySet behaves the same way it does in production. Returns a
+// cleanup function that removes the binding so sibling tests see a clean
+// state.
+func bindMaxTokensFlag(t *testing.T, args []string) func() {
+	t.Helper()
+	fs := pflag.NewFlagSet("test", pflag.ContinueOnError)
+	fs.Int("max-tokens", 8192, "")
+	if err := viper.BindPFlag("max-tokens", fs.Lookup("max-tokens")); err != nil {
+		t.Fatalf("BindPFlag: %v", err)
+	}
+	if err := fs.Parse(args); err != nil {
+		t.Fatalf("fs.Parse: %v", err)
+	}
+	return func() {
+		viper.Reset()
+	}
+}
+
+func TestRightSizeMaxTokens_RaisesWhenBelowCeiling(t *testing.T) {
+	cleanup := bindMaxTokensFlag(t, nil) // no args → flag.Changed = false
+	defer cleanup()
+
+	config := &ProviderConfig{MaxTokens: 8192}
+	modelInfo := &ModelInfo{
+		ID:    "claude-sonnet-4-5",
+		Limit: Limit{Context: 200000, Output: 64000},
+	}
+
+	rightSizeMaxTokens(config, modelInfo)
+
+	if config.MaxTokens != 32768 {
+		t.Errorf("expected MaxTokens raised to defaultRightSizeCap (32768), got %d", config.MaxTokens)
+	}
+}
+
+func TestRightSizeMaxTokens_CapsAtDefaultRightSizeCap(t *testing.T) {
+	cleanup := bindMaxTokensFlag(t, nil)
+	defer cleanup()
+
+	config := &ProviderConfig{MaxTokens: 8192}
+	// Mistral Devstral has 262144 output — we should still cap at 32768.
+	modelInfo := &ModelInfo{
+		ID:    "devstral-medium-latest",
+		Limit: Limit{Context: 262144, Output: 262144},
+	}
+
+	rightSizeMaxTokens(config, modelInfo)
+
+	if config.MaxTokens != defaultRightSizeCap {
+		t.Errorf("expected MaxTokens capped at %d, got %d", defaultRightSizeCap, config.MaxTokens)
+	}
+}
+
+func TestRightSizeMaxTokens_UsesExactOutputWhenBelowCap(t *testing.T) {
+	cleanup := bindMaxTokensFlag(t, nil)
+	defer cleanup()
+
+	config := &ProviderConfig{MaxTokens: 4096}
+	// Model with output limit smaller than the cap.
+	modelInfo := &ModelInfo{
+		ID:    "gpt-4",
+		Limit: Limit{Context: 8192, Output: 8192},
+	}
+
+	rightSizeMaxTokens(config, modelInfo)
+
+	if config.MaxTokens != 8192 {
+		t.Errorf("expected MaxTokens raised to model output ceiling (8192), got %d", config.MaxTokens)
+	}
+}
+
+func TestRightSizeMaxTokens_DoesNotLowerCurrentValue(t *testing.T) {
+	cleanup := bindMaxTokensFlag(t, nil)
+	defer cleanup()
+
+	// User (via per-model settings, applied earlier) already bumped MaxTokens
+	// above the cap — we must not clobber their choice.
+	config := &ProviderConfig{MaxTokens: 100000}
+	modelInfo := &ModelInfo{
+		ID:    "devstral-medium-latest",
+		Limit: Limit{Context: 262144, Output: 262144},
+	}
+
+	rightSizeMaxTokens(config, modelInfo)
+
+	if config.MaxTokens != 100000 {
+		t.Errorf("expected MaxTokens preserved at 100000, got %d", config.MaxTokens)
+	}
+}
+
+func TestRightSizeMaxTokens_RespectsExplicitFlag(t *testing.T) {
+	// Simulate `--max-tokens 4096` on the command line.
+	cleanup := bindMaxTokensFlag(t, []string{"--max-tokens", "4096"})
+	defer cleanup()
+
+	config := &ProviderConfig{MaxTokens: 4096}
+	modelInfo := &ModelInfo{
+		ID:    "claude-sonnet-4-5",
+		Limit: Limit{Context: 200000, Output: 64000},
+	}
+
+	rightSizeMaxTokens(config, modelInfo)
+
+	if config.MaxTokens != 4096 {
+		t.Errorf("expected explicit --max-tokens to be preserved (4096), got %d", config.MaxTokens)
+	}
+}
+
+func TestRightSizeMaxTokens_NilModelInfo(t *testing.T) {
+	cleanup := bindMaxTokensFlag(t, nil)
+	defer cleanup()
+
+	config := &ProviderConfig{MaxTokens: 8192}
+	// Custom model / Ollama / unknown provider → no model info.
+	rightSizeMaxTokens(config, nil)
+
+	if config.MaxTokens != 8192 {
+		t.Errorf("expected MaxTokens unchanged with nil modelInfo, got %d", config.MaxTokens)
+	}
+}
+
+func TestRightSizeMaxTokens_ZeroOutputLimit(t *testing.T) {
+	cleanup := bindMaxTokensFlag(t, nil)
+	defer cleanup()
+
+	config := &ProviderConfig{MaxTokens: 8192}
+	// Model present in catalog but with no known output limit.
+	modelInfo := &ModelInfo{
+		ID:    "unknown-model",
+		Limit: Limit{Context: 0, Output: 0},
+	}
+
+	rightSizeMaxTokens(config, modelInfo)
+
+	if config.MaxTokens != 8192 {
+		t.Errorf("expected MaxTokens unchanged with zero output limit, got %d", config.MaxTokens)
+	}
+}

internal/prompts/loader.go

@@ -2,11 +2,10 @@ package prompts
 
 import (
 	"fmt"
+	"log"
 	"os"
 	"path/filepath"
 	"strings"
-
-	"github.com/charmbracelet/log"
 )
 
 // LoadOptions configures how templates are discovered and loaded.
@@ -37,15 +36,17 @@ type Diagnostic struct {
 }
 
 // LoadAll discovers and loads all prompt templates from standard locations
-// and any extra paths. Templates are loaded in order of precedence (lowest
-// to highest), with later templates overriding earlier ones of the same name.
+// and any extra paths. Templates are loaded in order of precedence (highest
+// to lowest); the first source to define a given name wins, later definitions
+// of the same name are dropped with a diagnostic.
 //
 // Discovery paths searched in order:
 //  1. Default templates (if IncludeDefaults)
-//  2. ~/.kit/prompts/ (global user templates)
-//  3. .kit/prompts/ (project-local templates)
-//  4. ConfigPaths (from configuration)
-//  5. ExtraPaths (explicit paths, highest precedence)
+//  2. ~/.kit/prompts/                          (legacy global)
+//  3. $XDG_CONFIG_HOME/kit/prompts/            (XDG global, default ~/.config/kit/prompts/)
+//  4. <cwd>/.kit/prompts/                      (project-local templates)
+//  5. ConfigPaths                              (from configuration)
+//  6. ExtraPaths                               (explicit paths, lowest precedence)
 func LoadAll(opts LoadOptions) ([]*PromptTemplate, []Diagnostic, error) {
 	if opts.Cwd == "" {
 		opts.Cwd, _ = os.Getwd()
@@ -74,10 +75,7 @@ func LoadAll(opts LoadOptions) ([]*PromptTemplate, []Diagnostic, error) {
 					DroppedPath: tpl.FilePath,
 					Reason:      fmt.Sprintf("template from %s overridden by %s", source, existing.Source),
 				})
-				log.Debug("template collision",
-					"name", tpl.Name,
-					"dropped", tpl.FilePath,
-					"kept", existing.FilePath)
+				log.Printf("DEBUG template collision: name=%s dropped=%s kept=%s", tpl.Name, tpl.FilePath, existing.FilePath)
 			} else {
 				tpl.Source = source
 				seen[tpl.Name] = tpl
@@ -92,13 +90,21 @@ func LoadAll(opts LoadOptions) ([]*PromptTemplate, []Diagnostic, error) {
 		addTemplates(defaults, "default")
 	}
 
-	// 2. Global user templates: ~/.kit/prompts/
-	globalDir := filepath.Join(opts.HomeDir, ".kit", "prompts")
-	if templates, err := LoadFromDir(globalDir); err == nil {
+	// 2. Legacy global user templates: ~/.kit/prompts/
+	legacyGlobalDir := filepath.Join(opts.HomeDir, ".kit", "prompts")
+	if templates, err := LoadFromDir(legacyGlobalDir); err == nil {
 		addTemplates(templates, "global")
 	}
 
-	// 3. Project-local templates: .kit/prompts/
+	// 3. XDG global user templates: $XDG_CONFIG_HOME/kit/prompts/
+	//    Default: ~/.config/kit/prompts/. Aligns with extensions and skills.
+	if xdgDir := GlobalDir(); xdgDir != "" && xdgDir != legacyGlobalDir {
+		if templates, err := LoadFromDir(xdgDir); err == nil {
+			addTemplates(templates, "global")
+		}
+	}
+
+	// 4. Project-local templates: .kit/prompts/
 	localDir := filepath.Join(opts.Cwd, ".kit", "prompts")
 	if templates, err := LoadFromDir(localDir); err == nil {
 		addTemplates(templates, "local")
@@ -183,31 +189,6 @@ func LoadFromDir(dir string) ([]*PromptTemplate, error) {
 	return templates, nil
 }
 
-// Deduplicate removes duplicate templates by name, keeping the first occurrence.
-// It returns the deduplicated list and diagnostics for any collisions.
-// This is a standalone function for when you need to deduplicate an existing list.
-func Deduplicate(templates []*PromptTemplate) ([]*PromptTemplate, []Diagnostic) {
-	seen := make(map[string]*PromptTemplate)
-	var result []*PromptTemplate
-	var diagnostics []Diagnostic
-
-	for _, tpl := range templates {
-		if existing, ok := seen[tpl.Name]; ok {
-			diagnostics = append(diagnostics, Diagnostic{
-				Name:        tpl.Name,
-				KeptPath:    existing.FilePath,
-				DroppedPath: tpl.FilePath,
-				Reason:      "duplicate template name (first-match-wins)",
-			})
-		} else {
-			seen[tpl.Name] = tpl
-			result = append(result, tpl)
-		}
-	}
-
-	return result, diagnostics
-}
-
 // loadDefaultTemplates returns the built-in default templates.
 // These are embedded templates that ship with Kit.
 func loadDefaultTemplates() []*PromptTemplate {
@@ -215,3 +196,22 @@ func loadDefaultTemplates() []*PromptTemplate {
 	// For now, return an empty slice - users can define their own templates
 	return nil
 }
+
+// GlobalDir returns the XDG-aligned global prompts directory, respecting
+// $XDG_CONFIG_HOME. Defaults to ~/.config/kit/prompts/. Returns an empty
+// string if the user's home directory cannot be resolved.
+//
+// This is the canonical location for user-wide prompt templates and aligns
+// with the discovery paths used for extensions ($XDG_CONFIG_HOME/kit/extensions/)
+// and skills ($XDG_CONFIG_HOME/kit/skills/).
+func GlobalDir() string {
+	base := os.Getenv("XDG_CONFIG_HOME")
+	if base == "" {
+		home, err := os.UserHomeDir()
+		if err != nil {
+			return ""
+		}
+		base = filepath.Join(home, ".config")
+	}
+	return filepath.Join(base, "kit", "prompts")
+}

internal/prompts/template.go

@@ -7,10 +7,12 @@ import (
 	"regexp"
 	"strconv"
 	"strings"
+
+	"github.com/mark3labs/kit/internal/fences"
 )
 
 // PromptTemplate is a named prompt template with shell-style argument placeholders.
-// It supports Pi-style $1, $2, $@, $ARGUMENTS, ${@:N}, ${@:N:L} syntax.
+// It supports Pi-style $1, $2, $@, $+, $ARGUMENTS, ${@:N}, ${@:N:L} syntax.
 type PromptTemplate struct {
 	// Name is the human-readable identifier for this template.
 	Name string
@@ -120,19 +122,28 @@ func ParseCommandArgs(input string) []string {
 
 // argPlaceholder matches shell-style argument placeholders:
 //   - $1, $2, etc. - positional arguments
-//   - $@ - all arguments
+//   - $@ - all arguments (zero or more)
+//   - $+ - all arguments (one or more required)
 //   - $ARGUMENTS - all arguments (alias for $@)
 //   - ${@:N} - arguments from N onwards
 //   - ${@:N:L} - L arguments starting from N
-var argPlaceholder = regexp.MustCompile(`\$\{(\d+)\}|\$\{(\d+):(\d+)\}|\$\{ARGUMENTS\}|\$\{@(:\d+)?(:\d+)?\}|\$(\d+)|\$@|\$ARGUMENTS`)
+var argPlaceholder = regexp.MustCompile(`\$\{(\d+)\}|\$\{(\d+):(\d+)\}|\$\{ARGUMENTS\}|\$\{@(:\d+)?(:\d+)?\}|\$(\d+)|\$@|\$\+|\$ARGUMENTS`)
 
 // SubstituteArgs replaces argument placeholders in content with values from args.
 // Supported placeholders:
 //   - $N, ${N} - the Nth argument (1-indexed)
-//   - $@, $ARGUMENTS, ${ARGUMENTS} - all arguments joined with spaces
+//   - $@, $+, $ARGUMENTS, ${ARGUMENTS} - all arguments joined with spaces
 //   - ${@:N} - arguments from index N onwards (0-indexed)
 //   - ${@:N:L} - L arguments starting from index N (0-indexed)
 func SubstituteArgs(content string, args []string) string {
+	return fences.ReplaceOutside(content, func(segment string) string {
+		return substituteArgsInSegment(segment, args)
+	})
+}
+
+// substituteArgsInSegment performs argument substitution on a single text
+// segment that is known to be outside fenced code blocks.
+func substituteArgsInSegment(content string, args []string) string {
 	return argPlaceholder.ReplaceAllStringFunc(content, func(match string) string {
 		// Check for ${N} or ${N:M} format
 		if strings.HasPrefix(match, "${") && strings.Contains(match, "}") {
@@ -191,8 +202,8 @@ func SubstituteArgs(content string, args []string) string {
 		if strings.HasPrefix(match, "$") && !strings.HasPrefix(match, "${") {
 			suffix := match[1:]
 
-			// $@ or $ARGUMENTS
-			if suffix == "@" || suffix == "ARGUMENTS" {
+			// $@, $+, or $ARGUMENTS
+			if suffix == "@" || suffix == "+" || suffix == "ARGUMENTS" {
 				return strings.Join(args, " ")
 			}
 
@@ -266,6 +277,48 @@ func joinArgsRange(args []string, start, length int) string {
 	return strings.Join(args[start:end], " ")
 }
 
+// HasArgPlaceholders reports whether the template content contains any
+// argument placeholders ($1, $@, $ARGUMENTS, ${@:...}, etc.).
+// Placeholders inside fenced code blocks and inline code spans are ignored.
+func (t *PromptTemplate) HasArgPlaceholders() bool {
+	return argPlaceholder.MatchString(fences.StripCode(t.Content))
+}
+
+// RequiredArgs returns the number of positional arguments the template
+// expects. This is determined by the highest $N or ${N} placeholder found
+// in the content (1-indexed, so $2 means 2 args required). The $+
+// placeholder (required variadic) ensures at least 1. Optional wildcards
+// ($@, $ARGUMENTS) do not contribute to the count.
+func (t *PromptTemplate) RequiredArgs() int {
+	content := fences.StripCode(t.Content)
+	maxN := 0
+	hasRequiredVariadic := strings.Contains(content, "$+")
+	for _, match := range argPlaceholder.FindAllStringSubmatch(content, -1) {
+		// Group 1: ${N} format — the N value.
+		if match[1] != "" {
+			if n, err := strconv.Atoi(match[1]); err == nil && n > maxN {
+				maxN = n
+			}
+		}
+		// Group 2: ${N:M} format — the N value (start index).
+		if match[2] != "" {
+			if n, err := strconv.Atoi(match[2]); err == nil && n > maxN {
+				maxN = n
+			}
+		}
+		// Group 6: $N format (no braces) — the N value.
+		if match[6] != "" {
+			if n, err := strconv.Atoi(match[6]); err == nil && n > maxN {
+				maxN = n
+			}
+		}
+	}
+	if hasRequiredVariadic && maxN < 1 {
+		maxN = 1
+	}
+	return maxN
+}
+
 // Expand substitutes arguments into the template content and returns the result.
 // It first parses args from the input string, then substitutes them into the template.
 func (t *PromptTemplate) Expand(argsInput string) string {

internal/prompts/template_test.go

@@ -129,6 +129,48 @@ func TestSubstituteArgs(t *testing.T) {
 			args:     []string{},
 			expected: "Args: ",
 		},
+		{
+			name:     "$1 inside code block preserved",
+			content:  "Use $1 here\n```bash\necho $1\n```\ndone",
+			args:     []string{"foo"},
+			expected: "Use foo here\n```bash\necho $1\n```\ndone",
+		},
+		{
+			name:     "$@ inside code block preserved",
+			content:  "Run $@\n```\necho $@\n```\n",
+			args:     []string{"a", "b"},
+			expected: "Run a b\n```\necho $@\n```\n",
+		},
+		{
+			name:     "all placeholders inside code block",
+			content:  "Prompt\n```\n$1 $2 $@\n```\n",
+			args:     []string{"x"},
+			expected: "Prompt\n```\n$1 $2 $@\n```\n",
+		},
+		{
+			name:     "$1 inside inline code preserved",
+			content:  "Use `$1` here and $1 outside",
+			args:     []string{"foo"},
+			expected: "Use `$1` here and foo outside",
+		},
+		{
+			name:     "$+ required variadic",
+			content:  "Args: $+",
+			args:     []string{"a", "b", "c"},
+			expected: "Args: a b c",
+		},
+		{
+			name:     "$+ with empty args",
+			content:  "Args: $+",
+			args:     []string{},
+			expected: "Args: ",
+		},
+		{
+			name:     "all placeholders in inline code",
+			content:  "Use `$1` and `$@` for args",
+			args:     []string{"x"},
+			expected: "Use `$1` and `$@` for args",
+		},
 	}
 
 	for _, tt := range tests {
@@ -213,3 +255,78 @@ func TestPromptTemplateExpand(t *testing.T) {
 		})
 	}
 }
+
+func TestHasArgPlaceholders(t *testing.T) {
+	tests := []struct {
+		name    string
+		content string
+		want    bool
+	}{
+		{"no placeholders", "Just a plain prompt with no args", false},
+		{"$1 placeholder", "Create a $1 component", true},
+		{"$@ placeholder", "Run with args: $@", true},
+		{"$ARGUMENTS placeholder", "Features: $ARGUMENTS", true},
+		{"${1} placeholder", "Name: ${1}", true},
+		{"${ARGUMENTS} placeholder", "All: ${ARGUMENTS}", true},
+		{"${@:1} placeholder", "Rest: ${@:1}", true},
+		{"${@:1:2} placeholder", "Slice: ${@:1:2}", true},
+		{"dollar in text", "Cost is one hundred dollars", false},
+		{"empty content", "", false},
+		{"$1 inside code block only", "Prompt\n```\necho $1\n```\n", false},
+		{"$1 outside and inside code block", "Use $1 here\n```\necho $1\n```\n", true},
+		{"$@ inside code block only", "Prompt\n```bash\necho $@\n```\n", false},
+		{"$+ placeholder", "Run with args: $+", true},
+		{"$+ inside inline code only", "Use `$+` for required args", false},
+		{"$1 inside inline code only", "Use `$1` for positional args", false},
+		{"$1 outside and in inline code", "Create $1 (see `$1` syntax)", true},
+		{"$@ outside $1 in inline code", "Run $@ with `$1` syntax", true},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			tpl := &PromptTemplate{Content: tt.content}
+			if got := tpl.HasArgPlaceholders(); got != tt.want {
+				t.Errorf("HasArgPlaceholders() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestRequiredArgs(t *testing.T) {
+	tests := []struct {
+		name    string
+		content string
+		want    int
+	}{
+		{"no placeholders", "Just a plain prompt", 0},
+		{"$1 only", "Create a $1 component", 1},
+		{"$1 and $2", "Create $1 with $2", 2},
+		{"$3 skipping $2", "Use $1 and $3", 3},
+		{"${1} braced", "Name: ${1}", 1},
+		{"${2} braced", "Name: ${1} Desc: ${2}", 2},
+		{"$@ only", "Run with: $@", 0},
+		{"$ARGUMENTS only", "Features: $ARGUMENTS", 0},
+		{"${ARGUMENTS} only", "All: ${ARGUMENTS}", 0},
+		{"$1 and $@", "Create $1 with extras: $@", 1},
+		{"${@:1} slice only", "Rest: ${@:1}", 0},
+		{"${@:1:2} slice only", "Slice: ${@:1:2}", 0},
+		{"mixed $1 $2 and $@", "Create $1 named $2: $@", 2},
+		{"empty content", "", 0},
+		{"$2 inside code block only", "Prompt\n```\n$1 $2\n```\n", 0},
+		{"$1 outside $2 inside code block", "Use $1\n```\n$2 inside\n```\n", 1},
+		{"$+ only", "Run with: $+", 1},
+		{"$+ and $2", "Create $2 with: $+", 2},
+		{"$+ inside inline code only", "Use `$+` for required args", 0},
+		{"$1 and $2 in inline code only", "Use `$1` and `$2` for args", 0},
+		{"$1 outside $2 in inline code", "Create $1 (see `$2`)", 1},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			tpl := &PromptTemplate{Content: tt.content}
+			if got := tpl.RequiredArgs(); got != tt.want {
+				t.Errorf("RequiredArgs() = %d, want %d", got, tt.want)
+			}
+		})
+	}
+}

internal/session/compaction_cycle_test.go

@@ -0,0 +1,66 @@
+package session
+
+import (
+	"testing"
+
+	"github.com/mark3labs/kit/internal/message"
+)
+
+// TestCompactionParentCycleRegression tests that after multiple compactions,
+// newly appended messages always have a valid parent chain and BuildContext
+// returns the correct messages.
+func TestCompactionParentCycleRegression(t *testing.T) {
+	tm := InMemoryTreeSession("/test")
+
+	// Simulate a long conversation with multiple compactions.
+	msg1, _ := tm.AppendMessage(message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "msg1"}}})
+	msg2, _ := tm.AppendMessage(message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "msg2"}}})
+
+	// First compaction
+	comp1, _ := tm.AppendCompaction("Summary 1", msg1, 1000, 500, 1, []string{}, []string{})
+
+	msg3, _ := tm.AppendMessage(message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "msg3"}}})
+	msg4, _ := tm.AppendMessage(message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "msg4"}}})
+
+	// Second compaction
+	comp2, _ := tm.AppendCompaction("Summary 2", msg3, 1000, 500, 1, []string{}, []string{})
+
+	msg5, _ := tm.AppendMessage(message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "msg5"}}})
+	msg6, _ := tm.AppendMessage(message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "msg6"}}})
+
+	// Verify parent chain integrity
+	for _, id := range []string{msg1, msg2, comp1, msg3, msg4, comp2, msg5, msg6} {
+		entry := tm.GetEntry(id)
+		if entry == nil {
+			t.Fatalf("entry %s not found in index", id)
+		}
+	}
+
+	// Walk parent chain from msg6 — must reach root without cycles
+	visited := make(map[string]bool)
+	current := msg6
+	for current != "" {
+		if visited[current] {
+			t.Fatalf("cycle detected at entry %s", current)
+		}
+		visited[current] = true
+		entry := tm.GetEntry(current)
+		if entry == nil {
+			t.Fatalf("entry %s missing from index during parent walk", current)
+		}
+		parent := ""
+		switch e := entry.(type) {
+		case *MessageEntry:
+			parent = e.ParentID
+		case *CompactionEntry:
+			parent = e.ParentID
+		}
+		current = parent
+	}
+
+	// BuildContext should return: Summary2 + msg6 + msg5 + msg3 + msg4 = 5 messages
+	msgs, _, _ := tm.BuildContext()
+	if len(msgs) != 5 {
+		t.Fatalf("expected 5 messages, got %d: %+v", len(msgs), msgs)
+	}
+}

internal/session/compaction_test.go

@@ -0,0 +1,326 @@
+package session
+
+import (
+	"slices"
+	"testing"
+
+	"charm.land/fantasy"
+	"github.com/mark3labs/kit/internal/message"
+)
+
+// TestCompactionCreatesNewLeaf verifies that after compaction, the compaction
+// entry has no parent (creating a new root), and BuildContext returns only
+// the summary and kept messages, not the old compacted messages.
+func TestCompactionCreatesNewLeaf(t *testing.T) {
+	tm := InMemoryTreeSession("/test")
+
+	// Add some messages: M1, M2 (old, will be compacted), M3, M4 (kept)
+	msg1 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Message 1 - old"}}}
+	msg2 := message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "Message 2 - old"}}}
+	msg3 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Message 3 - kept"}}}
+	msg4 := message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "Message 4 - kept"}}}
+
+	_, _ = tm.AppendMessage(msg1)
+	_, _ = tm.AppendMessage(msg2)
+	id3, _ := tm.AppendMessage(msg3)
+	id4, _ := tm.AppendMessage(msg4)
+
+	// Verify initial state - all messages should be in context
+	messages, _, _ := tm.BuildContext()
+	if len(messages) != 4 {
+		t.Fatalf("expected 4 messages before compaction, got %d", len(messages))
+	}
+
+	// Verify entry IDs
+	entryIDs := tm.GetContextEntryIDs()
+	if len(entryIDs) != 4 {
+		t.Fatalf("expected 4 entry IDs before compaction, got %d", len(entryIDs))
+	}
+
+	// Now add a compaction entry, simulating that M3 is the first kept entry
+	summary := "Summary of old messages"
+	compactionID, err := tm.AppendCompaction(summary, id3, 1000, 500, 2, []string{}, []string{})
+	if err != nil {
+		t.Fatalf("failed to append compaction: %v", err)
+	}
+
+	// Verify the compaction entry has no parent (empty ParentID)
+	compactionEntry := tm.GetEntry(compactionID).(*CompactionEntry)
+	if compactionEntry.ParentID != "" {
+		t.Errorf("compaction entry should have no parent, got %q", compactionEntry.ParentID)
+	}
+
+	// Verify the leaf is now the compaction entry
+	if tm.GetLeafID() != compactionID {
+		t.Errorf("leaf should be compaction entry %q, got %q", compactionID, tm.GetLeafID())
+	}
+
+	// Now BuildContext should return: [summary] + [M3, M4]
+	messages, _, _ = tm.BuildContext()
+	if len(messages) != 3 {
+		t.Fatalf("expected 3 messages after compaction (summary + 2 kept), got %d", len(messages))
+	}
+
+	// First message should be the summary
+	if messages[0].Role != fantasy.MessageRoleSystem {
+		t.Errorf("first message should be system summary, got %s", messages[0].Role)
+	}
+	summaryText := messages[0].Content[0].(fantasy.TextPart).Text
+	if summaryText != "[Conversation summary — earlier messages were compacted]\n\n"+summary {
+		t.Errorf("unexpected summary text: %s", summaryText)
+	}
+
+	// Second message should be M3 (kept)
+	if messages[1].Role != fantasy.MessageRoleUser {
+		t.Errorf("second message should be user (M3), got %s", messages[1].Role)
+	}
+	m3Text := messages[1].Content[0].(fantasy.TextPart).Text
+	if m3Text != "Message 3 - kept" {
+		t.Errorf("unexpected M3 text: %s", m3Text)
+	}
+
+	// Third message should be M4 (kept)
+	if messages[2].Role != fantasy.MessageRoleAssistant {
+		t.Errorf("third message should be assistant (M4), got %s", messages[2].Role)
+	}
+	m4Text := messages[2].Content[0].(fantasy.TextPart).Text
+	if m4Text != "Message 4 - kept" {
+		t.Errorf("unexpected M4 text: %s", m4Text)
+	}
+
+	// Verify GetContextEntryIDs returns correct IDs
+	entryIDs = tm.GetContextEntryIDs()
+	if len(entryIDs) != 3 {
+		t.Fatalf("expected 3 entry IDs after compaction (empty for summary + 2 kept), got %d: %v", len(entryIDs), entryIDs)
+	}
+
+	// First entry ID should be empty (summary has no entry)
+	if entryIDs[0] != "" {
+		t.Errorf("first entry ID should be empty (summary), got %q", entryIDs[0])
+	}
+
+	// Second and third should be id3 and id4 (the kept messages)
+	if entryIDs[1] != id3 {
+		t.Errorf("second entry ID should be %q (M3), got %q", id3, entryIDs[1])
+	}
+	if entryIDs[2] != id4 {
+		t.Errorf("third entry ID should be %q (M4), got %q", id4, entryIDs[2])
+	}
+}
+
+// TestCompactionWithNewMessagesAfterCompaction verifies that messages appended
+// after compaction are correctly included in the context.
+func TestCompactionWithNewMessagesAfterCompaction(t *testing.T) {
+	tm := InMemoryTreeSession("/test")
+
+	// Add initial messages
+	msg1 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Message 1"}}}
+	msg2 := message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "Message 2"}}}
+	msg3 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Message 3 - kept"}}}
+
+	_, _ = tm.AppendMessage(msg1)
+	_, _ = tm.AppendMessage(msg2)
+	id3, _ := tm.AppendMessage(msg3)
+
+	// Compact, keeping only M3
+	_, _ = tm.AppendCompaction("Summary", id3, 1000, 500, 2, []string{}, []string{})
+
+	// Add a new message after compaction
+	msg4 := message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "Message 4 - after compaction"}}}
+	_, _ = tm.AppendMessage(msg4)
+
+	// BuildContext should return: [summary] + [M3 (kept)] + [M4 (new after compaction)]
+	// Kept messages must appear BEFORE post-compaction messages so the LLM
+	// sees the conversation in chronological order. Otherwise the latest
+	// post-compaction user message would be followed by an older kept user
+	// message, breaking user/assistant alternation and causing the model to
+	// respond as if the post-compaction turn never happened.
+	messages, _, _ := tm.BuildContext()
+	if len(messages) != 3 {
+		t.Fatalf("expected 3 messages (summary + M3 + M4), got %d: %+v", len(messages), messages)
+	}
+
+	// Verify order: summary, M3 (kept), M4 (new)
+	if messages[0].Role != fantasy.MessageRoleSystem {
+		t.Errorf("first message should be summary, got %s", messages[0].Role)
+	}
+	if messages[1].Role != fantasy.MessageRoleUser {
+		t.Errorf("second message should be user (M3 kept), got %s", messages[1].Role)
+	}
+	m3Text := messages[1].Content[0].(fantasy.TextPart).Text
+	if m3Text != "Message 3 - kept" {
+		t.Errorf("unexpected M3 text: %s", m3Text)
+	}
+	if messages[2].Role != fantasy.MessageRoleAssistant {
+		t.Errorf("third message should be assistant (M4 post-compact), got %s", messages[2].Role)
+	}
+	m4Text := messages[2].Content[0].(fantasy.TextPart).Text
+	if m4Text != "Message 4 - after compaction" {
+		t.Errorf("unexpected M4 text: %s", m4Text)
+	}
+
+	// Verify that M1 is NOT in the context
+	for i, msg := range messages {
+		if msg.Role == fantasy.MessageRoleUser {
+			text := msg.Content[0].(fantasy.TextPart).Text
+			if text == "Message 1" {
+				t.Errorf("Message 1 (compacted) should not be in context at index %d", i)
+			}
+		}
+	}
+}
+
+// TestCompactionWithNoKeptMessages verifies compaction when all messages are compacted.
+func TestCompactionWithNoKeptMessages(t *testing.T) {
+	tm := InMemoryTreeSession("/test")
+
+	// Add messages that will all be compacted
+	msg1 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Message 1"}}}
+	msg2 := message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "Message 2"}}}
+
+	if _, err := tm.AppendMessage(msg1); err != nil {
+		t.Fatalf("failed to append message: %v", err)
+	}
+	if _, err := tm.AppendMessage(msg2); err != nil {
+		t.Fatalf("failed to append message: %v", err)
+	}
+
+	// Compact with no kept messages (empty firstKeptEntryID)
+	summary := "All messages summarized"
+	compactionID, _ := tm.AppendCompaction(summary, "", 1000, 100, 2, []string{}, []string{})
+
+	// Verify the compaction entry has no parent
+	compactionEntry := tm.GetEntry(compactionID).(*CompactionEntry)
+	if compactionEntry.ParentID != "" {
+		t.Errorf("compaction entry should have no parent, got %q", compactionEntry.ParentID)
+	}
+
+	// BuildContext should return only the summary
+	messages, _, _ := tm.BuildContext()
+	if len(messages) != 1 {
+		t.Fatalf("expected 1 message (summary only), got %d: %+v", len(messages), messages)
+	}
+	if messages[0].Role != fantasy.MessageRoleSystem {
+		t.Errorf("message should be system summary, got %s", messages[0].Role)
+	}
+}
+
+// TestMultipleCompactions verifies that multiple compactions work correctly.
+func TestMultipleCompactions(t *testing.T) {
+	tm := InMemoryTreeSession("/test")
+
+	// First batch of messages
+	msg1 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Batch 1 - User"}}}
+	msg2 := message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "Batch 1 - Assistant"}}}
+	id1, _ := tm.AppendMessage(msg1)
+	id2, _ := tm.AppendMessage(msg2)
+
+	// First compaction
+	_, _ = tm.AppendCompaction("Summary 1", id1, 1000, 500, 1, []string{}, []string{})
+
+	// Second batch
+	msg3 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Batch 2 - User"}}}
+	msg4 := message.Message{Role: message.RoleAssistant, Parts: []message.ContentPart{message.TextContent{Text: "Batch 2 - Assistant"}}}
+	id3, _ := tm.AppendMessage(msg3)
+	id4, _ := tm.AppendMessage(msg4)
+
+	// Second compaction (compacting the first compaction + batch 2)
+	// Note: id3 is the first kept entry, so id3 and id4 should be preserved
+	compactionID2, _ := tm.AppendCompaction("Summary 2", id3, 1000, 500, 3, []string{}, []string{})
+
+	// Verify second compaction has no parent
+	compactionEntry2 := tm.GetEntry(compactionID2).(*CompactionEntry)
+	if compactionEntry2.ParentID != "" {
+		t.Errorf("second compaction entry should have no parent, got %q", compactionEntry2.ParentID)
+	}
+
+	// Add final message
+	msg5 := message.Message{Role: message.RoleUser, Parts: []message.ContentPart{message.TextContent{Text: "Final message"}}}
+	id5, _ := tm.AppendMessage(msg5)
+
+	// BuildContext should include:
+	// - Summary 2 (from second compaction)
+	// - msg5 (final message)
+	// - msg3, msg4 (kept from second compaction)
+	// But NOT Summary 1 or msg1, msg2 (they're before the first kept entry of compaction 2)
+	messages, _, _ := tm.BuildContext()
+
+	// Should have: Summary 2 + msg5 + msg3 + msg4 = 4 messages
+	if len(messages) != 4 {
+		t.Fatalf("expected 4 messages (Summary 2 + msg5 + msg3 + msg4), got %d: %+v", len(messages), messages)
+	}
+
+	// First should be Summary 2
+	if messages[0].Role != fantasy.MessageRoleSystem {
+		t.Errorf("first message should be system (Summary 2), got %s", messages[0].Role)
+	}
+	summaryText := messages[0].Content[0].(fantasy.TextPart).Text
+	if summaryText != "[Conversation summary — earlier messages were compacted]\n\nSummary 2" {
+		t.Errorf("unexpected summary: %s", summaryText)
+	}
+
+	// Verify msg5 is included
+	foundFinal := false
+	for _, msg := range messages {
+		if msg.Role == fantasy.MessageRoleUser {
+			text := msg.Content[0].(fantasy.TextPart).Text
+			if text == "Final message" {
+				foundFinal = true
+				break
+			}
+		}
+	}
+	if !foundFinal {
+		t.Error("Final message (msg5) should be in context")
+	}
+
+	// Verify msg1, msg2 are NOT included (compacted by first compaction, then second)
+	for _, msg := range messages {
+		if msg.Role == fantasy.MessageRoleUser || msg.Role == fantasy.MessageRoleAssistant {
+			text := msg.Content[0].(fantasy.TextPart).Text
+			if text == "Batch 1 - User" || text == "Batch 1 - Assistant" {
+				t.Errorf("Batch 1 messages should not be in context, found: %s", text)
+			}
+		}
+	}
+
+	// Verify entry IDs
+	entryIDs := tm.GetContextEntryIDs()
+	if len(entryIDs) != 4 {
+		t.Fatalf("expected 4 entry IDs, got %d: %v", len(entryIDs), entryIDs)
+	}
+
+	// First should be empty (summary)
+	if entryIDs[0] != "" {
+		t.Errorf("first entry ID should be empty (summary), got %q", entryIDs[0])
+	}
+
+	// Check that id5 is in the list
+	if !slices.Contains(entryIDs, id5) {
+		t.Errorf("id5 (final message) should be in entry IDs, got %v", entryIDs)
+	}
+
+	// Verify id3 and id4 ARE in the list (they were kept)
+	foundID3, foundID4 := false, false
+	for _, id := range entryIDs {
+		if id == id3 {
+			foundID3 = true
+		}
+		if id == id4 {
+			foundID4 = true
+		}
+	}
+	if !foundID3 {
+		t.Errorf("id3 (kept message) should be in entry IDs, got %v", entryIDs)
+	}
+	if !foundID4 {
+		t.Errorf("id4 (kept message) should be in entry IDs, got %v", entryIDs)
+	}
+
+	// Verify id1 and id2 are NOT in the list (they were compacted away)
+	for _, id := range entryIDs {
+		if id == id1 || id == id2 {
+			t.Errorf("id1 or id2 (compacted) should not be in entry IDs, found %q in %v", id, entryIDs)
+		}
+	}
+}

internal/session/entry.go

@@ -24,6 +24,7 @@ const (
 	EntryTypeSessionInfo   EntryType = "session_info"
 	EntryTypeExtensionData EntryType = "extension_data"
 	EntryTypeCompaction    EntryType = "compaction"
+	EntryTypeSystemPrompt  EntryType = "system_prompt"
 )
 
 // CurrentVersion is the session format version for JSONL tree sessions.
@@ -117,6 +118,19 @@ type CompactionEntry struct {
 	ModifiedFiles    []string `json:"modified_files,omitempty"`
 }
 
+// SystemPromptEntry records the system prompt and model used for the session.
+// This is primarily for sharing/debugging to see what instructions were
+// active during the conversation. It does NOT participate in the tree
+// structure (no ParentID) and is not used when building LLM context.
+type SystemPromptEntry struct {
+	Type      EntryType `json:"type"`      // always "system_prompt"
+	ID        string    `json:"id"`        // unique entry ID
+	Timestamp time.Time `json:"timestamp"` // when captured
+	Content   string    `json:"content"`   // the system prompt text
+	Model     string    `json:"model"`     // the model used (e.g., "claude-sonnet-4-5")
+	Provider  string    `json:"provider"`  // the provider used (e.g., "anthropic")
+}
+
 // GenerateEntryID creates a unique entry identifier (16 hex chars).
 func GenerateEntryID() string {
 	bytes := make([]byte, 8)
@@ -217,6 +231,18 @@ func NewCompactionEntry(parentID, summary, firstKeptEntryID string, tokensBefore
 	}
 }
 
+// NewSystemPromptEntry creates a SystemPromptEntry.
+func NewSystemPromptEntry(content, model, provider string) *SystemPromptEntry {
+	return &SystemPromptEntry{
+		Type:      EntryTypeSystemPrompt,
+		ID:        GenerateEntryID(),
+		Timestamp: time.Now(),
+		Content:   content,
+		Model:     model,
+		Provider:  provider,
+	}
+}
+
 // --- JSONL marshaling helpers ---
 
 // MarshalEntry serializes any entry to a JSON line (no trailing newline).
@@ -295,6 +321,13 @@ func UnmarshalEntry(data []byte) (any, error) {
 		}
 		return &e, nil
 
+	case EntryTypeSystemPrompt:
+		var e SystemPromptEntry
+		if err := json.Unmarshal(data, &e); err != nil {
+			return nil, fmt.Errorf("failed to unmarshal system_prompt entry: %w", err)
+		}
+		return &e, nil
+
 	default:
 		return nil, fmt.Errorf("unknown entry type: %q", env.Type)
 	}