🐛 fix(server): rehydrate subagent runs from DB on cold replica (#15788)

* 🐛 fix(server): rehydrate subagent runs from DB on cold replica

Server-side hetero persistence kept per-operation state in a module-level
map. On a cold serverless replica (or any cross-replica batch), the main
agent state is rebuilt from DB but `MainAgentRunState.subagents` was seeded
empty. A continuing subagent event then hit the `!existing` branch of
`ensureRun` and forked a brand-new isolation thread for a parentToolCallId
that already had one — producing piles of generic "Subagent" threads that
were never attached to the right thread. Desktop never hit this (one
long-lived run-state closure).

Rebuild `state.main.subagents` from DB the same way the main half is
rehydrated: add `rehydrateSubagentRunsState` to @lobechat/heterogeneous-agents
and call a new `refreshSubagentRunsFromDb` each ingest. Only runs MISSING
from memory are rehydrated (warm accumulators win); finalized (Active)
threads are excluded so completed spawns are never resurrected.

Sibling of #15783 (main message chaining) — same root cause, subagent half.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* 🐛 fix(server): scope subagent rehydration to operation + de-dupe inner tools

Two follow-up fixes on the cold-replica subagent rehydration:

- P1: de-dupe inner tool creation against the run-lifetime tool set, not just
  the per-turn `persistedIds`. Per-turn state is reset on every turn boundary
  and starts empty after a rehydration, so a replayed / continued tools_calling
  on a cold replica minted a SECOND tool message for an id the run already
  wrote. `lifetimeToolCallIds` survives boundaries and is restored from DB, so
  it is the durable de-dupe key. Mirrors the main-agent retry protection.

- P2: scope `refreshSubagentRunsFromDb` to the current operation. Topics are
  reused across turns; a prior crashed/cancelled run can leave a subagent
  thread stuck `Processing`. Rehydrating purely by topic+status would merge
  that unrelated thread into the new operation's reducer state and finalize it
  on the new run's terminal drain. Stamp `operationId` on the subagent thread
  metadata at creation and filter rehydration by it.

Adds regression cases for both (each verified to fail without its fix).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

.agents/skills/add-provider-doc/SKILL.md

@@ -1,6 +1,8 @@
 ---
 name: add-provider-doc
-description: Guide for adding new AI provider documentation. Use when adding documentation for a new AI provider (like OpenAI, Anthropic, etc.), including usage docs, environment variables, Docker config, and image resources. Triggers on provider documentation tasks.
+description: Add documentation for a new AI provider — usage docs, env vars, Docker config, image resources.
+disable-model-invocation: true
+argument-hint: '[provider-name]'
 ---
 
 # Adding New AI Provider Documentation

.agents/skills/add-setting-env/SKILL.md

@@ -1,6 +1,8 @@
 ---
 name: add-setting-env
-description: Guide for adding environment variables to configure user settings. Use when implementing server-side environment variables that control default values for user settings. Triggers on env var configuration or setting default value tasks.
+description: Add server-side environment variables that control default values for user settings.
+disable-model-invocation: true
+argument-hint: '[setting-name]'
 ---
 
 # Adding Environment Variable for User Settings
@@ -49,7 +51,7 @@ export interface GlobalServerConfig {
 
 ### 3. Assemble Server Config (if new domain)
 
-In `src/server/globalConfig/index.ts`:
+In `apps/server/src/globalConfig/index.ts`:
 
 ```typescript
 import { <domain>Env } from '@/envs/<domain>';
@@ -95,7 +97,7 @@ AI_IMAGE_DEFAULT_IMAGE_NUM: z.coerce.number().min(1).max(20).optional(),
 // packages/types/src/serverConfig.ts
 image?: PartialDeep<UserImageConfig>;
 
-// src/server/globalConfig/index.ts
+// apps/server/src/globalConfig/index.ts
 image: cleanObject({ defaultImageNum: imageEnv.AI_IMAGE_DEFAULT_IMAGE_NUM }),
 
 // src/store/user/slices/common/action.ts

.agents/skills/agent-runtime-hooks/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: agent-runtime-hooks
-description: "Agent runtime lifecycle hooks for observing and intercepting agent execution. Use when adding hooks to agent operations, mocking tool calls, logging step events, handling human intervention, sub-agent calls, context compression, or building eval/tracing integrations. Triggers on 'hooks', 'beforeToolCall', 'afterToolCall', 'beforeStep', 'afterStep', 'onComplete', 'onError', 'tool mock', 'agent lifecycle', 'human intervention', 'callAgent', 'compact'."
+description: 'Agent runtime lifecycle hooks. Use for before/after tool or step hooks, tool mocks, human intervention, sub-agent calls, context compression, evals, tracing, callAgent, or lifecycle events.'
 user-invocable: false
 ---
 
@@ -50,14 +50,14 @@ execAgent({ hooks })
 
 ## Key Files
 
-| File                                                       | Role                                                   |
-| ---------------------------------------------------------- | ------------------------------------------------------ |
-| `packages/agent-runtime/src/types/hooks.ts`                | Type definitions (AgentHookType, all event interfaces) |
-| `src/server/services/agentRuntime/hooks/types.ts`          | Server-side types (AgentHook, re-exports)              |
-| `src/server/services/agentRuntime/hooks/HookDispatcher.ts` | Registration, dispatch, dispatchBeforeToolCall         |
-| `src/server/modules/AgentRuntime/RuntimeExecutors.ts`      | Tool/Compact/HumanIntervention hook dispatch           |
-| `src/server/services/agentRuntime/AgentRuntimeService.ts`  | Step hooks + HumanIntervention resume/reject           |
-| `src/server/services/aiAgent/index.ts`                     | CallAgent hook dispatch                                |
+| File                                                            | Role                                                   |
+| --------------------------------------------------------------- | ------------------------------------------------------ |
+| `packages/agent-runtime/src/types/hooks.ts`                     | Type definitions (AgentHookType, all event interfaces) |
+| `apps/server/src/services/agentRuntime/hooks/types.ts`          | Server-side types (AgentHook, re-exports)              |
+| `apps/server/src/services/agentRuntime/hooks/HookDispatcher.ts` | Registration, dispatch, dispatchBeforeToolCall         |
+| `apps/server/src/modules/AgentRuntime/RuntimeExecutors.ts`      | Tool/Compact/HumanIntervention hook dispatch           |
+| `apps/server/src/services/agentRuntime/AgentRuntimeService.ts`  | Step hooks + HumanIntervention resume/reject           |
+| `apps/server/src/services/aiAgent/index.ts`                     | CallAgent hook dispatch                                |
 
 ## Registration Flow
 

.agents/skills/agent-signal/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: agent-signal
-description: Build or extend LobeHub Agent Signal pipelines for background or quiet agent work driven by event sources, semantic signals, and action handlers. Use when adding a new Agent Signal source, signal or action type, policy, middleware handler, workflow handoff, dedupe or scope behavior, or observability around `src/server/services/agentSignal/**`, `packages/agent-signal`, or `packages/observability-otel/src/modules/agent-signal`.
+description: 'Build or extend LobeHub Agent Signal pipelines. Use for signal sources, signal/action types, policies, middleware, workflow handoff, dedupe, scope behavior, or observability.'
 ---
 
 # Agent Signal
@@ -26,9 +26,9 @@ Agent Signal has one consistent shape:
 
 Read:
 
-- `src/server/services/agentSignal/index.ts`
-- `src/server/workflows/agentSignal/index.ts`
-- `src/server/workflows/agentSignal/run.ts`
+- `apps/server/src/services/agentSignal/index.ts`
+- `apps/server/src/workflows/agentSignal/index.ts`
+- `apps/server/src/workflows/agentSignal/run.ts`
 
 ## Core Model
 
@@ -48,11 +48,11 @@ Keep the boundaries strict:
 ## Implementation Workflow
 
 1. Decide whether the use case is synchronous or quiet background work.
-2. Define or reuse a source type in `src/server/services/agentSignal/sourceTypes.ts`.
-3. Define or reuse signal and action types in `src/server/services/agentSignal/policies/types.ts`.
+2. Define or reuse a source type in `apps/server/src/services/agentSignal/sourceTypes.ts`.
+3. Define or reuse signal and action types in `apps/server/src/services/agentSignal/policies/types.ts`.
 4. Implement handlers with `defineSourceHandler`, `defineSignalHandler`, or `defineActionHandler`.
 5. Bundle handlers with `defineAgentSignalHandlers(...)`.
-6. Register the policy in `src/server/services/agentSignal/policies/index.ts` and pass it into the runtime factory if needed.
+6. Register the policy in `apps/server/src/services/agentSignal/policies/index.ts` and pass it into the runtime factory if needed.
 7. Add or update ingress code that emits or enqueues the source event.
 8. Add observability and tests before considering the flow complete.
 
@@ -63,19 +63,19 @@ Keep the boundaries strict:
   `packages/agent-signal/src/base/builders.ts`
   `packages/agent-signal/src/base/types.ts`
 - Server-owned runtime and middleware:
-  `src/server/services/agentSignal/runtime/AgentSignalRuntime.ts`
-  `src/server/services/agentSignal/runtime/AgentSignalScheduler.ts`
-  `src/server/services/agentSignal/runtime/middleware.ts`
-  `src/server/services/agentSignal/runtime/context.ts`
+  `apps/server/src/services/agentSignal/runtime/AgentSignalRuntime.ts`
+  `apps/server/src/services/agentSignal/runtime/AgentSignalScheduler.ts`
+  `apps/server/src/services/agentSignal/runtime/middleware.ts`
+  `apps/server/src/services/agentSignal/runtime/context.ts`
 - Existing policy example:
-  `src/server/services/agentSignal/policies/analyzeIntent/index.ts`
-  `src/server/services/agentSignal/policies/analyzeIntent/feedbackSatisfaction.ts`
-  `src/server/services/agentSignal/policies/analyzeIntent/feedbackDomain.ts`
-  `src/server/services/agentSignal/policies/analyzeIntent/feedbackAction.ts`
-  `src/server/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`
+  `apps/server/src/services/agentSignal/policies/analyzeIntent/index.ts`
+  `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackSatisfaction.ts`
+  `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackDomain.ts`
+  `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackAction.ts`
+  `apps/server/src/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`
 - Observability:
-  `src/server/services/agentSignal/observability/projector.ts`
-  `src/server/services/agentSignal/observability/traceEvents.ts`
+  `apps/server/src/services/agentSignal/observability/projector.ts`
+  `apps/server/src/services/agentSignal/observability/traceEvents.ts`
   `packages/observability-otel/src/modules/agent-signal/index.ts`
 
 ## Implementation Rules
@@ -86,7 +86,7 @@ Keep the boundaries strict:
 - Use stable ids and idempotency keys when the same source can arrive more than once.
 - Preserve scope discipline. The runtime uses `scopeKey` to serialize related background work.
 - Prefer the dedicated shared package types and builders from `@lobechat/agent-signal` for normalized nodes and result contracts.
-- Add focused tests near the touched runtime, policy, or store module. Existing tests under `src/server/services/agentSignal/**/__tests__` are the reference pattern.
+- Add focused tests near the touched runtime, policy, or store module. Existing tests under `apps/server/src/services/agentSignal/**/__tests__` are the reference pattern.
 
 ## References
 

.agents/skills/agent-signal/references/architecture.md

@@ -32,9 +32,9 @@ source node
 
 Read:
 
-- `src/server/services/agentSignal/index.ts`
-- `src/server/services/agentSignal/sources/index.ts`
-- `src/server/services/agentSignal/runtime/AgentSignalScheduler.ts`
+- `apps/server/src/services/agentSignal/index.ts`
+- `apps/server/src/services/agentSignal/sources/index.ts`
+- `apps/server/src/services/agentSignal/runtime/AgentSignalScheduler.ts`
 
 ## Package Boundaries
 
@@ -56,7 +56,7 @@ Read:
 - `packages/agent-signal/src/types/events.ts`
 - `packages/agent-signal/src/types/builtin.ts`
 
-### `src/server/services/agentSignal`
+### `apps/server/src/services/agentSignal`
 
 Treat this as the server-owned implementation layer.
 
@@ -89,11 +89,11 @@ Examples:
 
 Define source payloads in:
 
-- `src/server/services/agentSignal/sourceTypes.ts`
+- `apps/server/src/services/agentSignal/sourceTypes.ts`
 
 Build normalized sources in:
 
-- `src/server/services/agentSignal/sources/buildSource.ts`
+- `apps/server/src/services/agentSignal/sources/buildSource.ts`
 - `packages/agent-signal/src/base/builders.ts`
 
 ### Signal
@@ -109,7 +109,7 @@ Examples from `analyzeIntent`:
 
 Define server-owned signal types in:
 
-- `src/server/services/agentSignal/policies/types.ts`
+- `apps/server/src/services/agentSignal/policies/types.ts`
 
 ### Action
 
@@ -157,9 +157,9 @@ When a user asks for "the procedure", document the flow above and point to the e
 
 Read:
 
-- `src/server/services/agentSignal/sources/index.ts`
-- `src/server/services/agentSignal/runtime/context.ts`
-- `src/server/services/agentSignal/constants.ts`
+- `apps/server/src/services/agentSignal/sources/index.ts`
+- `apps/server/src/services/agentSignal/runtime/context.ts`
+- `apps/server/src/services/agentSignal/constants.ts`
 
 Use `enqueueAgentSignalSourceEvent(...)` when the work should stay quiet and out-of-band. That path:
 
@@ -172,8 +172,8 @@ This is the preferred path when the UI request should finish immediately and the
 
 Read:
 
-- `src/server/workflows/agentSignal/index.ts`
-- `src/server/workflows/agentSignal/run.ts`
+- `apps/server/src/workflows/agentSignal/index.ts`
+- `apps/server/src/workflows/agentSignal/run.ts`
 
 ## Existing Example: `analyzeIntent`
 
@@ -192,8 +192,8 @@ agent.user.message
 
 Read:
 
-- `src/server/services/agentSignal/policies/analyzeIntent/index.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/feedbackSatisfaction.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/feedbackDomain.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/feedbackAction.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/index.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackSatisfaction.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackDomain.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackAction.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`

.agents/skills/agent-signal/references/handlers.md

@@ -2,7 +2,7 @@
 
 ## Fluent Registration API
 
-Use the middleware helpers in `src/server/services/agentSignal/runtime/middleware.ts`.
+Use the middleware helpers in `apps/server/src/services/agentSignal/runtime/middleware.ts`.
 
 They provide:
 
@@ -32,7 +32,7 @@ The context gives you:
 
 Read:
 
-- `src/server/services/agentSignal/runtime/context.ts`
+- `apps/server/src/services/agentSignal/runtime/context.ts`
 
 ## Return Contracts
 
@@ -48,7 +48,7 @@ Return one of these shapes:
 Read:
 
 - `packages/agent-signal/src/base/types.ts`
-- `src/server/services/agentSignal/runtime/AgentSignalScheduler.ts`
+- `apps/server/src/services/agentSignal/runtime/AgentSignalScheduler.ts`
 
 ## Policy Composition Pattern
 
@@ -72,8 +72,8 @@ That bundle is later passed into the runtime via:
 
 Read:
 
-- `src/server/services/agentSignal/policies/index.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/index.ts`
+- `apps/server/src/services/agentSignal/policies/index.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/index.ts`
 
 ## Source Handler Pattern
 
@@ -81,7 +81,7 @@ Use a source handler when you are interpreting a producer event into semantic si
 
 Reference:
 
-- `src/server/services/agentSignal/policies/analyzeIntent/feedbackSatisfaction.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackSatisfaction.ts`
 
 Pattern:
 
@@ -114,8 +114,8 @@ Use a signal handler when one semantic state should branch into more semantic st
 
 References:
 
-- `src/server/services/agentSignal/policies/analyzeIntent/feedbackDomain.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/feedbackAction.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackDomain.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/feedbackAction.ts`
 
 Pattern:
 
@@ -148,7 +148,7 @@ Use an action handler when the runtime should do actual work.
 
 Reference:
 
-- `src/server/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`
 
 Pattern:
 
@@ -186,9 +186,9 @@ Keep these rules:
 Use this split:
 
 - external event payloads:
-  `src/server/services/agentSignal/sourceTypes.ts`
+  `apps/server/src/services/agentSignal/sourceTypes.ts`
 - policy-owned signal and action payloads:
-  `src/server/services/agentSignal/policies/types.ts`
+  `apps/server/src/services/agentSignal/policies/types.ts`
 - normalized shared node contracts:
   `packages/agent-signal/src/base/types.ts`
 
@@ -216,10 +216,10 @@ Prefer focused tests near the touched code.
 
 Useful references:
 
-- `src/server/services/agentSignal/runtime/__tests__/AgentSignalRuntime.test.ts`
-- `src/server/services/agentSignal/__tests__/index.integration.test.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/__tests__/*`
-- `src/server/services/agentSignal/policies/analyzeIntent/actions/__tests__/*`
+- `apps/server/src/services/agentSignal/runtime/__tests__/AgentSignalRuntime.test.ts`
+- `apps/server/src/services/agentSignal/__tests__/index.integration.test.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/__tests__/*`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/actions/__tests__/*`
 
 Test at the smallest level that proves the behavior:
 

.agents/skills/agent-signal/references/observability.md

@@ -24,9 +24,9 @@ After runtime execution, the service projects one compact observability model fr
 
 Read:
 
-- `src/server/services/agentSignal/observability/projector.ts`
-- `src/server/services/agentSignal/observability/traceEvents.ts`
-- `src/server/services/agentSignal/observability/store.ts`
+- `apps/server/src/services/agentSignal/observability/projector.ts`
+- `apps/server/src/services/agentSignal/observability/traceEvents.ts`
+- `apps/server/src/services/agentSignal/observability/store.ts`
 
 Projection outputs:
 
@@ -58,7 +58,7 @@ Workflow-triggered runs do not naturally pass through the normal foreground runt
 
 Read:
 
-- `src/server/workflows/agentSignal/run.ts`
+- `apps/server/src/workflows/agentSignal/run.ts`
 
 Use that path when:
 
@@ -77,8 +77,8 @@ Check:
 
 Read:
 
-- `src/server/services/agentSignal/index.ts`
-- `src/server/services/agentSignal/sources/index.ts`
+- `apps/server/src/services/agentSignal/index.ts`
+- `apps/server/src/services/agentSignal/sources/index.ts`
 
 ### The signal exists but no action runs
 
@@ -98,8 +98,8 @@ Check:
 
 Reference:
 
-- `src/server/services/agentSignal/policies/actionIdempotency.ts`
-- `src/server/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`
+- `apps/server/src/services/agentSignal/policies/actionIdempotency.ts`
+- `apps/server/src/services/agentSignal/policies/analyzeIntent/actions/userMemory.ts`
 
 ### Background runs are hard to discover
 

.agents/skills/agent-testing/SKILL.md

@@ -0,0 +1,376 @@
+---
+name: agent-testing
+description: >
+  Agentic end-to-end testing for LobeHub: backend verification via the CLI,
+  frontend verification via agent-browser (Electron), full-stack verification in
+  the browser, and bot-channel verification via osascript. Local-first today,
+  designed to extend to cloud automation. Triggers on 'cli test', 'test with cli',
+  'verify with cli', 'backend test with cli', 'local test', 'test in electron',
+  'test desktop', 'test bot', 'bot test', 'test in discord', 'test in telegram',
+  'test in slack', 'test in wechat', 'test in weixin', 'test in lark', 'test in feishu',
+  'test in qq', 'manual test', 'osascript', 'test report', or any local
+  end-to-end verification task.
+---
+
+# Agent Testing (Agentic End-to-End Verification)
+
+One skill for all agentic end-to-end testing — local-first today, designed to
+also run as full cloud automation. Every test session follows the same
+four-step contract:
+
+```
+Step -1: Plan approval  →  Step 0: Env + Auth  →  Step 1: Pick surface  →  Step 2: Run  →  Step 3: Structured report
+```
+
+## Step -1 — Plan approval for non-trivial tests
+
+Skip directly to Step 0 if: the test is a single re-run after a fix, the plan
+was already agreed on, or the user gave exact commands.
+
+Otherwise, propose a test plan (surface, cases, expected evidence, assumptions)
+and use the runtime structured question tool (`request_user_input` /
+ask-user-question equivalent) with two fixed choices:
+
+1. `开始执行 (Recommended)` — 测试方案没问题，开始执行
+2. `先讨论下` — 方案有问题，先讨论下
+
+Wait for the user's choice before proceeding.
+
+## Step 0 — Environment setup + auth check (mandatory)
+
+Step 0 is about getting the environment ready: **dependencies are healthy**
+and **auth is green**. A test run that dies halfway on a missing dependency or
+a login wall wastes the whole session — clear both gates BEFORE writing a
+single test step.
+
+### 0.0 Resolve the current test environment
+
+Before starting a dev server, checking auth, opening agent-browser, or writing
+test steps, print and confirm the current local test environment:
+
+```bash
+./.agents/skills/agent-testing/scripts/test-env.sh
+```
+
+This command is the source of truth for local test ports. It reads the current
+shell plus `.env` files using the same precedence as `scripts/runWithEnv.mts`,
+then prints:
+
+- `APP_URL`
+- `PORT`
+- `SERVER_URL`
+- `AUTH_TRUSTED_ORIGINS`
+- `SPA_PORT`
+- `MOBILE_SPA_PORT`
+- `DESKTOP_PORT`
+
+For commands that need these values, export them from the same resolver:
+
+```bash
+eval "$(./.agents/skills/agent-testing/scripts/test-env.sh --exports)"
+```
+
+Do not rely on hard-coded port tables. If the printed values do not match the
+running dev server, fix/export the env first, then continue.
+
+### 0.1 Dependencies are installed — root AND standalone apps
+
+The root pnpm workspace does **NOT** cover every app: `pnpm-workspace.yaml`
+lists `packages/**`, `e2e`, `apps/server`, and only `apps/desktop/src/main` —
+**`apps/desktop` and `apps/cli` are standalone**, each keeping its own
+`node_modules` with its own links into `packages/`. A root install does not
+refresh them, so install in every app the test will touch:
+
+```bash
+pnpm install                    # root workspace
+cd apps/desktop && pnpm install # Electron surface
+cd apps/cli && pnpm install     # CLI surface
+```
+
+Symptom of a stale standalone install: the build/launch fails to resolve a
+recently added workspace package — `Rolldown failed to resolve import
+"@lobechat/<pkg>"` (Electron) or `Cannot find module '@lobechat/<pkg>'` (CLI).
+
+### 0.2 Run scripts from the repo root
+
+All paths in this skill (`./.agents/skills/agent-testing/...`) are
+repo-root-relative, and background commands inherit the current working
+directory — a script launched while `cwd` is `apps/desktop` fails with
+`No such file or directory`. Verify `pwd` is the repo root before launching
+long-running scripts.
+
+### 0.3 Init local dev env without `.env`
+
+For Web smoke against local code, start a **normal local dev environment**.
+First check the repo root for `.env`:
+
+- If `.env` exists, use the existing local configuration and start the dev
+  server normally.
+- If `.env` does not exist, use the agent-testing env bootstrap.
+
+Do not start the standalone e2e server as the product under test.
+
+Use `scripts/init-dev-env.sh`. It follows the e2e setup pattern — Postgres,
+migrations, auth/key-vault/S3 test env, seed user — but it is owned by this
+skill and starts the repo's dev server (`pnpm run dev:next` / `bun run dev`),
+not `e2e/scripts/setup.ts --start`. The script hard-blocks when root `.env`
+exists, so it cannot accidentally override a user's local config. When `.env`
+exists, do not call any `init-dev-env.sh` subcommand.
+
+Decision flow:
+
+```bash
+if [[ -f .env ]]; then
+  bun run dev
+else
+  ./.agents/skills/agent-testing/scripts/init-dev-env.sh setup-db
+  ./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user
+  ./.agents/skills/agent-testing/scripts/init-dev-env.sh dev
+fi
+```
+
+Bootstrap flow when no `.env` exists:
+
+```bash
+# From repo root. Managed DB flow requires Docker Desktop.
+./.agents/skills/agent-testing/scripts/init-dev-env.sh setup-db
+./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user
+./.agents/skills/agent-testing/scripts/init-dev-env.sh dev
+```
+
+If using an existing Postgres instead of the managed Docker DB, set
+`DATABASE_URL` and skip `setup-db`:
+
+```bash
+DATABASE_URL=postgresql://... ./.agents/skills/agent-testing/scripts/init-dev-env.sh migrate
+DATABASE_URL=postgresql://... ./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user
+DATABASE_URL=postgresql://... ./.agents/skills/agent-testing/scripts/init-dev-env.sh dev
+```
+
+For backend-only checks, `dev-next` is available, but Web smoke needs the
+full-stack `dev` command so Next can proxy the SPA HTML from Vite:
+
+```bash
+./.agents/skills/agent-testing/scripts/init-dev-env.sh dev-next
+```
+
+Useful subcommands:
+
+```bash
+./.agents/skills/agent-testing/scripts/init-dev-env.sh env       # print exports
+./.agents/skills/agent-testing/scripts/init-dev-env.sh write     # write .records/env/agent-testing-dev.env
+./.agents/skills/agent-testing/scripts/init-dev-env.sh migrate   # migrations only
+./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user # seed user + CLI API key
+./.agents/skills/agent-testing/scripts/init-dev-env.sh qstash    # local QStash for workflow paths
+./.agents/skills/agent-testing/scripts/init-dev-env.sh clean-db  # remove managed DB container
+```
+
+Default script env:
+
+- `APP_URL=http://localhost:3010`
+- `DATABASE_URL=postgresql://postgres:postgres@localhost:5433/postgres`
+- `DATABASE_DRIVER=node`
+- `FEATURE_FLAGS=-agent_self_iteration` so local smoke does not require QStash
+- Local QStash defaults (`QSTASH_URL`, `QSTASH_TOKEN`, signing keys) are exported;
+  run `init-dev-env.sh qstash` in a separate terminal when the path under test
+  triggers QStash/Workflow.
+- `KEY_VAULTS_SECRET`, `AUTH_SECRET`, auth verification off
+- S3 mock vars
+- Managed DB container: `lobehub-agent-testing-postgres`
+
+`seed-user` creates `agent-testing@lobehub.com` / `TestPassword123!` with
+onboarding already completed, plus a local API key in
+`.records/env/agent-testing-cli.env` for CLI automation. When running Cucumber
+against this dev server, pass the same script env into the test process too;
+Cucumber has its own `BeforeAll` seed path and it must see `DATABASE_URL`
+instead of silently skipping setup:
+
+```bash
+cd e2e
+# Only in the no-.env branch.
+eval "$(../.agents/skills/agent-testing/scripts/init-dev-env.sh env)"
+BASE_URL=http://localhost:3010 HEADLESS=true bun run test:smoke
+```
+
+### 0.4 Auth is green for the selected surface
+
+**Auth is the gate for automated testing, but the gate is surface-scoped.**
+Pick the intended surface first when it is already clear from the task, then
+check only that surface. Do not block a Web test on CLI device-code auth or an
+Electron login state unless the test spans those surfaces.
+
+```bash
+./.agents/skills/agent-testing/scripts/setup-auth.sh status --surface web
+```
+
+Use `status` with no `--surface` only for cross-surface test plans.
+
+| Surface  | Mechanism                                     | One-key path             | Standard check                            |
+| -------- | --------------------------------------------- | ------------------------ | ----------------------------------------- |
+| CLI      | Seeded API key, device-code fallback          | `setup-auth.sh cli-seed` | `setup-auth.sh status --surface cli`      |
+| Web      | Seeded better-auth login into `agent-browser` | `setup-auth.sh web-seed` | `setup-auth.sh status --surface web`      |
+| Electron | App's own persistent login state              | Log in once in the app   | `setup-auth.sh status --surface electron` |
+| Bot      | Native apps already logged in                 | —                        | per-platform screenshot                   |
+
+Login-state checks are standardized — do NOT hand-roll `window.__LOBE_STORES`
+eval snippets; use `scripts/app-probe.sh auth` (returns `{ isSignedIn, userId }`,
+works for Electron CDP and web sessions via `AB_TARGET`).
+
+For Web tests, the test surface is always `agent-browser --session lobehub-dev`.
+Use `setup-auth.sh web-seed` first in the seeded local env. The user's normal
+Chrome is only a source for copying the Cookie header when seed auth is not
+available or `status --surface web` still fails. If Chrome is already logged in,
+do not open a login page; verify agent-browser first, then request the Network
+`Cookie:` header only if that verification fails. Full background and failure modes:
+[references/auth.md](./references/auth.md).
+
+## Step 1 — Pick the surface by change scope
+
+| Change scope                                            | Default surface                      | Why                                                               | Guide                              |
+| ------------------------------------------------------- | ------------------------------------ | ----------------------------------------------------------------- | ---------------------------------- |
+| **Backend** (TRPC router / service / model / migration) | **CLI**                              | Fastest loop, text-assertable output, zero UI flakiness           | [cli/index.md](./cli/index.md)     |
+| **Pure frontend** (components, store, styles, UX)       | **Electron** (agent-browser + CDP)   | Primary product shape; `__LOBE_STORES` state introspection        | [ui/electron.md](./ui/electron.md) |
+| **Full-stack** (new API + UI consuming it)              | **Web** (browser + local dev server) | One surface where network requests and UI are observable together | [ui/web.md](./ui/web.md)           |
+| **Bot channels** (Discord / WeChat / Lark / …)          | Native app via osascript / bridge    | Only way to exercise the real channel end-to-end                  | `bot/<platform>/index.md`          |
+
+Escalate, don't duplicate: verify a backend change with the CLI first; only add
+a UI pass when the change actually affects the UI.
+
+### Environment support (local macOS vs cloud Linux)
+
+The decisive constraint per surface is **how evidence (screenshots) is
+captured**: CDP-based capture (`agent-browser screenshot`) renders from the
+browser engine and needs no real display; OS-level capture (`screencapture`,
+osascript) is macOS-only.
+
+| Surface  | macOS (local) | Linux / cloud (headless)                                  | Screenshot mechanism                                   |
+| -------- | ------------- | --------------------------------------------------------- | ------------------------------------------------------ |
+| CLI      | ✅            | ✅                                                        | n/a — text output                                      |
+| Web      | ✅            | ✅ headless Chromium works natively                       | CDP — no display needed                                |
+| Electron | ✅            | ⚠️ runs, but needs a display server: wrap with `xvfb-run` | CDP works under Xvfb; `capture-app-window.sh` does NOT |
+| Bot      | ✅            | ❌ osascript + native apps are macOS-only                 | macOS `screencapture` only                             |
+
+When a test must stay cloud-portable, prefer CDP-based evidence over
+OS-level capture wherever both exist.
+
+### Bot platforms
+
+| Platform      | Guide                                            | Quick switcher        |
+| ------------- | ------------------------------------------------ | --------------------- |
+| Discord       | [bot/discord/index.md](./bot/discord/index.md)   | `Cmd+K`               |
+| Slack         | [bot/slack/index.md](./bot/slack/index.md)       | `Cmd+K`               |
+| Telegram      | [bot/telegram/index.md](./bot/telegram/index.md) | `Cmd+F`               |
+| WeChat / 微信 | [bot/wechat/index.md](./bot/wechat/index.md)     | `Cmd+F`               |
+| Lark / 飞书   | [bot/lark/index.md](./bot/lark/index.md)         | `Cmd+K`               |
+| QQ            | [bot/qq/index.md](./bot/qq/index.md)             | `Cmd+F`               |
+| iMessage      | [bot/imessage/index.md](./bot/imessage/index.md) | bridge (no osascript) |
+
+Each platform folder contains an `index.md` (activation, navigation,
+send-message, verification snippets) and a `test-<platform>-bot.sh` script
+sharing the interface:
+
+```bash
+./.agents/skills/agent-testing/bot/<platform>/test-<platform>-bot.sh <channel_or_contact> <message> [wait_seconds] [screenshot_path]
+```
+
+New to osascript automation? Read
+[references/osascript.md](./references/osascript.md) first — it is a general
+macOS-automation asset (activate, type, paste, screenshot, accessibility reads,
+gotchas), not bot-specific.
+
+## Step 2 — Run
+
+Surface guides above carry the detailed workflows. Shared infrastructure:
+
+| Need                                 | Where                                                                |
+| ------------------------------------ | -------------------------------------------------------------------- |
+| Start / restart the local dev server | [references/dev-server.md](./references/dev-server.md)               |
+| `agent-browser` command reference    | [references/agent-browser.md](./references/agent-browser.md)         |
+| osascript patterns (general macOS)   | [references/osascript.md](./references/osascript.md)                 |
+| Agent gateway probing                | [references/agent-gateway.md](./references/agent-gateway.md)         |
+| Screen recording                     | [references/record-app-screen.md](./references/record-app-screen.md) |
+
+### Scripts
+
+All under `.agents/skills/agent-testing/scripts/`:
+
+| Script                    | Usage                                                                        |
+| ------------------------- | ---------------------------------------------------------------------------- |
+| `test-env.sh`             | Print/export the resolved local test env and ports                           |
+| `setup-auth.sh`           | One-stop auth setup & status check (`status` / `cli` / `web`)                |
+| `init-dev-env.sh`         | Self-contained local dev env (`setup-db` / `seed-user` / `dev-next` / `dev`) |
+| `app-probe.sh`            | LobeHub app probes: `auth` / `route` / `ops` / `goto <path>` / `errors`      |
+| `record-gif.sh`           | Frame-sequence → GIF for time-based behavior (streaming, timers, animations) |
+| `report-init.sh`          | Scaffold a structured test report (Step 3)                                   |
+| `electron-dev.sh`         | Manage Electron dev env (start/stop/status/restart, CDP 9222)                |
+| `capture-app-window.sh`   | Screenshot a specific app window (general; used by bot tests)                |
+| `record-app-screen.sh`    | Record app screen (video + periodic screenshots)                             |
+| `record-electron-demo.sh` | Record Electron app demo with ffmpeg                                         |
+| `agent-gateway/`          | Gateway probe / dump / analyze tools                                         |
+
+`app-probe.sh` is the LobeHub-specific fast path into app state — auth check,
+current route, running operations, and `goto <path>` quick navigation
+(`/agent/<agentId>/<topicId>`, `/task/<taskId>`, `/settings`, …) so a test can
+jump straight to the state under test instead of clicking through the UI. See
+[ui/electron.md](./ui/electron.md#lobehub-probes--quick-navigation) for usage.
+
+## Step 3 — Structured report (mandatory deliverable)
+
+Every automated test session ends with a structured, evidence-backed report —
+not a chat-only summary. Scaffold it up front and fill it as you test:
+
+```bash
+DIR=$(./.agents/skills/agent-testing/scripts/report-init.sh my-feature "Verify my feature")
+# ... test, saving screenshots / CLI transcripts into $DIR/assets/ ...
+# fill $DIR/report.md (scope, case table with inline evidence, verdict, score) and $DIR/result.json
+```
+
+Reports live in `.records/reports/<timestamp>-<slug>/` (gitignored): `report.md`
+(human-readable, with screenshots/GIFs embedded directly in the case table),
+`result.json` (machine-readable pass/fail + score), `assets/` (evidence).
+Format spec and evidence rules:
+[references/report.md](./references/report.md).
+
+Two hard rules worth front-loading:
+
+- **Report language = the user's conversation language.** Write the ENTIRE
+  `report.md` (headings included) in the language the user is conversing in —
+  no mixed English. `result.json` keys/status values stay English.
+- **The case table is the main reading surface.** Prefer the compact
+  `# | case | result | key observation | evidence` shape and embed the
+  screenshot/GIF in the evidence cell. Use separate evidence sections only for
+  long CLI transcripts, HAR summaries, or supplemental detail.
+- **Visual evidence must render inline.** Screenshots and GIFs in `report.md`
+  must use Markdown image syntax like `![case 1](assets/case1.png)`. Do not
+  use bare file paths, Markdown links, or local file links as the primary
+  visual evidence; those make the report unreadable without opening each asset.
+- **Final replies must include visual evidence links.** When a run includes UI
+  screenshots or GIFs, include the report directory and the most important
+  visual artifacts in the final chat response. Each item must include a stable
+  label, an evidence caption describing the observed UI outcome, and a
+  repo-relative path, for example:
+  `[Image #1 - error toast shows provider auth failure](<report-dir>/assets/foo.png)`.
+  Use repo-relative paths, not absolute paths.
+- **Time-based behavior needs a GIF, not a screenshot.** If a case asserts
+  change over time (streaming output, a ticking timer, loading states,
+  animations), record it with `scripts/record-gif.sh` and embed the GIF —
+  a static screenshot cannot prove the behavior.
+
+## Directory map
+
+```
+agent-testing/
+├── SKILL.md            # this router
+├── cli/index.md        # backend verification via the LobeHub CLI
+├── ui/electron.md      # pure-frontend verification in the desktop app
+├── ui/web.md           # full-stack verification in the browser
+├── bot/<platform>/     # bot-channel verification (osascript / bridge)
+├── references/         # shared knowledge: auth, dev-server, agent-browser, osascript, report
+└── scripts/            # setup-auth, report-init, electron-dev, capture, recording, gateway
+```
+
+## Gotchas
+
+- agent-browser: see [references/agent-browser.md](./references/agent-browser.md#gotchas)
+- Electron: see [ui/electron.md](./ui/electron.md#electron-gotchas)
+- osascript: see [references/osascript.md](./references/osascript.md#gotchas)

.agents/skills/agent-testing/bot/discord/index.md

@@ -0,0 +1,97 @@
+# Discord Bot Testing
+
+**App name:** `Discord` | **Process name:** `Discord`
+
+See [references/osascript.md](../../references/osascript.md) for shared patterns.
+
+## Activate & Navigate
+
+```bash
+# Activate Discord
+osascript -e 'tell application "Discord" to activate'
+sleep 1
+
+# Open Quick Switcher (Cmd+K) to navigate to a channel
+osascript -e 'tell application "System Events" to keystroke "k" using command down'
+sleep 0.5
+osascript -e 'tell application "System Events" to keystroke "bot-testing"'
+sleep 1
+osascript -e 'tell application "System Events" to key code 36' # Enter
+sleep 2
+```
+
+## Send Message to Bot
+
+```bash
+# The message input is focused after navigating to a channel
+# Type a message
+osascript -e 'tell application "System Events" to keystroke "/hello"'
+sleep 0.5
+osascript -e 'tell application "System Events" to key code 36' # Enter
+```
+
+## Send Long Message (via clipboard)
+
+```bash
+osascript -e '
+tell application "Discord" to activate
+delay 0.5
+set the clipboard to "Write a 3000 word essay about space exploration"
+tell application "System Events"
+    keystroke "v" using command down
+    delay 0.3
+    key code 36  -- Enter
+end tell
+'
+```
+
+## Verify Bot Response
+
+```bash
+# Wait for bot to respond, then screenshot
+sleep 10
+screencapture /tmp/discord-bot-response.png
+# Read with the Read tool for visual verification
+```
+
+## Full Bot Test Example
+
+```bash
+#!/usr/bin/env bash
+# test-discord-bot.sh — Send message and verify bot response
+
+# 1. Activate Discord and navigate to channel
+osascript -e '
+tell application "Discord" to activate
+delay 1
+-- Quick Switcher
+tell application "System Events" to keystroke "k" using command down
+delay 0.5
+tell application "System Events" to keystroke "bot-testing"
+delay 1
+tell application "System Events" to key code 36
+delay 2
+'
+
+# 2. Send test message
+osascript -e '
+set the clipboard to "!ping"
+tell application "System Events"
+    keystroke "v" using command down
+    delay 0.3
+    key code 36
+end tell
+'
+
+# 3. Wait for response and capture
+sleep 5
+screencapture /tmp/discord-test-result.png
+echo "Screenshot saved to /tmp/discord-test-result.png"
+```
+
+## Script
+
+```bash
+./.agents/skills/agent-testing/bot/discord/test-discord-bot.sh "bot-testing" "!ping"
+./.agents/skills/agent-testing/bot/discord/test-discord-bot.sh "bot-testing" "/ask Tell me a joke" 30
+```

.agents/skills/agent-testing/bot/discord/test-discord-bot.sh

@@ -60,5 +60,5 @@ echo "[$APP] Waiting ${WAIT}s for bot response..."
 sleep "$WAIT"
 
 echo "[$APP] Capturing screenshot..."
-"$SCRIPT_DIR/capture-app-window.sh" "$APP" "$SCREENSHOT"
+"$SCRIPT_DIR/../../scripts/capture-app-window.sh" "$APP" "$SCREENSHOT"
 echo "[$APP] Done! Screenshot saved to $SCREENSHOT"

.agents/skills/agent-testing/bot/imessage/index.md

@@ -0,0 +1,232 @@
+# iMessage Desktop bridge regression test
+
+The iMessage channel is different from the other bot platforms: there is **no
+native app to drive with osascript**. Instead the Desktop app runs a local
+**BlueBubbles bridge** — a small HTTP server in the Electron main process that
+registers a webhook on a local [BlueBubbles](https://bluebubbles.app/) server,
+receives iMessage events, and forwards them to LobeHub Cloud.
+
+So the test surface is three layers:
+
+1. **Electron main IPC** — `imessageBridge.*` handlers (`getStatus`,
+   `testConfig`, `upsertConfig`, `removeConfig`, `start`, `stop`)
+2. **Local bridge HTTP server** — `http://127.0.0.1:<port>/webhooks/bluebubbles/<appId>?secret=<secret>`
+3. **BlueBubbles REST API** — `http://127.0.0.1:1234/api/v1/*` (webhook + server/info)
+
+## Prerequisites
+
+- A running **BlueBubbles server** (macOS, default `http://127.0.0.1:1234`) with
+  a known password. Sanity check:
+  ```bash
+  curl -sS -m4 -o /dev/null -w '%{http_code}\n' \
+    "http://127.0.0.1:1234/api/v1/server/info?password=<PW>" # expect 200
+  ```
+- **Electron dev running with CDP**: `./.agents/skills/agent-testing/scripts/electron-dev.sh start`
+- The **iMessage Desktop branch** checked out (the `imessageBridge` IPC group
+  and `@lobechat/chat-adapter-imessage` must be compiled into the main bundle).
+  Run `pnpm install --ignore-scripts` at the repo root **and** in `apps/desktop/`
+  after switching branches — the new workspace package must be linked or the
+  main build fails to resolve `@lobechat/chat-adapter-imessage`.
+
+## Fast path: automated script
+
+```bash
+./.agents/skills/agent-testing/bot/imessage/test-imessage-bridge.sh '<bluebubbles_password>' [bb_url] [cdp_port]
+```
+
+Asserts the whole flow and self-cleans (unique `applicationId` per run, removes
+its bridge config + BlueBubbles webhook on exit). Exit 0 = all green. It covers:
+
+- BlueBubbles reachable + password valid; Electron CDP reachable; IPC available
+- `testConfig` happy path → success
+- `testConfig` wrong password → rejected; unreachable URL → rejected
+- `upsertConfig` **first-time save → success** (Bug #1 regression guard, below)
+- `getStatus` → `running:true`, config persisted, password redacted (`blueBubblesPasswordSet`)
+- BlueBubbles webhook actually registered for the appId
+- Local bridge HTTP server: wrong secret → 401; valid secret → past auth
+
+The password is passed as argv (visible in `ps`) — local dev only, don't use a
+real secret on a shared machine.
+
+## Layer 1 — IPC probes (no UI)
+
+The renderer exposes the main-process handlers via `window.electronAPI.invoke`.
+This is the quickest way to exercise the bridge without clicking:
+
+```bash
+# baseline
+agent-browser --cdp 9222 eval \
+  "(async()=>JSON.stringify(await window.electronAPI.invoke('imessageBridge.getStatus',{})))()"
+
+# test a connection (note: password as a JS string)
+agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
+(async function () {
+  try {
+    var r = await window.electronAPI.invoke('imessageBridge.testConfig', {
+      applicationId: 'probe',
+      blueBubblesServerUrl: 'http://127.0.0.1:1234',
+      blueBubblesPassword: 'PASTE_PW',
+      enabled: true,
+      webhookSecret: 'probe-secret',
+    });
+    return JSON.stringify(r);            // { success: true }
+  } catch (e) { return 'ERR: ' + (e.message || e); }
+})()
+EVALEOF
+```
+
+`upsertConfig` persists to the Electron store, starts the local HTTP server, and
+registers the BlueBubbles webhook. `removeConfig` + `stop` reverse it.
+
+## Layer 2 — full UI flow (agent-browser)
+
+The bridge settings only render in Desktop (`isDesktop` guard) under the agent's
+**Channel → iMessage** screen. The platform tile only appears as a real (non
+"Coming Soon") entry once the server registers `imessage` **and** the frontend
+drops it from `COMING_SOON_PLATFORMS` (`src/routes/(main)/agent/channel/const.ts`).
+
+```bash
+agent-browser --cdp 9222 open "http://localhost:5173/agent/<aid>/channel"
+agent-browser --cdp 9222 wait --load networkidle && agent-browser --cdp 9222 wait 1500
+
+# confirm the remote backend lists imessage (it must be registered + deployed)
+agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
+(async function(){
+  var url='lobe-backend://lobe/trpc/lambda/agentBotProvider.listPlatforms?input='+encodeURIComponent('{"json":null,"meta":{"values":["undefined"],"v":1}}');
+  var d=await (await fetch(url,{credentials:'include'})).json();
+  var p=d.result?.data?.json||d;
+  return JSON.stringify(p.map(function(x){return x.id;}));
+})()
+EVALEOF
+
+# click the iMessage tile, then fill the form by ref
+agent-browser --cdp 9222 eval "(()=>{var b=[...document.querySelectorAll('aside button')].find(x=>/imessage/i.test(x.textContent));b&&b.click();})()"
+agent-browser --cdp 9222 wait 1500
+agent-browser --cdp 9222 snapshot -i | grep -iE "127.0.0.1:1234|Application ID|Webhook Secret|Test BlueBubbles|Save Bridge"
+```
+
+Field refs (from the snapshot): Application ID, Webhook Secret, BlueBubbles
+Server URL (`placeholder="http://127.0.0.1:1234"`), and a **nested** textbox right
+under the URL one is the BlueBubbles Password. Fill with `fill` (real input
+events — `eval`-setting React inputs won't fire onChange), click **Test
+BlueBubbles**, then **Save Bridge**. Read the antd toast immediately (it
+auto-dismisses):
+
+```bash
+agent-browser --cdp 9222 eval \
+  "JSON.stringify([...new Set([...document.querySelectorAll('.ant-message-custom-content')].map(n=>n.textContent.trim()))])"
+# Test  → "BlueBubbles connection passed"
+# Save  → "iMessage Desktop bridge saved"
+```
+
+Verify the end state via BlueBubbles + IPC:
+
+```bash
+curl -sS "http://127.0.0.1:1234/api/v1/webhook?password=<PW>" # webhook for the appId present
+agent-browser --cdp 9222 eval "(async()=>JSON.stringify(await window.electronAPI.invoke('imessageBridge.getStatus',{})))()"
+# running:true, serverUrl: http://127.0.0.1:33270, configs[].blueBubblesPasswordSet:true
+```
+
+Cleanup: `removeConfig` + `stop` via IPC, then `DELETE /api/v1/webhook/<id>` on
+BlueBubbles.
+
+## Outbound send test (desktop → BlueBubbles → iMessage)
+
+Verifies the leg the bridge uses to _reply_: `BlueBubblesApiClient.sendText`
+→ `POST /api/v1/message/text`. Run the helper against your own number:
+
+```bash
+./.agents/skills/agent-testing/bot/imessage/send-imessage-test.sh '<bb_password>' '+<E164>' # e.g. +15551234567
+```
+
+**Gotcha that bites everyone:** with `method=apple-script` and a _new_
+conversation, the HTTP POST often **times out** even though the message is
+sent. Never judge success by the HTTP response. Instead poll
+`POST /api/v1/message/query` and read the matching `isFromMe:true` row's
+`error` field:
+
+- `error: 0` (or null) → sent OK
+- non-zero `error` → real send failure
+
+The script does exactly this: fires the send, ignores the timeout, then matches
+its marker text in the message store and asserts `error == 0`.
+
+Two more notes:
+
+- Use a full E.164 handle (`iMessage;-;+<countrycode><number>`) or an Apple ID
+  email. Looking the chat up by guid afterwards may 404 if BB filed the message
+  under a differently-formatted guid — that's a lookup quirk, not a send failure.
+- Sending to _your own_ number round-trips: BB records both the outgoing
+  (`fromMe:true`) and an incoming copy (`fromMe:false`).
+
+## Inbound e2e test (iMessage → cloud agent → reply)
+
+Full inbound chain: a message arrives → BlueBubbles fires its `new-message`
+webhook → local bridge (`:33270`) → `forwardWebhook` POSTs to
+`<remote>/api/agent/webhooks/imessage/<appId>?secret=…` → cloud agent → reply
+flows back via Device Gateway → BB `sendText`.
+
+Prerequisites:
+
+- A cloud bot provider for the same `applicationId` exists and is **connected**
+  (Save Configuration + the device gateway connected — a _disconnected_ gateway
+  yields `DEVICE_NOT_FOUND` on connect and blocks the reply leg).
+- The `imessage` Labs toggle is on (otherwise the channel is gated to "Coming
+  Soon"), and `webhookSecret` matches on both ends (auto-generated on save).
+
+Two ways to drive it:
+
+1. **Second device / Apple ID (recommended).** Have _another_ Apple ID message
+   the BB-hosted number (e.g. "please reply pong"). The bot replies; you see it
+   on the other device. **No loop risk** — the reply goes to the other party,
+   not back to itself.
+2. **Send to your own number (quick, loop-aware).** `sendText` to the hosted
+   number; the loopback _incoming_ copy (`isFromMe:false`) triggers the bot.
+   Watch the reply land in `message/query` as a `fromMe:true` row.
+
+**Loop guard — why a self-send doesn't spin forever:** the Chat SDK adapter
+drops any `isFromMe` message before dispatch
+(`packages/chat-adapter-imessage/src/adapter.ts`: `if (message.isFromMe) return`).
+The bot's own reply (`isFromMe:true`) is never re-processed, so in the normal
+case (someone else → bot → reply to them) there is no loop. The self-send case
+is a **test-only edge**: the bot's reply also round-trips to your number, and
+only the adapter's `isFromMe` check stops a second pass. Keep the prompt
+conversational (so the bot doesn't keep finding something to answer), and
+**turn the `imessage` lab off / remove the config when done** — never leave a
+self-send bot running unattended.
+
+Watch the chain live:
+
+```bash
+tail -f /tmp/electron-dev.log | grep -iE "imessage|bridge|forward|Message API"
+# the agent reply shows up as a fromMe:true row with the bot's text:
+curl -sS -X POST "http://127.0.0.1:1234/api/v1/message/query?password=<PW>" \
+  -H 'Content-Type: application/json' -d '{"limit":5,"sort":"DESC"}'
+```
+
+`startTyping` will log a Private-API error unless BlueBubbles has the Private
+API helper set up (needs a jailbroken / SIP-disabled Mac) — it's logged and
+ignored; text replies still work.
+
+## Known bugs / gotchas
+
+- **Bug #1 — first-time save (fixed; guarded by the script).** BlueBubbles'
+  `GET /api/v1/webhook?url=<unregistered>` returns **HTTP 500**
+  (`Cannot read properties of null (reading 'events')`). The bridge must list
+  **all** webhooks and match client-side, never pass the `?url=` filter. If you
+  see `upsertConfig` fail with "An unhandled error has occurred!" originating in
+  `listWebhooks`, this regressed.
+- **Save leaves a half-state on webhook failure.** `upsertConfig` writes the
+  config + starts the HTTP server _before_ registering the webhook, so a webhook
+  failure still reports `running:true` with the config persisted but no
+  BlueBubbles webhook. Always assert the BlueBubbles webhook list, not just IPC
+  status.
+- **Unknown appId / forward failure → 500.** Posting to the local bridge for an
+  unknown appId, or when no cloud bot is bound, returns 500 (BlueBubbles retries
+  on 5xx). Auth (wrong secret → 401) is enforced before that.
+- **Backend deploy lag.** Desktop dev proxies tRPC through `lobe-backend://` to
+  the _remote_ server. iMessage only appears in `listPlatforms` once the server
+  registration is deployed there, regardless of local branch.
+- **Restart to load main-process fixes.** Editing `imessageBridgeSrv.ts` /
+  `@lobechat/chat-adapter-imessage` needs `electron-dev.sh restart` — main isn't
+  hot-replaced. On restart, enabled configs auto-register their webhook again.

.agents/skills/agent-testing/bot/imessage/send-imessage-test.sh

@@ -0,0 +1,81 @@
+#!/usr/bin/env bash
+#
+# send-imessage-test.sh — Verify the outbound leg: desktop → BlueBubbles → iMessage
+#
+# Sends one real iMessage via the same REST call the Desktop bridge uses
+# (`POST /api/v1/message/text`, which BlueBubblesApiClient.sendText wraps) and
+# confirms it actually went out.
+#
+# KEY GOTCHA: with method=apple-script and a NEW conversation, the HTTP request
+# often TIMES OUT even though the message is sent. Do NOT treat the timeout as a
+# failure — instead poll `POST /api/v1/message/query` and check the message's
+# `error` field (0 = sent OK). This script does that for you.
+#
+# This sends a REAL message, so it has side effects. Target your own number.
+#
+# Usage:
+#   ./send-imessage-test.sh <bb_password> <target_e164> [message] [bb_url]
+#
+# Example (send to your own phone, E.164 with country code):
+#   ./send-imessage-test.sh 'my-bb-pass' '+15551234567'
+#
+set -euo pipefail
+
+BB_PASS="${1:?Usage: $0 <bb_password> <target_e164(+countrycode)> [message] [bb_url]}"
+TARGET="${2:?Need a target handle in E.164, e.g. +15551234567 (or an Apple ID email)}"
+MARKER="lobe-imsg-test-$(date +%s)"
+MESSAGE="${3:-[${MARKER}] desktop bridge → BlueBubbles → iMessage outbound check}"
+BB_URL="${4:-http://127.0.0.1:1234}"
+
+CHAT_GUID="iMessage;-;${TARGET}"
+
+echo "[send-test] target=${TARGET}  marker=${MARKER}"
+
+# 1) Fire the send. apple-script on a new chat may hang the HTTP response, so we
+#    cap it short and ignore a timeout — step 2 is the source of truth.
+python3 - "$BB_PASS" "$BB_URL" "$CHAT_GUID" "$MESSAGE" <<'PY' || true
+import json,sys,urllib.request,urllib.parse,uuid
+pw,base,guid,msg=sys.argv[1:5]
+url=base+"/api/v1/message/text?password="+urllib.parse.quote(pw)
+body={"chatGuid":guid,"message":msg,"method":"apple-script","tempGuid":str(uuid.uuid4())}
+req=urllib.request.Request(url,data=json.dumps(body).encode("utf-8"),
+    headers={"Content-Type":"application/json"},method="POST")
+try:
+    r=urllib.request.urlopen(req,timeout=8)
+    print("[send-test] HTTP",r.status,"(immediate response)")
+except urllib.error.HTTPError as e:
+    print("[send-test] HTTP",e.code,e.read().decode()[:200])
+except Exception as e:
+    print("[send-test] HTTP request returned no body (likely apple-script delay):",type(e).__name__)
+PY
+
+# 2) Source of truth: find our marker in the message store and read its error.
+echo "[send-test] verifying via message/query (the HTTP timeout above is expected)…"
+sleep 3
+python3 - "$BB_PASS" "$BB_URL" "$MARKER" <<'PY'
+import json,sys,time,urllib.request,urllib.parse
+pw,base,marker=sys.argv[1:4]
+url=base+"/api/v1/message/query?password="+urllib.parse.quote(pw)
+def query():
+    body={"limit":15,"offset":0,"with":["chats"],"sort":"DESC"}
+    req=urllib.request.Request(url,data=json.dumps(body).encode(),
+        headers={"Content-Type":"application/json"},method="POST")
+    return json.load(urllib.request.urlopen(req,timeout=12)).get("data") or []
+hit=None
+for _ in range(5):
+    for m in query():
+        if marker in (m.get("text") or "") and m.get("isFromMe"):
+            hit=m; break
+    if hit: break
+    time.sleep(2)
+if not hit:
+    print("[send-test] ✗ outbound message not found in BB store — send likely failed")
+    sys.exit(1)
+err=hit.get("error")
+if err in (0,None):
+    print("[send-test] ✓ outbound message sent (fromMe=True, error=%s)"%err)
+    print("[send-test]   → confirm it arrived in the Messages app on the target device")
+else:
+    print("[send-test] ✗ BlueBubbles reported send error=%s"%err)
+    sys.exit(1)
+PY

.agents/skills/agent-testing/bot/imessage/test-imessage-bridge.sh

@@ -0,0 +1,187 @@
+#!/usr/bin/env bash
+#
+# test-imessage-bridge.sh — Regression test for the iMessage Desktop bridge
+#
+# Drives the Electron main-process `imessageBridge.*` IPC handlers plus the
+# local bridge HTTP server and the BlueBubbles server, asserting the full
+# connect/configure flow. Use it to regression-test PR work on the iMessage
+# channel (BlueBubbles bridge) without clicking through the UI every time.
+#
+# Prerequisites:
+#   1. BlueBubbles server running and reachable (default http://127.0.0.1:1234)
+#   2. Electron dev running with CDP — `electron-dev.sh start`
+#   3. `agent-browser` on PATH, connected to the same CDP port
+#
+# Usage:
+#   ./test-imessage-bridge.sh <bluebubbles_password> [bb_url] [cdp_port]
+#
+# Example:
+#   ./test-imessage-bridge.sh 'my-bb-password'
+#   ./test-imessage-bridge.sh 'my-bb-password' http://127.0.0.1:1234 9222
+#
+# Notes:
+#   - The password is passed as an argv, so it is visible in `ps`. This is a
+#     local dev tool; do not run it on shared machines with a real secret.
+#   - It uses a unique applicationId per run (imsg-regression-$$) and cleans up
+#     its own bridge config + BlueBubbles webhook on exit, so it is safe to
+#     re-run and does not disturb real configs.
+set -euo pipefail
+
+BB_PASS="${1:?Usage: $0 <bluebubbles_password> [bb_url] [cdp_port]}"
+BB_URL="${2:-http://127.0.0.1:1234}"
+CDP_PORT="${3:-9222}"
+
+APP_ID="imsg-regression-$$"
+SECRET="regression-secret-$$"
+
+PASS=0
+FAIL=0
+
+# ── Output helpers ───────────────────────────────────────────────────
+ok()   { echo "  ✓ $1"; PASS=$((PASS + 1)); }
+bad()  { echo "  ✗ $1 — $2"; FAIL=$((FAIL + 1)); }
+note() { echo "[imsg-test] $1"; }
+
+# ── BlueBubbles REST helpers ─────────────────────────────────────────
+bb_get_webhooks() {
+  curl -sS -m 8 "${BB_URL}/api/v1/webhook?password=${BB_PASS}"
+}
+
+# Delete every webhook whose URL mentions our APP_ID (cleanup is idempotent).
+bb_cleanup_webhooks() {
+  local ids
+  ids=$(bb_get_webhooks | python3 -c '
+import json,sys
+try: d=json.load(sys.stdin)
+except Exception: sys.exit(0)
+for w in (d.get("data") or []):
+    if "'"$APP_ID"'" in (w.get("url") or ""): print(w["id"])
+' 2>/dev/null || true)
+  for id in $ids; do
+    curl -sS -m 8 -X DELETE "${BB_URL}/api/v1/webhook/${id}?password=${BB_PASS}" >/dev/null 2>&1 || true
+  done
+}
+
+# ── IPC helper (drives the Electron renderer's electronAPI bridge) ───
+# Runs a JS snippet that returns a string token; prints the raw token.
+# The BlueBubbles password is base64-injected (atob) so special chars in the
+# secret never need shell/JS quoting.
+ipc_eval() {
+  local js="$1"
+  agent-browser --cdp "$CDP_PORT" eval -b "$(printf '%s' "$js" | base64)" 2>/dev/null
+}
+
+PASS_B64=$(printf '%s' "$BB_PASS" | base64)
+
+# Emit an inline JS object literal for the bridge config. $1 overrides the
+# password expression (defaults to atob of the real password); pass a JS string
+# literal like "'wrong'" to test the rejection path.
+ipc_config_js() {
+  local pwexpr="${1:-atob('${PASS_B64}')}"
+  printf "{applicationId:'%s',blueBubblesServerUrl:'%s',blueBubblesPassword:%s,enabled:true,webhookSecret:'%s'}" \
+    "$APP_ID" "$BB_URL" "$pwexpr" "$SECRET"
+}
+
+# ── Preflight ────────────────────────────────────────────────────────
+note "BlueBubbles: ${BB_URL}   CDP: ${CDP_PORT}   appId: ${APP_ID}"
+
+code=$(curl -sS -m 6 -o /dev/null -w '%{http_code}' \
+  "${BB_URL}/api/v1/server/info?password=${BB_PASS}" || echo 000)
+if [ "$code" = "200" ]; then ok "BlueBubbles reachable + password valid"; else
+  bad "BlueBubbles preflight" "HTTP $code (is BlueBubbles running on ${BB_URL}?)"
+  echo "Aborting — fix BlueBubbles first."; exit 1
+fi
+
+if ! curl -sf --max-time 3 "http://localhost:${CDP_PORT}/json/version" >/dev/null 2>&1; then
+  bad "Electron CDP preflight" "CDP ${CDP_PORT} unreachable — run electron-dev.sh start"
+  echo "Aborting."; exit 1
+fi
+ok "Electron CDP reachable"
+
+# Bridge must expose the IPC group (built from this branch's code).
+probe=$(ipc_eval "(async()=>{try{var s=await window.electronAPI.invoke('imessageBridge.getStatus',{});return 'OK:'+JSON.stringify(s);}catch(e){return 'ERR:'+(e.message||e);}})()")
+case "$probe" in
+  *OK:*) ok "imessageBridge IPC available" ;;
+  *) bad "imessageBridge IPC" "got: $probe (is the iMessage Desktop branch checked out?)"; echo "Aborting."; exit 1 ;;
+esac
+
+# Start clean: remove any leftover config for this appId + BB webhooks.
+ipc_eval "(async()=>{try{await window.electronAPI.invoke('imessageBridge.removeConfig',{applicationId:'${APP_ID}'});}catch(e){}return 'done';})()" >/dev/null
+bb_cleanup_webhooks
+
+# ── testConfig: happy path ───────────────────────────────────────────
+r=$(ipc_eval "(async()=>{try{var c=$(ipc_config_js);var x=await window.electronAPI.invoke('imessageBridge.testConfig',c);return 'OK:'+JSON.stringify(x);}catch(e){return 'ERR:'+(e.message||e);}})()")
+case "$r" in
+  *OK:*success*true*) ok "testConfig with valid password → success" ;;
+  *) bad "testConfig (valid)" "got: $r" ;;
+esac
+
+# ── testConfig: wrong password rejects ───────────────────────────────
+r=$(ipc_eval "(async()=>{try{var c=$(ipc_config_js "'definitely-wrong-password'");var x=await window.electronAPI.invoke('imessageBridge.testConfig',c);return 'OK:'+JSON.stringify(x);}catch(e){return 'ERR:'+(e.message||e);}})()")
+case "$r" in
+  *ERR:*) ok "testConfig with wrong password → rejected" ;;
+  *) bad "testConfig (wrong password)" "expected rejection, got: $r" ;;
+esac
+
+# ── testConfig: unreachable URL rejects ──────────────────────────────
+r=$(ipc_eval "(async()=>{try{var x=await window.electronAPI.invoke('imessageBridge.testConfig',{applicationId:'${APP_ID}',blueBubblesServerUrl:'http://127.0.0.1:65530',blueBubblesPassword:atob('${PASS_B64}'),enabled:true,webhookSecret:'${SECRET}'});return 'OK:'+JSON.stringify(x);}catch(e){return 'ERR:'+(e.message||e);}})()")
+case "$r" in
+  *ERR:*) ok "testConfig with unreachable URL → rejected" ;;
+  *) bad "testConfig (unreachable)" "expected rejection, got: $r" ;;
+esac
+
+# ── upsertConfig: FIRST-TIME registration (Bug #1 regression guard) ──
+# BlueBubbles' GET /webhook?url=<unregistered> returns HTTP 500. The bridge
+# must list ALL webhooks and match client-side, otherwise this first save
+# fails. This assertion guards that fix.
+r=$(ipc_eval "(async()=>{try{var c=$(ipc_config_js);var x=await window.electronAPI.invoke('imessageBridge.upsertConfig',c);return 'OK:'+JSON.stringify(x);}catch(e){return 'ERR:'+(e.message||e);}})()")
+case "$r" in
+  *OK:*success*true*) ok "upsertConfig first-time save → success (Bug #1 guard)" ;;
+  *) bad "upsertConfig (first-time)" "got: $r" ;;
+esac
+
+# ── getStatus: bridge running + config persisted ─────────────────────
+# Return a quote-free token so grep isn't tripped up by agent-browser's
+# JSON-string escaping of the eval result.
+r=$(ipc_eval "(async()=>{var s=await window.electronAPI.invoke('imessageBridge.getStatus',{});var c=(s.configs||[]).find(function(x){return x.applicationId==='${APP_ID}';});return 'RUN='+(s.running?'Y':'N')+' CFG='+(c?'Y':'N')+' PW='+((c&&c.blueBubblesPasswordSet)?'Y':'N');})()")
+echo "$r" | grep -q 'RUN=Y' && ok "bridge running" || bad "bridge running" "got: $r"
+echo "$r" | grep -q 'CFG=Y' && ok "config persisted" || bad "config persisted" "got: $r"
+echo "$r" | grep -q 'PW=Y'  && ok "password stored (redacted in status)" || bad "password stored" "got: $r"
+
+# ── BlueBubbles webhook actually registered ──────────────────────────
+if bb_get_webhooks | grep -q "${APP_ID}"; then
+  ok "BlueBubbles webhook registered for appId"
+else
+  bad "BlueBubbles webhook" "no webhook URL containing ${APP_ID}"
+fi
+
+# ── Local bridge HTTP server: secret enforcement ─────────────────────
+BRIDGE_URL=$(ipc_eval "(async()=>{var s=await window.electronAPI.invoke('imessageBridge.getStatus',{});return s.serverUrl||'';})()" | tr -d '"')
+if [ -n "$BRIDGE_URL" ]; then
+  # wrong secret → 401
+  code=$(curl -sS -m 6 -o /dev/null -w '%{http_code}' -X POST \
+    -H 'Content-Type: application/json' \
+    "${BRIDGE_URL}/webhooks/bluebubbles/${APP_ID}?secret=WRONG" \
+    -d '{"type":"new-message","data":{"guid":"x"}}' || echo 000)
+  [ "$code" = "401" ] && ok "local bridge rejects wrong secret (401)" || bad "local bridge wrong secret" "expected 401, got $code"
+
+  # right secret → passes auth (reaches forward; without a bound cloud bot it
+  # returns 5xx — that's fine, we're only asserting auth + routing here)
+  code=$(curl -sS -m 6 -o /dev/null -w '%{http_code}' -X POST \
+    -H 'Content-Type: application/json' \
+    "${BRIDGE_URL}/webhooks/bluebubbles/${APP_ID}?secret=${SECRET}" \
+    -d '{"type":"new-message","data":{"guid":"x","text":"hi"}}' || echo 000)
+  [ "$code" != "401" ] && ok "local bridge accepts valid secret (HTTP $code, past auth)" || bad "local bridge valid secret" "got 401 with correct secret"
+else
+  bad "local bridge URL" "getStatus returned no serverUrl"
+fi
+
+# ── Cleanup ──────────────────────────────────────────────────────────
+ipc_eval "(async()=>{try{await window.electronAPI.invoke('imessageBridge.removeConfig',{applicationId:'${APP_ID}'});await window.electronAPI.invoke('imessageBridge.stop',{});}catch(e){}return 'cleaned';})()" >/dev/null
+bb_cleanup_webhooks
+note "cleaned up config + BlueBubbles webhook for ${APP_ID}"
+
+# ── Summary ──────────────────────────────────────────────────────────
+echo ""
+echo "[imsg-test] PASS=${PASS}  FAIL=${FAIL}"
+[ "$FAIL" -eq 0 ] || exit 1

.agents/skills/agent-testing/bot/lark/index.md

@@ -0,0 +1,61 @@
+# Lark / 飞书 Bot Testing
+
+**App name:** `Lark` or `飞书` | **Process name:** `Lark` or `飞书`
+
+See [references/osascript.md](../../references/osascript.md) for shared patterns.
+
+## Activate & Navigate
+
+```bash
+# Activate Lark (auto-detects Lark or 飞书)
+osascript -e 'tell application "Lark" to activate' 2> /dev/null \
+  || osascript -e 'tell application "飞书" to activate'
+sleep 1
+
+# Quick Switcher / Search (Cmd+K)
+osascript -e 'tell application "System Events" to keystroke "k" using command down'
+sleep 0.5
+osascript -e '
+set the clipboard to "bot-testing"
+tell application "System Events"
+    keystroke "v" using command down
+    delay 1.5
+    key code 36  -- Enter
+end tell
+'
+sleep 2
+```
+
+## Send Message to Bot
+
+```bash
+osascript -e '
+set the clipboard to "@MyBot help me with this task"
+tell application "System Events"
+    keystroke "v" using command down
+    delay 0.3
+    key code 36  -- Enter
+end tell
+'
+```
+
+## Verify Response
+
+```bash
+sleep 10
+screencapture /tmp/lark-bot-response.png
+```
+
+## Lark-Specific Notes
+
+- App name varies: `Lark` (international) vs `飞书` (China mainland) — the script auto-detects
+- Uses `Cmd+K` for quick search (same as Discord/Slack)
+- Enter sends message by default
+- Always use clipboard paste for CJK characters
+
+## Script
+
+```bash
+./.agents/skills/agent-testing/bot/lark/test-lark-bot.sh "bot-testing" "@MyBot hello"
+./.agents/skills/agent-testing/bot/lark/test-lark-bot.sh "bot-testing" "Help me with this" 30
+```

.agents/skills/agent-testing/bot/lark/test-lark-bot.sh

@@ -80,5 +80,5 @@ echo "[$APP] Waiting ${WAIT}s for bot response..."
 sleep "$WAIT"
 
 echo "[$APP] Capturing screenshot..."
-"$SCRIPT_DIR/capture-app-window.sh" "$APP" "$SCREENSHOT"
+"$SCRIPT_DIR/../../scripts/capture-app-window.sh" "$APP" "$SCREENSHOT"
 echo "[$APP] Done! Screenshot saved to $SCREENSHOT"

.agents/skills/agent-testing/bot/qq/index.md

@@ -0,0 +1,62 @@
+# QQ Bot Testing
+
+**App name:** `QQ` | **Process name:** `QQ`
+
+See [references/osascript.md](../../references/osascript.md) for shared patterns.
+
+## Activate & Navigate
+
+```bash
+osascript -e 'tell application "QQ" to activate'
+sleep 1
+
+# Search for contact/group (Cmd+F)
+osascript -e '
+tell application "System Events"
+    keystroke "f" using command down
+    delay 0.8
+end tell
+'
+osascript -e '
+set the clipboard to "bot-testing"
+tell application "System Events"
+    keystroke "v" using command down
+    delay 1.5
+    key code 36  -- Enter
+end tell
+'
+sleep 2
+```
+
+## Send Message to Bot
+
+```bash
+osascript -e '
+set the clipboard to "Hello bot!"
+tell application "System Events"
+    keystroke "v" using command down
+    delay 0.3
+    key code 36  -- Enter
+end tell
+'
+```
+
+## Verify Response
+
+```bash
+sleep 10
+screencapture /tmp/qq-bot-response.png
+```
+
+## QQ-Specific Notes
+
+- Enter sends message by default; Shift+Enter for newlines
+- Uses `Cmd+F` for search (not `Cmd+K` like Discord/Slack/Lark)
+- Always use clipboard paste for CJK characters
+
+## Script
+
+```bash
+./.agents/skills/agent-testing/bot/qq/test-qq-bot.sh "bot-testing" "Hello bot" 15
+./.agents/skills/agent-testing/bot/qq/test-qq-bot.sh "MyBot" "/help" 10
+```

.agents/skills/agent-testing/bot/qq/test-qq-bot.sh

@@ -72,5 +72,5 @@ echo "[$APP] Waiting ${WAIT}s for bot response..."
 sleep "$WAIT"
 
 echo "[$APP] Capturing screenshot..."
-"$SCRIPT_DIR/capture-app-window.sh" "$APP" "$SCREENSHOT"
+"$SCRIPT_DIR/../../scripts/capture-app-window.sh" "$APP" "$SCREENSHOT"
 echo "[$APP] Done! Screenshot saved to $SCREENSHOT"

.agents/skills/agent-testing/bot/slack/index.md

@@ -0,0 +1,73 @@
+# Slack Bot Testing
+
+**App name:** `Slack` | **Process name:** `Slack`
+
+See [references/osascript.md](../../references/osascript.md) for shared patterns.
+
+## Activate & Navigate
+
+```bash
+# Activate Slack
+osascript -e 'tell application "Slack" to activate'
+sleep 1
+
+# Quick Switcher (Cmd+K)
+osascript -e 'tell application "System Events" to keystroke "k" using command down'
+sleep 0.5
+osascript -e 'tell application "System Events" to keystroke "bot-testing"'
+sleep 1
+osascript -e 'tell application "System Events" to key code 36' # Enter
+sleep 2
+```
+
+## Send Message to Bot
+
+```bash
+# Direct message input (focused after channel nav)
+osascript -e 'tell application "System Events" to keystroke "@mybot hello"'
+sleep 0.3
+osascript -e 'tell application "System Events" to key code 36'
+```
+
+## Send Long Message
+
+```bash
+osascript -e '
+tell application "Slack" to activate
+delay 0.5
+set the clipboard to "A long test message for the bot..."
+tell application "System Events"
+    keystroke "v" using command down
+    delay 0.3
+    key code 36
+end tell
+'
+```
+
+## Slash Command Test
+
+```bash
+osascript -e '
+tell application "Slack" to activate
+delay 0.5
+tell application "System Events"
+    keystroke "/ask What is the meaning of life?"
+    delay 0.5
+    key code 36
+end tell
+'
+```
+
+## Verify Response
+
+```bash
+sleep 10
+screencapture /tmp/slack-bot-response.png
+```
+
+## Script
+
+```bash
+./.agents/skills/agent-testing/bot/slack/test-slack-bot.sh "bot-testing" "@mybot hello"
+./.agents/skills/agent-testing/bot/slack/test-slack-bot.sh "bot-testing" "/ask What is 2+2?" 20
+```

.agents/skills/agent-testing/bot/slack/test-slack-bot.sh

@@ -60,5 +60,5 @@ echo "[$APP] Waiting ${WAIT}s for bot response..."
 sleep "$WAIT"
 
 echo "[$APP] Capturing screenshot..."
-"$SCRIPT_DIR/capture-app-window.sh" "$APP" "$SCREENSHOT"
+"$SCRIPT_DIR/../../scripts/capture-app-window.sh" "$APP" "$SCREENSHOT"
 echo "[$APP] Done! Screenshot saved to $SCREENSHOT"

.agents/skills/agent-testing/bot/telegram/index.md

@@ -0,0 +1,80 @@
+# Telegram Bot Testing
+
+**App name:** `Telegram` | **Process name:** `Telegram`
+
+See [references/osascript.md](../../references/osascript.md) for shared patterns.
+
+## Activate & Navigate
+
+```bash
+# Activate Telegram
+osascript -e 'tell application "Telegram" to activate'
+sleep 1
+
+# Search for a bot (Cmd+F or click search)
+osascript -e '
+tell application "System Events"
+    keystroke "f" using command down
+    delay 0.5
+    keystroke "MyTestBot"
+    delay 1
+    key code 36  -- Enter to select
+end tell
+'
+sleep 2
+```
+
+## Send Message to Bot
+
+```bash
+# After navigating to bot chat, input is focused
+osascript -e '
+tell application "System Events"
+    keystroke "/start"
+    delay 0.3
+    key code 36
+end tell
+'
+```
+
+## Send Long Message
+
+```bash
+osascript -e '
+tell application "Telegram" to activate
+delay 0.5
+set the clipboard to "Tell me about quantum computing in detail"
+tell application "System Events"
+    keystroke "v" using command down
+    delay 0.3
+    key code 36
+end tell
+'
+```
+
+## Verify Response
+
+```bash
+sleep 10
+screencapture /tmp/telegram-bot-response.png
+```
+
+## Telegram Bot API (programmatic alternative)
+
+For sending messages directly to the bot's chat without UI:
+
+```bash
+# Send message as the bot (for testing webhooks/responses)
+curl -s "https://api.telegram.org/bot$TELEGRAM_BOT_TOKEN/sendMessage" \
+  -d "chat_id=$CHAT_ID&text=test message"
+
+# Get recent updates
+curl -s "https://api.telegram.org/bot$TELEGRAM_BOT_TOKEN/getUpdates?limit=5" | jq .
+```
+
+## Script
+
+```bash
+./.agents/skills/agent-testing/bot/telegram/test-telegram-bot.sh "MyTestBot" "/start"
+./.agents/skills/agent-testing/bot/telegram/test-telegram-bot.sh "GPTBot" "Hello" 60
+```

.agents/skills/agent-testing/bot/telegram/test-telegram-bot.sh

@@ -75,5 +75,5 @@ echo "[$APP] Waiting ${WAIT}s for bot response..."
 sleep "$WAIT"
 
 echo "[$APP] Capturing screenshot..."
-"$SCRIPT_DIR/capture-app-window.sh" "$APP" "$SCREENSHOT"
+"$SCRIPT_DIR/../../scripts/capture-app-window.sh" "$APP" "$SCREENSHOT"
 echo "[$APP] Done! Screenshot saved to $SCREENSHOT"

.agents/skills/agent-testing/bot/wechat/index.md

@@ -0,0 +1,81 @@
+# WeChat / 微信 Bot Testing
+
+**App name:** `微信` or `WeChat` | **Process name:** `WeChat`
+
+See [references/osascript.md](../../references/osascript.md) for shared patterns.
+
+## Activate & Navigate
+
+```bash
+# Activate WeChat
+osascript -e 'tell application "微信" to activate'
+sleep 1
+
+# Search for a contact/bot (Cmd+F)
+osascript -e '
+tell application "System Events"
+    keystroke "f" using command down
+    delay 0.5
+    keystroke "TestBot"
+    delay 1
+    key code 36  -- Enter to select
+end tell
+'
+sleep 2
+```
+
+## Send Message
+
+```bash
+# After navigating to a chat, the input is focused
+osascript -e '
+tell application "System Events"
+    keystroke "Hello bot!"
+    delay 0.3
+    key code 36
+end tell
+'
+```
+
+## Send Long Message (clipboard)
+
+```bash
+osascript -e '
+tell application "微信" to activate
+delay 0.5
+set the clipboard to "Please help me with this task..."
+tell application "System Events"
+    keystroke "v" using command down
+    delay 0.3
+    key code 36
+end tell
+'
+```
+
+## Verify Response
+
+```bash
+sleep 10
+screencapture /tmp/wechat-bot-response.png
+```
+
+## WeChat-Specific Notes
+
+- WeChat macOS app name can be `微信` or `WeChat` depending on system language. Try both:
+  ```bash
+  osascript -e 'tell application "微信" to activate' 2> /dev/null \
+    || osascript -e 'tell application "WeChat" to activate'
+  ```
+- WeChat uses **Enter** to send (not Cmd+Enter by default, but configurable)
+- For multi-line messages without sending, use **Shift+Enter**:
+  ```bash
+  osascript -e 'tell application "System Events" to key code 36 using shift down'
+  ```
+- Always use clipboard paste for CJK characters — `keystroke` mangles non-ASCII
+
+## Script
+
+```bash
+./.agents/skills/agent-testing/bot/wechat/test-wechat-bot.sh "文件传输助手" "test message" 5
+./.agents/skills/agent-testing/bot/wechat/test-wechat-bot.sh "MyBot" "Tell me a joke" 30
+```

.agents/skills/agent-testing/bot/wechat/test-wechat-bot.sh

@@ -81,5 +81,5 @@ echo "[$APP] Waiting ${WAIT}s for bot response..."
 sleep "$WAIT"
 
 echo "[$APP] Capturing screenshot..."
-"$SCRIPT_DIR/capture-app-window.sh" "$APP" "$SCREENSHOT"
+"$SCRIPT_DIR/../../scripts/capture-app-window.sh" "$APP" "$SCREENSHOT"
 echo "[$APP] Done! Screenshot saved to $SCREENSHOT"

.agents/skills/agent-testing/cli/index.md

@@ -0,0 +1,152 @@
+# CLI Backend Verification
+
+Default surface for verifying **backend changes** (TRPC routers, services,
+models, migrations) end-to-end: fastest loop, text-assertable output, zero UI
+flakiness.
+
+## When to use
+
+- Verifying TRPC router / service / model changes end-to-end
+- Testing new API fields or response structure changes
+- Validating CLI command output after backend modifications
+- Debugging data flow issues between server and CLI
+
+## Prerequisites
+
+| Requirement  | Details                                                                                                                                        |
+| ------------ | ---------------------------------------------------------------------------------------------------------------------------------------------- |
+| Dev server   | `localhost:3010` — see [../references/dev-server.md](../references/dev-server.md)                                                              |
+| CLI source   | `apps/cli/` — runs from source, no rebuild; standalone `node_modules` — run `pnpm install` inside `apps/cli/` (root install does not cover it) |
+| CLI dev mode | `LOBEHUB_CLI_HOME=.lobehub-dev` for isolated settings                                                                                          |
+| Auth         | Seeded API key first; Device Code Flow only as fallback — see [../references/auth.md](../references/auth.md)                                   |
+
+All CLI dev commands run from `apps/cli/`. Subsequent examples use `$CLI`:
+
+```bash
+source ../../.records/env/agent-testing-cli.env
+CLI="bun src/index.ts"
+```
+
+## Workflow
+
+### Step 1 — Server up?
+
+See [../references/dev-server.md](../references/dev-server.md) for the health
+check, start, and restart commands. Server-side code changes require a restart.
+
+### Step 2 — Auth ready?
+
+```bash
+./.agents/skills/agent-testing/scripts/setup-auth.sh status
+```
+
+If the CLI is not ready in the seeded local environment:
+
+```bash
+./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user
+source .records/env/agent-testing-cli.env
+./.agents/skills/agent-testing/scripts/setup-auth.sh cli-seed
+```
+
+If the target environment is not seeded, use the interactive fallback:
+
+```bash
+cd apps/cli && LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts login --server http://localhost:3010
+```
+
+Seeded API-key auth does not store credentials. It writes local settings under
+`$HOME/.lobehub-dev` and requires the generated env file to be sourced before
+CLI commands. Details:
+[../references/auth.md](../references/auth.md).
+
+### Step 3 — Test with CLI commands
+
+CLI runs from source, so CLI-side code changes take effect immediately without
+rebuilding:
+
+```bash
+cd apps/cli
+$CLI <command>
+```
+
+Capture output for the report as you go (e.g. `$CLI task list | tee "$DIR/assets/task-list.txt"`).
+
+### Step 4 — Clean up test data
+
+```bash
+$CLI task delete < id > -y
+$CLI agent delete < id > -y
+```
+
+### Step 5 — Report
+
+Finish with a structured report —
+[../references/report.md](../references/report.md). CLI evidence = exact
+command + trimmed output.
+
+## Common testing patterns
+
+### Task system
+
+```bash
+$CLI task list
+$CLI task create -n "Root Task" -i "Test instruction"
+$CLI task create -n "Child Task" -i "Sub instruction" --parent T-1
+$CLI task view T-1
+$CLI task tree T-1
+$CLI task edit T-1 --status running
+$CLI task comment T-1 -m "Test comment"
+$CLI task delete T-1 -y
+```
+
+### Agent system
+
+```bash
+$CLI agent list
+$CLI agent view <agent-id>
+$CLI agent run <agent-id> -m "Test prompt"
+```
+
+### Document & knowledge base
+
+```bash
+$CLI doc list
+$CLI doc create -t "Test Doc" -c "Content here"
+$CLI doc view <doc-id>
+$CLI kb list
+$CLI kb tree <kb-id>
+```
+
+### Model & provider
+
+```bash
+$CLI model list
+$CLI provider list
+$CLI provider test <provider-id>
+```
+
+## Dev-test cycle
+
+```
+1. Make code changes (service/model/router/type)
+         |
+2. Run unit tests (fast feedback)
+   bunx vitest run --silent='passed-only' '<test-file>'
+         |
+3. Restart dev server (if server-side changes — see dev-server.md)
+         |
+4. CLI verification (end-to-end)
+   $CLI <command>
+         |
+5. Clean up test data + write the report
+```
+
+## Troubleshooting
+
+| Issue                       | Solution                                                                                               |
+| --------------------------- | ------------------------------------------------------------------------------------------------------ |
+| `No authentication found`   | Source `.records/env/agent-testing-cli.env`, or run device-code `login --server http://localhost:3010` |
+| `UNAUTHORIZED` on API calls | Re-run `init-dev-env.sh seed-user` and re-source the env file; for device-code fallback, re-run login  |
+| `ECONNREFUSED`              | Dev server not running — see dev-server.md                                                             |
+| CLI shows old data/behavior | Server needs restart to pick up code changes                                                           |
+| Login opens wrong server    | Must use `--server` flag (env var doesn't work)                                                        |

.agents/skills/agent-testing/references/agent-browser.md

@@ -0,0 +1,257 @@
+# agent-browser CLI Reference
+
+Generic reference for the `agent-browser` CLI — automate Chromium-based apps (Electron, Chrome, web) via Chrome DevTools Protocol. LobeHub-specific patterns live in [../ui/electron.md](../ui/electron.md) and [../ui/web.md](../ui/web.md); authentication recipes live in [auth.md](./auth.md).
+
+Use `agent-browser` to automate Chromium-based apps via Chrome DevTools Protocol.
+
+Install via `npm i -g agent-browser`, `brew install agent-browser`, or `cargo install agent-browser`. Run `agent-browser install` to download Chrome. Run `agent-browser upgrade` to update.
+
+## Core Workflow
+
+Every browser automation follows this pattern:
+
+1. **Navigate**: `agent-browser open <url>`
+2. **Snapshot**: `agent-browser snapshot -i` (get element refs like `@e1`, `@e2`)
+3. **Interact**: Use refs to click, fill, select
+4. **Re-snapshot**: After navigation or DOM changes, get fresh refs
+
+```bash
+agent-browser open https://example.com/form
+agent-browser snapshot -i
+# Output: @e1 [input type="email"], @e2 [input type="password"], @e3 [button] "Submit"
+
+agent-browser fill @e1 "user@example.com"
+agent-browser fill @e2 "password123"
+agent-browser click @e3
+agent-browser wait --load networkidle
+agent-browser snapshot -i # Check result
+```
+
+## Command Chaining
+
+```bash
+# Chain open + wait + snapshot in one call
+agent-browser open https://example.com && agent-browser wait --load networkidle && agent-browser snapshot -i
+```
+
+Use `&&` when you don't need to read intermediate output. Run commands separately when you need to parse output first (e.g., snapshot to discover refs, then interact).
+
+## Essential Commands
+
+```bash
+# Navigation
+agent-browser open <url>              # Navigate (aliases: goto, navigate)
+agent-browser close                   # Close browser
+agent-browser close --all             # Close all active sessions
+
+# Snapshot
+agent-browser snapshot -i             # Interactive elements with refs (recommended)
+agent-browser snapshot -s "#selector" # Scope to CSS selector
+
+# Interaction (use @refs from snapshot)
+agent-browser click @e1               # Click element
+agent-browser click @e1 --new-tab     # Click and open in new tab
+agent-browser fill @e2 "text"         # Clear and type text
+agent-browser type @e2 "text"         # Type without clearing
+agent-browser select @e1 "option"     # Select dropdown option
+agent-browser check @e1               # Check checkbox
+agent-browser press Enter             # Press key
+agent-browser keyboard type "text"    # Type at current focus (no selector)
+agent-browser keyboard inserttext "text"  # Insert without key events
+agent-browser scroll down 500         # Scroll page
+agent-browser scroll down 500 --selector "div.content"  # Scroll within container
+
+# Get information
+agent-browser get text @e1            # Get element text
+agent-browser get url                 # Get current URL
+agent-browser get title               # Get page title
+agent-browser get cdp-url             # Get CDP WebSocket URL
+
+# Wait
+agent-browser wait @e1                # Wait for element
+agent-browser wait --load networkidle # Wait for network idle
+agent-browser wait --url "**/page"    # Wait for URL pattern
+agent-browser wait 2000               # Wait milliseconds
+agent-browser wait --text "Welcome"   # Wait for text to appear
+agent-browser wait --fn "!document.body.innerText.includes('Loading...')"  # Wait for text to disappear
+agent-browser wait "#spinner" --state hidden  # Wait for element to disappear
+
+# Downloads
+agent-browser download @e1 ./file.pdf          # Click element to trigger download
+agent-browser wait --download ./output.zip     # Wait for any download to complete
+
+# Network
+agent-browser network requests                 # Inspect tracked requests
+agent-browser network requests --type xhr,fetch  # Filter by resource type
+agent-browser network requests --method POST   # Filter by HTTP method
+agent-browser network route "**/api/*" --abort # Block matching requests
+agent-browser network har start                # Start HAR recording
+agent-browser network har stop ./capture.har   # Stop and save HAR file
+
+# Viewport & Device Emulation
+agent-browser set viewport 1920 1080          # Set viewport size (default: 1280x720)
+agent-browser set viewport 1920 1080 2        # 2x retina
+agent-browser set device "iPhone 14"          # Emulate device (viewport + user agent)
+
+# Capture
+agent-browser screenshot              # Screenshot to temp dir
+agent-browser screenshot --full       # Full page screenshot
+agent-browser screenshot --annotate   # Annotated screenshot with numbered element labels
+agent-browser pdf output.pdf          # Save as PDF
+
+# Clipboard
+agent-browser clipboard read          # Read text from clipboard
+agent-browser clipboard write "text"  # Write text to clipboard
+agent-browser clipboard copy          # Copy current selection
+agent-browser clipboard paste         # Paste from clipboard
+
+# Dialogs (alert, confirm, prompt, beforeunload)
+agent-browser dialog accept           # Accept dialog
+agent-browser dialog accept "input"   # Accept prompt dialog with text
+agent-browser dialog dismiss          # Dismiss/cancel dialog
+agent-browser dialog status           # Check if dialog is open
+
+# Diff (compare page states)
+agent-browser diff snapshot                        # Compare current vs last snapshot
+agent-browser diff screenshot --baseline before.png  # Visual pixel diff
+agent-browser diff url <url1> <url2>               # Compare two pages
+
+# Streaming
+agent-browser stream enable           # Start WebSocket streaming
+agent-browser stream status           # Inspect streaming state
+agent-browser stream disable          # Stop streaming
+```
+
+## Batch Execution
+
+```bash
+echo '[
+  ["open", "https://example.com"],
+  ["snapshot", "-i"],
+  ["click", "@e1"],
+  ["screenshot", "result.png"]
+]' | agent-browser batch --json
+```
+
+## Authentication
+
+```bash
+# Option 1: Auth vault (credentials stored encrypted)
+echo "$PASSWORD" | agent-browser auth save myapp --url https://app.example.com/login --username user --password-stdin
+agent-browser auth login myapp
+
+# Option 2: Session name (auto-save/restore cookies + localStorage)
+agent-browser --session-name myapp open https://app.example.com/login
+agent-browser close                                                       # State auto-saved
+agent-browser --session-name myapp open https://app.example.com/dashboard # Auto-restored
+
+# Option 3: Persistent profile
+agent-browser --profile ~/.myapp open https://app.example.com/login
+
+# Option 4: State file
+agent-browser state save auth.json
+agent-browser state load auth.json
+```
+
+### LobeHub dev server — inject better-auth cookie
+
+`agent-browser --headed` on macOS can create an off-screen Chromium window, blocking manual login. For a local LobeHub dev server (e.g. `localhost:3010`), copy the `better-auth.session_token` cookie out of a **Network request** in the user's own Chrome DevTools and load it via `state load`. See [auth.md](./auth.md) for the full recipe.
+
+## Semantic Locators (Alternative to Refs)
+
+```bash
+agent-browser find text "Sign In" click
+agent-browser find label "Email" fill "user@test.com"
+agent-browser find role button click --name "Submit"
+agent-browser find placeholder "Search" type "query"
+agent-browser find testid "submit-btn" click
+```
+
+## JavaScript Evaluation (eval)
+
+```bash
+# Simple expressions
+agent-browser eval 'document.title'
+
+# Complex JS: use --stdin with heredoc (RECOMMENDED)
+agent-browser eval --stdin << 'EVALEOF'
+JSON.stringify(
+  Array.from(document.querySelectorAll("img"))
+    .filter(i => !i.alt)
+    .map(i => ({ src: i.src.split("/").pop(), width: i.width }))
+)
+EVALEOF
+
+# Base64 encoding (avoids all shell escaping issues)
+agent-browser eval -b "$(echo -n 'document.title' | base64)"
+```
+
+## Ref Lifecycle
+
+Refs (`@e1`, `@e2`, etc.) are invalidated when the page changes. Always re-snapshot after clicking links/buttons that navigate, form submissions, or dynamic content loading.
+
+## Annotated Screenshots (Vision Mode)
+
+```bash
+agent-browser screenshot --annotate
+# Output includes the image path and a legend:
+#   [1] @e1 button "Submit"
+#   [2] @e2 link "Home"
+agent-browser click @e2 # Click using ref from annotated screenshot
+```
+
+## Parallel Sessions
+
+```bash
+agent-browser --session site1 open https://site-a.com
+agent-browser --session site2 open https://site-b.com
+agent-browser session list
+```
+
+## Connect to Existing Chrome
+
+```bash
+agent-browser --auto-connect snapshot # Auto-discover running Chrome
+agent-browser --cdp 9222 snapshot     # Explicit CDP port
+```
+
+## iOS Simulator (Mobile Safari)
+
+```bash
+agent-browser device list
+agent-browser -p ios --device "iPhone 16 Pro" open https://example.com
+agent-browser -p ios snapshot -i
+agent-browser -p ios tap @e1
+agent-browser -p ios swipe up
+agent-browser -p ios screenshot mobile.png
+agent-browser -p ios close
+```
+
+## Observability Dashboard
+
+```bash
+agent-browser dashboard install
+agent-browser dashboard start # Background server on port 4848
+agent-browser dashboard stop
+```
+
+## Cloud Providers
+
+Use `-p <provider>` to run against cloud browsers: `agentcore`, `browserbase`, `browserless`, `browseruse`, `kernel`.
+
+## Browser Engine Selection
+
+```bash
+agent-browser --engine lightpanda open example.com # 10x faster, 10x less memory
+```
+
+## Gotchas
+
+- **Daemon can get stuck** — if commands hang, `agent-browser close --all` or `pkill -f agent-browser` to reset
+- **HMR invalidates everything** — after code changes, refs break. Re-snapshot or restart
+- **`snapshot -i` doesn't find contenteditable** — use `snapshot -i -C` for rich text editors
+- **`fill` doesn't work on contenteditable** — use `type` for chat inputs
+- **Screenshots go to `~/.agent-browser/tmp/screenshots/`** — read them with the `Read` tool
+- **Dialogs block all commands** — if commands time out, check `agent-browser dialog status`
+- **Default timeout is 25s** — override with `AGENT_BROWSER_DEFAULT_TIMEOUT` (ms) or use explicit waits
+- **Shell quoting corrupts eval** — use `eval --stdin <<'EVALEOF'` for complex JS

.agents/skills/agent-testing/references/agent-gateway.md

@@ -0,0 +1,93 @@
+# LobeHub gateway streaming + tab-switch test harness
+
+Captures store + DOM state at 200ms intervals so we can prove or disprove
+claims like "切回 tab 后消息回到了很早以前". Built for gateway-mode chat but
+works for any LobeHub streaming session.
+
+## Files
+
+`scripts/agent-gateway/`
+
+| File            | Role                                                             |
+| --------------- | ---------------------------------------------------------------- |
+| `probe.js`      | Injects a 200ms sampler + `__PROBE_EVENT` marker + `__switchTab` |
+| `probe-dump.js` | Stops the sampler and returns `{events, samples}` as JSON string |
+| `tab-switch.js` | Runs N round-trip switches between two tabs, marks each step     |
+| `analyze.mjs`   | Node post-processor: timeline + regression detection             |
+
+## Standard workflow
+
+```bash
+# 1. Start Electron with CDP
+./.agents/skills/agent-testing/scripts/electron-dev.sh start
+
+# 2. Navigate to a chat, switch runtime to Cloud Sandbox (gateway mode)
+
+# 3. Install the probe + helpers
+agent-browser --cdp 9222 eval --stdin \
+  < .agents/skills/agent-testing/scripts/agent-gateway/probe.js
+
+# 4. Send a tool-call message — manually or via type+press
+agent-browser --cdp 9222 eval "window.__PROBE_EVENT('SENT')"
+
+# 5. Run the multi-switch driver (auto-picks active tab as BACK and the
+#    rightmost inactive tab as AWAY — edit ROUND_TRIPS / DWELL_MS in the
+#    file if you want different timing)
+agent-browser --cdp 9222 eval --stdin \
+  < .agents/skills/agent-testing/scripts/agent-gateway/tab-switch.js
+
+# 6. Wait for streaming to finish, then dump
+agent-browser --cdp 9222 eval --stdin \
+  < .agents/skills/agent-testing/scripts/agent-gateway/probe-dump.js \
+  > /tmp/probe.json
+
+# 7. Analyze
+node .agents/skills/agent-testing/scripts/agent-gateway/analyze.mjs /tmp/probe.json
+```
+
+The analyzer prints three sections: EVENTS, TIMELINE, REGRESSIONS. If
+REGRESSIONS is non-empty it means content/reasoning/childN dropped on the
+same topic — the symptom users describe.
+
+## What the probe tracks (and why)
+
+`chat.messagesMap` only stores the top-level `assistantGroup` shell. The
+actual streamed content, reasoning, and tool calls live in
+`assistantGroup.children: AssistantContentBlock[]`. Any probe that only
+reads `m.content` / `m.reasoning` will see zeros throughout streaming and
+miss everything that matters. probe.js walks both levels and sums:
+
+- `cT` total content length
+- `rT` total reasoning length
+- `toolT` total tool-call count
+- `childN` number of content blocks
+
+Plus DOM-side signals (`domLen`, search/crawl indicator counts) so you can
+tell store-side regressions apart from render-side regressions.
+
+## Gotchas
+
+- **Optimistic new-topic state.** Before the first chunk lands, messages
+  live under the `<scope>_new` key with `tmp_*` ids and no `topicId` field.
+  probe.js falls back to those when `activeTopicId` is null.
+- **Reasoning resets to 0 are not bugs.** When the assistant finishes
+  thinking and starts tool-use or text, the streaming reasoning buffer
+  empties and the finalised reasoning gets sealed into a completed block.
+  Filter these out manually if needed.
+- **DOM length jitters by a handful of chars** because counters like "(10)"
+  in tool-call labels change as results arrive. analyze.mjs only flags
+  `domLen` drops greater than 100 chars to ignore that noise.
+- **Never identify tabs by innerText.** The active tab's text embeds a
+  ` · <agent name>` suffix, so a search like `'LobeHub Growth'` matches the
+  active tab when the active agent happens to be LobeHub Growth — and you
+  end up clicking the tab you're already on. probe.js uses the stable
+  `data-contextmenu-trigger` attribute (a React `useId()` value that's set
+  per-tab and survives focus changes) plus `data-active="true"` to mark
+  the active one. Helpers exposed:
+  `__listTabs()` / `__clickTabByKey(key)` / `__clickTabByIndex(i)` /
+  `__activeTabKey()`.
+- **`tab-switch.js` fires-and-forgets.** The IIFE kicks off an async loop
+  and returns immediately so the agent-browser CLI eval doesn't blow past
+  its default 25 s timeout. Wait on the `SWITCH_LOOP_DONE` event marker
+  before dumping. Re-running while a loop is in flight is refused — the
+  chaotic data from overlapping runs is not worth debugging.

.agents/skills/agent-testing/references/auth.md

@@ -0,0 +1,166 @@
+# Auth Setup for Local Agent Testing
+
+**Auth is the gate for all automated testing.** Complete
+[Step 0.0](../SKILL.md#00-resolve-the-current-test-environment) first so
+`SERVER_URL` and ports are resolved, then verify auth before writing any test
+step.
+
+Initialize helpers first:
+
+```bash
+SCRIPT="./.agents/skills/agent-testing/scripts/setup-auth.sh"
+TEST_ENV="./.agents/skills/agent-testing/scripts/test-env.sh"
+eval "$($TEST_ENV --exports)"
+```
+
+Quick reference after initialization:
+
+| Command                        | Purpose                                            |
+| ------------------------------ | -------------------------------------------------- |
+| `$SCRIPT status`               | Check all surfaces (server + CLI + web + Electron) |
+| `$SCRIPT status --surface web` | Check only the Web surface gate                    |
+| `$SCRIPT cli-seed`             | Configure CLI API-key auth from the seeded key     |
+| `$SCRIPT cli`                  | Interactive CLI device-code login (user must run)  |
+| `$SCRIPT open-chrome`          | Open Chrome at `SERVER_URL` with DevTools          |
+| `$SCRIPT web-seed`             | Sign in the seeded user and inject cookies         |
+| `pbpaste \| $SCRIPT web`       | Inject a copied Cookie header into agent-browser   |
+| `$SCRIPT web-verify`           | Live-check agent-browser session auth              |
+
+Use `localhost` for Web auth; better-auth cookies are stored for `localhost`,
+not `127.0.0.1`.
+
+## Per-surface overview
+
+| Surface  | Mechanism                                | Persistence                                                       | Human interaction                              |
+| -------- | ---------------------------------------- | ----------------------------------------------------------------- | ---------------------------------------------- |
+| CLI      | Seeded API key or OIDC Device Code Flow  | `.records/env/agent-testing-cli.env` + `$HOME/.lobehub-dev`       | No for seed path; yes for device-code fallback |
+| Web      | Seeded better-auth login or cookie copy  | `~/.lobehub-agent-testing/web-state.json` + agent-browser session | No for seed path; copy cookie only as fallback |
+| Electron | App's own login state                    | Electron user-data dir                                            | Log in once manually in the app                |
+| Bot      | Native apps (Discord/WeChat/…) logged in | Each app's own session                                            | Once per app                                   |
+
+## CLI — Seeded API key
+
+For the self-contained no-root-`.env` dev environment, seed the baseline user
+and API key once:
+
+```bash
+./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user
+source .records/env/agent-testing-cli.env
+./.agents/skills/agent-testing/scripts/setup-auth.sh cli-seed
+```
+
+The seed step writes `LOBE_API_KEY` for humans and maps it to the CLI's current
+auth variable, `LOBEHUB_CLI_API_KEY`. It also sets `LOBEHUB_SERVER` so CLI
+commands hit the local server without needing a stored device-code token.
+
+Use this for automated CLI verification:
+
+```bash
+cd apps/cli
+source ../../.records/env/agent-testing-cli.env
+bun src/index.ts <command>
+```
+
+## CLI — Device Code Flow fallback
+
+Use device-code login only when testing against a non-seeded environment.
+Credentials are isolated from the user's real CLI config via
+`LOBEHUB_CLI_HOME=.lobehub-dev`, which the current CLI stores under
+`$HOME/.lobehub-dev`.
+
+```bash
+cd apps/cli && LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts login --server http://localhost:3010
+```
+
+- The `--server` flag is required — an env var does NOT work and login will hit
+  the wrong server without it.
+- Check state without logging in: `setup-auth.sh status` (verifies
+  `LOBEHUB_CLI_API_KEY` when present, otherwise checks the stored server URL).
+- `UNAUTHORIZED` on API calls means the token expired — re-run login.
+
+## Web — seeded better-auth login
+
+The Web test surface is `agent-browser --session lobehub-dev`. The user's
+ordinary Chrome is only a cookie source; Chrome screenshots, Chrome Network
+records, and Chrome logged-in state do not prove the agent-browser test session
+is authenticated.
+
+For the seeded local dev environment, use the automatic path:
+
+```bash
+./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user
+./.agents/skills/agent-testing/scripts/setup-auth.sh web-seed
+```
+
+`web-seed` posts the seeded email/password to
+`/api/auth/sign-in/email`, stores the returned cookie jar under
+`~/.lobehub-agent-testing/`, converts it to Playwright `storageState`, loads it
+into the `agent-browser` session, and verifies the session does not land on
+`/signin`.
+
+## Web — manual cookie injection fallback
+
+`agent-browser --headed` on macOS often creates the Chromium window off-screen —
+the user can't see or interact with it, so manual login inside the agent-browser
+session fails. Instead, copy the **better-auth session cookie** out of the
+user's own logged-in Chrome and inject it as a Playwright-style state file.
+
+Do **not** use this on production URLs — only local dev. Treat the cookie as a
+secret: don't paste it into shared logs, PRs, or commit it anywhere.
+
+### Web — decision flow
+
+1. `$SCRIPT status --surface web` — green? Start testing. Do not ask for a Cookie header.
+2. Not green and using the seeded local env → `$SCRIPT web-seed`.
+3. Still not green or not using the seed env → `$SCRIPT open-chrome` opens Chrome at `SERVER_URL` with DevTools.
+4. User copies the `Cookie:` header from Network tab → any same-origin request → Request Headers → right-click `Cookie:` → **Copy value**. Must be from Network, NOT `document.cookie` (HttpOnly cookies are invisible to `document.cookie`).
+5. `pbpaste | $SCRIPT web` — filters to better-auth cookies (`session_token`, `session_data`, `state`), builds Playwright `storageState`, loads it into the `agent-browser` session (`lobehub-dev`), opens `SERVER_URL`, and asserts the URL is not `/signin`.
+
+### Using the authenticated session
+
+```bash
+agent-browser --session lobehub-dev open "$SERVER_URL/"
+agent-browser --session lobehub-dev snapshot -i | head -20
+```
+
+### Notes
+
+- `storageState` doesn't enforce the HttpOnly flag on load — the script stores
+  cookies with `httpOnly: false`, which is fine for local dev and sidesteps a
+  CDP-context quirk where HttpOnly cookies sometimes fail to attach.
+- The state file is kept at `~/.lobehub-agent-testing/web-state.json` so
+  `setup-auth.sh status` can report web-auth readiness across sessions.
+
+### Common failure modes
+
+| Symptom                                       | Cause                                                                     | Fix                                                                                            |
+| --------------------------------------------- | ------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- |
+| Still redirects to `/signin` after injection  | User pasted from `document.cookie` → missed HttpOnly session              | Re-pull from Network request Headers, not console                                              |
+| Script reports `no better-auth cookies found` | User pasted the wrong value, or the cookie parser regressed               | Keep the raw `Cookie:` header as-is; run `scripts/setup-auth.test.sh` if the input looks valid |
+| Login works briefly then expires              | `better-auth.session_token` rotated (user logged out / signed in again)   | Re-copy and re-inject                                                                          |
+| Domain mismatch                               | Cookie domain must be `localhost` literally, no leading dot for local dev | —                                                                                              |
+
+## Electron
+
+The desktop app keeps its own persistent login state in its user-data
+directory — log in once manually inside the app and it survives restarts of
+`electron-dev.sh`. No injection needed. The standard check (do NOT hand-roll a
+store eval) once Electron is up with CDP:
+
+```bash
+./.agents/skills/agent-testing/scripts/app-probe.sh auth
+# → {"ok":true,"isSignedIn":true,"userId":"user_xxx"}
+```
+
+`setup-auth.sh status` runs this probe automatically when CDP 9222 is
+reachable.
+
+## Scope
+
+These recipes only cover **local dev** authentication. They do not:
+
+- Work for production — production cookies are `Secure; HttpOnly; Domain=.lobehub.com`
+  and must be delivered over HTTPS.
+- Replace real OAuth flows — tests that must exercise the login UI itself need a
+  real Chromium with `--remote-debugging-port` or a bot account.
+- Flow cookies back to the user's Chrome — injection is one-way.

.agents/skills/agent-testing/references/dev-server.md

@@ -0,0 +1,98 @@
+# Local Dev Server
+
+Single source of truth for starting / restarting the backend that all test
+surfaces (CLI, Electron, Web) hit.
+
+## Resolve ports first
+
+Run `test-env.sh` as described in
+[SKILL.md Step 0.0](../SKILL.md#00-resolve-the-current-test-environment)
+before starting or probing any local test surface.
+
+## Ports & modes
+
+| Command             | What it runs                                              | Port source         |
+| ------------------- | --------------------------------------------------------- | ------------------- |
+| `pnpm run dev:next` | Next.js backend (API + auth)                              | `PORT`              |
+| `bun run dev`       | Full-stack (Next.js + Vite SPA, via `devStartupSequence`) | `PORT` + `SPA_PORT` |
+| `bun run dev:spa`   | Vite SPA only, proxies API to `PORT`                      | `SPA_PORT`          |
+
+In the **cloud repo** (where this repo is the `lobehub/` submodule), local
+worktree names map to fallback defaults only when `.env` and shell env do not
+provide values:
+
+| Workspace directory | Default `SERVER_URL`             |
+| ------------------- | -------------------------------- |
+| `lobehub`           | `http://localhost:3010`          |
+| `lobehub-cloud`     | `http://localhost:3020`          |
+| `lobehub-cloud-1`   | `http://localhost:3021`          |
+| `lobehub-cloud-N`   | `http://localhost:$((3020 + N))` |
+
+`test-env.sh` and `setup-auth.sh` both use the resolved env first and these
+worktree defaults only as fallback. Treat the dev-server terminal output as the
+final source of truth when testing a non-standard port, then export it for every
+agent-testing command:
+
+```bash
+export SERVER_URL=http://localhost:<port-from-dev-output>
+```
+
+## Health check
+
+```bash
+curl -s -o /dev/null -w '%{http_code}' "$SERVER_URL/"
+```
+
+## Start / restart
+
+```bash
+# Start backend only.
+# With root .env: use the existing local config.
+pnpm run dev:next
+
+# Without root .env: use the self-contained agent-testing env.
+./.agents/skills/agent-testing/scripts/init-dev-env.sh dev-next
+
+# Full-stack SPA + backend. Required for Web smoke.
+# With root .env:
+bun run dev
+
+# Without root .env:
+./.agents/skills/agent-testing/scripts/init-dev-env.sh dev
+
+# Local QStash. Run in a separate terminal only when testing workflow paths.
+./.agents/skills/agent-testing/scripts/init-dev-env.sh qstash
+
+# Restart — required to pick up server-side code changes
+lsof -ti:"$PORT" | xargs kill
+pnpm run dev:next
+# or, when no root .env exists:
+# ./.agents/skills/agent-testing/scripts/init-dev-env.sh dev-next
+```
+
+## When a server restart is needed
+
+Next.js hot-reload may not pick up changes in workspace packages — restart when
+in doubt.
+
+| Change location                                 | Restart? |
+| ----------------------------------------------- | -------- |
+| `apps/server/src/` (routers, services, modules) | Yes      |
+| `src/server/` (agent-hono, workflows-hono)      | Yes      |
+| `packages/database/` (models)                   | Yes      |
+| `packages/types/`                               | Yes      |
+| `packages/prompts/`                             | Yes      |
+| `apps/cli/` (CLI runs from source)              | No       |
+
+## Troubleshooting
+
+| Issue                     | Solution                                                                                      |
+| ------------------------- | --------------------------------------------------------------------------------------------- |
+| `ECONNREFUSED`            | Server not running — start it                                                                 |
+| `EADDRINUSE` on the port  | Already running — `lsof -ti:<port> \| xargs kill` first                                       |
+| Stale data / old behavior | Server needs a restart to pick up code changes                                                |
+| QStash workflow failures  | Start `init-dev-env.sh qstash` and make sure dev server inherited the script's `QSTASH_*` env |
+
+Marketplace/community endpoints are not part of the local agent-testing auth
+gate. Do not block local product-chain verification on marketplace API auth
+unless the change explicitly targets marketplace behavior.

.agents/skills/agent-testing/references/record-app-screen.md

@@ -12,13 +12,13 @@ General-purpose screen recording tool for the Electron app. Captures CDP screens
 
 ```bash
 # Start recording (Electron must be running with CDP)
-.agents/skills/local-testing/scripts/record-app-screen.sh start [output_name]
+.agents/skills/agent-testing/scripts/record-app-screen.sh start [output_name]
 
 # Stop recording and assemble video
-.agents/skills/local-testing/scripts/record-app-screen.sh stop
+.agents/skills/agent-testing/scripts/record-app-screen.sh stop
 
 # Check if recording is active
-.agents/skills/local-testing/scripts/record-app-screen.sh status
+.agents/skills/agent-testing/scripts/record-app-screen.sh status
 ```
 
 ### Arguments
@@ -74,10 +74,10 @@ The `.records/` directory is at the project root and is gitignored.
 
 ```bash
 # Start Electron
-.agents/skills/local-testing/scripts/electron-dev.sh start
+.agents/skills/agent-testing/scripts/electron-dev.sh start
 
 # Start recording
-.agents/skills/local-testing/scripts/record-app-screen.sh start my-test
+.agents/skills/agent-testing/scripts/record-app-screen.sh start my-test
 
 # Run automation
 agent-browser --cdp 9222 click @e61
@@ -86,14 +86,14 @@ agent-browser --cdp 9222 press Enter
 sleep 10
 
 # Stop and get results
-.agents/skills/local-testing/scripts/record-app-screen.sh stop
+.agents/skills/agent-testing/scripts/record-app-screen.sh stop
 # → .records/my-test.mp4 + .records/my-test/*.png
 ```
 
 ### Gateway Streaming Demo
 
 ```bash
-.agents/skills/local-testing/scripts/electron-dev.sh start
+.agents/skills/agent-testing/scripts/electron-dev.sh start
 
 # Inject gateway URL
 agent-browser --cdp 9222 eval --stdin << 'EOF'
@@ -106,19 +106,19 @@ agent-browser --cdp 9222 eval --stdin << 'EOF'
 EOF
 
 # Record
-.agents/skills/local-testing/scripts/record-app-screen.sh start gateway-demo
+.agents/skills/agent-testing/scripts/record-app-screen.sh start gateway-demo
 
 # Navigate to agent, send message, wait for completion...
 # (automation commands here)
 
-.agents/skills/local-testing/scripts/record-app-screen.sh stop
+.agents/skills/agent-testing/scripts/record-app-screen.sh stop
 open .records/gateway-demo.mp4
 ```
 
 ### Check Active Recording
 
 ```bash
-.agents/skills/local-testing/scripts/record-app-screen.sh status
+.agents/skills/agent-testing/scripts/record-app-screen.sh status
 # [record] Active recording
 #   Frames:      42 captured (running: yes)
 #   Screenshots: 14 captured (running: yes)

.agents/skills/agent-testing/references/report.md

@@ -0,0 +1,186 @@
+# Structured Test Reports
+
+Every automated test session ends with a structured, evidence-backed report.
+A chat-only summary is not an acceptable deliverable: the report is what the
+user (or a reviewer, or a later agent) audits without replaying the session.
+
+## Location & layout
+
+Reports live under `.records/reports/` (gitignored, like all `.records/`
+output):
+
+```
+.records/reports/<YYYYMMDD-HHMMSS>-<slug>/
+├── report.md      # human-readable report (case table with inline screenshots, verdict)
+├── result.json    # machine-readable results (pass/fail counts, score)
+└── assets/        # evidence: screenshots, HAR files, CLI transcripts
+```
+
+## Workflow
+
+1. **Scaffold up front** — before running the first test step:
+
+   ```bash
+   DIR=$(./.agents/skills/agent-testing/scripts/report-init.sh < slug > "<title>")
+   ```
+
+   The script creates the directory, pre-fills branch / commit / date in both
+   files, and prints the directory path. The scaffold uses the compact report
+   shape below; translate its headings and table labels to the user's language
+   before delivery if needed.
+
+2. **Collect evidence as you test** — every asserted behavior gets one evidence
+   item in `$DIR/assets/`:
+   - UI (static state): `agent-browser screenshot` or `capture-app-window.sh`,
+     then **verify the screenshot with the Read tool before citing it** —
+     never cite an image you haven't looked at.
+
+   - UI (time-based behavior): **screenshot vs GIF is a judgment you must
+     make per case.** If the assertion is about change over time — streaming
+     output, a ticking timer, loading/progress states, animations,
+     appear/disappear transitions — a static screenshot cannot prove it.
+     Record a frame sequence and synthesize a GIF:
+
+     ```bash
+     # start recording (background), trigger the behavior, wait for it to finish
+     ../scripts/record-gif.sh "$DIR/assets/case2-streaming.gif" 12 2 &
+     GIF_PID=$!
+     # ... drive the scenario ...
+     wait $GIF_PID
+     ```
+
+     Embed it like an image: `![case 2](assets/case2-streaming.gif)`. Verify
+     at least the first/last frames visually (Read the GIF) before citing.
+
+   - CLI: exact command + trimmed output (`$CLI task list | tee "$DIR/assets/task-list.txt"`).
+
+   - Network: `agent-browser network requests` dumps or HAR files.
+
+3. **Fill `report.md` as you go** — don't reconstruct from memory at the end.
+   The primary evidence belongs in the case table itself: each row should pair
+   the assertion with the screenshot/GIF or non-visual artifact that proves it,
+   so readers can scan the result without jumping between sections. UI evidence
+   must render inline with Markdown image syntax; a plain link or file path is
+   not acceptable as primary visual evidence.
+
+4. **Set the verdict** in both `report.md` and `result.json`, then link the
+   report directory in your final answer to the user. If UI evidence exists,
+   list the key screenshot/GIF links in the final chat response. Use Markdown
+   link text as the evidence caption, for example:
+   `[Image #1 - observed outcome](<report-dir>/assets/case1.png)`.
+
+## Report language (hard rule)
+
+**`report.md` MUST be written in the language the user is conversing in** —
+the whole file, headings included. If the conversation is in Chinese, the
+report is in Chinese; do not mix English prose into it. The scaffold headings
+are placeholders — translate them when filling if the user is not conversing in
+the scaffold language. Exceptions that stay as-is: code/commands, identifiers,
+log excerpts, and `result.json` (its keys and status values are machine-read
+and stay English; the `title` and case `name` fields follow the user's
+language).
+
+## report.md sections
+
+Default report shape:
+
+| Section          | Content                                                                                      |
+| ---------------- | -------------------------------------------------------------------------------------------- |
+| **Scope**        | What changed / what is being verified; branch, commit, date, surface, entry URL/page, focus  |
+| **Cases**        | Compact table: `# \| Case \| Result \| Key observation \| Evidence`                          |
+| **Verdict**      | Overall verdict first (`pass` / `partial` / `fail`), then the concise reasons and follow-ups |
+| **Verification** | Commands or automated checks run in this session, with trimmed results                       |
+| **Score**        | Pass/fail/blocked counts, optional 0–100 score                                               |
+
+The case table is the main reading surface. Prefer one clear row per user
+scenario or regression assertion, and put the screenshot/GIF directly in the
+`Evidence` cell:
+
+```markdown
+| #   | Case                     | Result | Key observation                                                   | Evidence                                         |
+| --- | ------------------------ | ------ | ----------------------------------------------------------------- | ------------------------------------------------ |
+| 1   | Create a new page        | pass   | Title and body persisted after refresh                            | ![created page](assets/new-page-created.png)     |
+| 2   | Respect requested length | fail   | Requested about 600 Chinese characters; final body was about 1286 | ![final article](assets/write-article-final.png) |
+```
+
+## Inline visual evidence
+
+Screenshots and GIFs must be embedded so the report shows the image inline:
+
+```markdown
+![case 1 result](assets/case1-result.png)
+![streaming response](assets/case2-streaming.gif)
+```
+
+Do **not** use these as the primary evidence for UI cases:
+
+```markdown
+[case 1 result](assets/case1-result.png)
+assets/case1-result.png
+file:///tmp/case1-result.png
+```
+
+Links are acceptable for non-visual artifacts such as CLI transcripts, HAR
+files, or long logs. For videos, embed a representative screenshot/GIF inline in
+the case row and link the full video as supplemental evidence.
+
+Avoid the old wide table with separate `steps`, `expected`, and `actual`
+columns unless the test is purely non-visual and truly needs that breakdown.
+For UI reports, those columns make screenshot-backed reading harder. Put
+procedural detail in the row's key observation only when it changes the
+interpretation of the result.
+
+Use an extra evidence/detail section only when the inline table cannot carry
+the material cleanly, such as long CLI transcripts, HAR summaries, or multiple
+screenshots for one case. In that situation, keep the table evidence cell as an
+inline visual proof for UI cases or a concise link for non-visual artifacts,
+then put the longer material under `Verification` or a brief
+`Additional Evidence` section.
+
+Status values: `pass` / `fail` / `blocked` (couldn't run — e.g. auth or env
+missing; a blocked case is not a pass).
+
+## result.json schema
+
+```json
+{
+  "branch": "feat/task-tree",
+  "cases": [
+    {
+      "id": "1",
+      "name": "task tree returns nested children",
+      "surface": "cli",
+      "status": "pass",
+      "evidence": ["assets/task-tree.txt"]
+    }
+  ],
+  "commit": "abc1234",
+  "createdAt": "2026-06-11T15:30:00+08:00",
+  "summary": {
+    "total": 1,
+    "passed": 1,
+    "failed": 0,
+    "blocked": 0,
+    "score": 100,
+    "verdict": "pass"
+  },
+  "surfaces": ["cli"],
+  "title": "Verify task tree API"
+}
+```
+
+`score` is optional — use it when the verdict has a subjective component (UI
+polish, copy quality); omit it for purely binary runs. `verdict` is the single
+word the user reads first: `pass`, `fail`, or `partial`.
+
+## Rules
+
+- **No evidence, no claim** — every `pass`/`fail` in the case table must link
+  at least one asset. UI cases must inline-embed their primary screenshot/GIF;
+  non-visual CLI/network cases may link transcripts, HAR files, or logs.
+- **Screenshots must be visually verified** with the Read tool before being
+  cited.
+- **Report failures faithfully** — a failing case with clear evidence is a good
+  report; a vague green one is not.
+- If coverage was cut (cases skipped, surfaces not exercised), say so in the
+  Verdict section — silent truncation reads as "covered everything".

.agents/skills/agent-testing/scripts/agent-gateway/analyze-events.ts

@@ -0,0 +1,243 @@
+// Analyzer for probe-events dumps. Reads a JSON file produced by `run.ts dump`
+// and prints a layered breakdown:
+//
+//   1. STREAM EVENTS — every non-chunk WS/SSE event in receipt order
+//   2. CHUNKS SUMMARY — collapsed per-step chunk counts (otherwise floods)
+//   3. ACTION CALLS — replaceMessages / refreshMessages / MARK:* with stack
+//   4. CORRELATION — calls ↔ nearest stream event within ±300ms
+//   5. PER-KEY ASSISTANT GROWTH — for each messagesMap key, when the leading
+//      assistant message's cLen / rLen actually moves (this is what reveals
+//      "chunks arrived but the message never grew" regressions)
+//   6. ROLLBACKS — msgN / childN / role drops in the active-topic timeline
+//
+// Usage:
+//   bun run .agents/skills/agent-testing/scripts/agent-gateway/analyze-events.ts <dump.json>
+
+import { readFileSync } from 'node:fs';
+
+import type {
+  ProbeActionCall,
+  ProbeDump,
+  ProbeMessageSummary,
+  ProbeStreamEvent,
+  ProbeTimelineSample,
+} from './types';
+
+const file = process.argv[2];
+if (!file) {
+  console.error('usage: bun run analyze-events.ts <dump.json>');
+  process.exit(1);
+}
+
+const raw = readFileSync(file, 'utf8');
+// agent-browser eval --stdin wraps return values in quotes when the value is
+// a string — so the JSON file may be double-encoded depending on how it was
+// captured. Handle both.
+const parsedOnce = JSON.parse(raw) as ProbeDump | string;
+const dump: ProbeDump = typeof parsedOnce === 'string' ? JSON.parse(parsedOnce) : parsedOnce;
+
+const { streamEvents = [], actionCalls = [], timeline = [] } = dump;
+
+const pad = (v: unknown, n: number) => String(v).padStart(n);
+
+// ── META ───────────────────────────────────────────────────────────
+console.log('=== META ===');
+console.log(`  events:    ${streamEvents.length}`);
+console.log(`  calls:     ${actionCalls.length}`);
+console.log(`  timeline:  ${timeline.length}`);
+
+// ── 1. STREAM EVENTS (non-chunk) ───────────────────────────────────
+const nonChunkEvents = streamEvents.filter((e) => e.type !== 'stream_chunk');
+const chunkEvents = streamEvents.filter((e) => e.type === 'stream_chunk');
+
+console.log(
+  `\n=== STREAM EVENTS (${nonChunkEvents.length} non-chunk + ${chunkEvents.length} chunks elided) ===`,
+);
+for (const e of nonChunkEvents) {
+  const dataStr = e.dataKeys?.length ? ` [${e.dataKeys.join(',')}]` : '';
+  const data = e.data as Record<string, unknown> | undefined;
+  const uiHint = data?.uiMessagesPreview
+    ? ` uiPreview=${JSON.stringify(data.uiMessagesPreview)}`
+    : data?.uiMessagesTotal
+      ? ` uiTotal=${data.uiMessagesTotal}`
+      : '';
+  const phaseHint = data?.phase ? ` phase=${data.phase}` : '';
+  const extra = e.serverType ? ` serverType=${e.serverType}` : '';
+  console.log(
+    `  t=${pad(e.t, 7)}  [${(e.transport ?? '?').padEnd(3)}]  step=${pad(e.stepIndex ?? '-', 2)}  ` +
+      `type=${(e.type ?? '').padEnd(22)}  op=${e.opIdTail ?? '-'}${phaseHint}${uiHint}${extra}${dataStr}`,
+  );
+}
+
+// ── 2. CHUNK SUMMARY ───────────────────────────────────────────────
+console.log('\n=== CHUNKS SUMMARY (per step / chunkType) ===');
+const chunkBuckets = new Map<string, { count: number; firstT: number; lastT: number }>();
+for (const c of chunkEvents) {
+  const data = c.data as Record<string, unknown> | undefined;
+  const ct = (data?.chunkType as string | undefined) ?? '?';
+  const key = `step=${c.stepIndex ?? '-'}  chunkType=${ct.padEnd(8)}  op=${c.opIdTail}`;
+  const slot = chunkBuckets.get(key);
+  if (slot) {
+    slot.count += 1;
+    slot.lastT = c.t;
+  } else {
+    chunkBuckets.set(key, { count: 1, firstT: c.t, lastT: c.t });
+  }
+}
+for (const [k, v] of chunkBuckets) {
+  console.log(`  ${k}  count=${pad(v.count, 4)}  t=${pad(v.firstT, 7)}..${pad(v.lastT, 7)}`);
+}
+
+// ── 3. ACTION CALLS ───────────────────────────────────────────────
+console.log('\n=== ACTION CALLS (replace/refresh/MARK) ===');
+for (const c of actionCalls) {
+  if (c.name?.startsWith('MARK:')) {
+    console.log(`  t=${pad(c.t, 7)}  ${c.name}`);
+    continue;
+  }
+  const snapshot = (c.args as any)?.snapshot as
+    | Array<{ id: string; role: string; cLen: number; rLen: number }>
+    | undefined;
+  const snapStr = snapshot?.length
+    ? '  snapshot=' + snapshot.map((m) => `${m.id}:${m.role}/c${m.cLen}/r${m.rLen}`).join(' | ')
+    : '';
+  const summary =
+    c.name === 'replaceMessages'
+      ? `count=${c.args?.count} action=${(c.args?.params as any)?.action ?? '-'}${snapStr}`
+      : c.name === 'refreshMessages'
+        ? `ctx=${JSON.stringify(c.args?.context)}`
+        : c.error
+          ? `error=${c.error}`
+          : '';
+  console.log(`  t=${pad(c.t, 7)}  ${c.name.padEnd(20)} ${summary}`);
+  if (c.stack) {
+    const frames = c.stack
+      .split(' ← ')
+      .filter((f) => !!f && !f.includes('Object.<anonymous>'))
+      .slice(0, 3);
+    for (const f of frames) console.log(`             ↳ ${f}`);
+  }
+}
+
+// ── 4. CORRELATION ────────────────────────────────────────────────
+function nearestEventForCall(
+  call: ProbeActionCall,
+  windowMs = 300,
+): { event: ProbeStreamEvent; delta: number } | null {
+  let best: ProbeStreamEvent | null = null;
+  let bestDelta = Infinity;
+  for (const e of streamEvents) {
+    const d = Math.abs(e.t - call.t);
+    if (d < bestDelta && d <= windowMs) {
+      bestDelta = d;
+      best = e;
+    }
+  }
+  return best ? { event: best, delta: bestDelta } : null;
+}
+
+console.log('\n=== CORRELATION (replace/refresh ↔ nearest event within ±300ms) ===');
+for (const c of actionCalls) {
+  if (c.name !== 'refreshMessages' && c.name !== 'replaceMessages') continue;
+  const hit = nearestEventForCall(c);
+  if (hit) {
+    const phase = (hit.event.data as Record<string, unknown> | undefined)?.phase;
+    console.log(
+      `  t=${pad(c.t, 7)}  ${c.name.padEnd(16)} ← Δ${pad(hit.delta, 4)}ms ${hit.event.type}` +
+        (phase ? ` phase=${phase}` : ''),
+    );
+  } else {
+    console.log(`  t=${pad(c.t, 7)}  ${c.name.padEnd(16)} ← (no event nearby — external trigger)`);
+  }
+}
+
+// ── 5. PER-KEY ASSISTANT GROWTH ───────────────────────────────────
+// For each messagesMap key, find the trailing assistant message and report
+// the points in time where its cLen / rLen actually changed. If the timeline
+// shows chunks arriving but the assistant cLen never moves, that's the
+// signature of "dispatch queue blocked / messageId mismatch".
+console.log('\n=== PER-KEY ASSISTANT GROWTH ===');
+const keysEverSeen = new Set<string>();
+for (const s of timeline) for (const k of Object.keys(s.byKey ?? {})) keysEverSeen.add(k);
+
+for (const key of keysEverSeen) {
+  console.log(`\n  key=${key}`);
+  let lastSig: string | null = null;
+  for (const s of timeline) {
+    const slot = s.byKey?.[key];
+    if (!slot) continue;
+    const last = slot.msgs.at(-1) as ProbeMessageSummary | undefined;
+    if (!last) continue;
+    const sig = `${last.id}|c${last.cLen}|r${last.rLen}|n${slot.n}`;
+    if (sig === lastSig) continue;
+    lastSig = sig;
+    console.log(
+      `    t=${pad(s.t, 7)}  msgN=${pad(slot.n, 3)}  ` +
+        `lastAssistant=${last.id}  cLen=${pad(last.cLen, 5)}  rLen=${pad(last.rLen, 5)}` +
+        `  runOps=${s.runOps}`,
+    );
+  }
+}
+
+// ── 6. ROLLBACKS (active-topic msgN / childN / role drops) ─────────
+console.log('\n=== ROLLBACKS (active-topic msgN / childN / role drops) ===');
+let prev: ProbeTimelineSample | null = null;
+const rollbacks: Array<{ t: number; topic: string | null; drops: string[] }> = [];
+
+const flatten = (s: ProbeTimelineSample) => {
+  if (!s.activeTopic) return [];
+  return Object.entries(s.byKey ?? {})
+    .filter(([k]) => k.includes(s.activeTopic!))
+    .flatMap(([, v]) => v.msgs);
+};
+
+for (const s of timeline) {
+  if (s.err) {
+    prev = null;
+    continue;
+  }
+  if (!prev || prev.activeTopic !== s.activeTopic) {
+    prev = s;
+    continue;
+  }
+  const prevMsgs = flatten(prev);
+  const curMsgs = flatten(s);
+  const drops: string[] = [];
+
+  if (curMsgs.length < prevMsgs.length) drops.push(`msgN ${prevMsgs.length}→${curMsgs.length}`);
+
+  let prevChild = 0;
+  let curChild = 0;
+  for (const m of prevMsgs) prevChild += m.chN ?? 0;
+  for (const m of curMsgs) curChild += m.chN ?? 0;
+  if (curChild < prevChild) drops.push(`childN ${prevChild}→${curChild}`);
+
+  const prevById = new Map(prevMsgs.map((m) => [m.id, m]));
+  for (const m of curMsgs) {
+    const pr = prevById.get(m.id);
+    if (!pr) continue;
+    if (m.cLen < pr.cLen) drops.push(`cLen[${m.id}] ${pr.cLen}→${m.cLen}`);
+    if (m.rLen < pr.rLen) drops.push(`rLen[${m.id}] ${pr.rLen}→${m.rLen}`);
+  }
+
+  if (drops.length) rollbacks.push({ t: s.t, topic: s.activeTopic, drops });
+  prev = s;
+}
+
+if (rollbacks.length === 0) {
+  console.log('  (none)');
+} else {
+  for (const r of rollbacks) {
+    const nearEvent = streamEvents
+      .filter((e) => Math.abs(e.t - r.t) <= 300)
+      .map((e) => `${e.type}${(e.data as any)?.phase ? ':' + (e.data as any).phase : ''}`);
+    const nearCall = actionCalls
+      .filter((c) => Math.abs(c.t - r.t) <= 300 && !c.name?.startsWith('MARK:'))
+      .map((c) => c.name);
+    console.log(
+      `  t=${pad(r.t, 7)}  topic=${r.topic}  ${r.drops.join(' | ')}` +
+        (nearEvent.length ? `  near-event:[${nearEvent.join(',')}]` : '') +
+        (nearCall.length ? `  near-call:[${nearCall.join(',')}]` : ''),
+    );
+  }
+}

.agents/skills/agent-testing/scripts/agent-gateway/analyze.mjs

@@ -0,0 +1,119 @@
+#!/usr/bin/env node
+// Analyze a probe dump captured by probe.js + probe-dump.js.
+//
+//   node analyze.mjs /tmp/probe.json
+//
+// Prints:
+//   1. EVENTS — user-action markers with their relative timestamps
+//   2. TIMELINE — periodic samples (~1 per second + event-adjacent samples)
+//      showing every interesting field; columns:
+//        t(ms) | runOps | msgN | childN | content | reasoning | tools | domLen | search | crawl | topic | event
+//   3. REGRESSIONS — every place a tracked counter *dropped* on the same
+//      topic between adjacent samples. A "true" UI rollback shows up as a
+//      drop in content/reasoning/tools/childN/domLen without a topic change.
+//
+// Whitelisted transitions (not flagged):
+//   - topic change → all drops expected (focus moved away)
+//   - reasoning length 0 after content starts → reasoning gets sealed into a
+//     completed sub-block; the parent's running reasoning resets to ''.
+//   - msgN drop when topic transitions from `_new` placeholder to a real id.
+
+import fs from 'node:fs';
+
+const file = process.argv[2];
+if (!file) {
+  console.error('usage: node analyze.mjs <probe.json>');
+  process.exit(1);
+}
+
+const raw = JSON.parse(fs.readFileSync(file, 'utf8'));
+// probe-dump.js wraps the payload in JSON.stringify so agent-browser returns
+// it as a single quoted string. Unwrap.
+const data = typeof raw === 'string' ? JSON.parse(raw) : raw;
+const { events, samples } = data;
+
+const fmt = {
+  pad(v, n) {
+    return String(v).padStart(n);
+  },
+};
+
+console.log('=== EVENTS ===');
+for (const e of events) console.log(`  t=${fmt.pad(e.t, 7)}  ${e.name}`);
+
+console.log(
+  '\n=== TIMELINE (~1s cadence, plus event-adjacent samples) ===\n' +
+    '  t(ms)   runOps  msgN childN  content reasoning tools  domLen  search crawl  topic     event',
+);
+
+let lastSampledAt = -1e9;
+const eventBuckets = events.map((e) => e.t);
+for (let i = 0; i < samples.length; i++) {
+  const s = samples[i];
+  const nearEvent = eventBuckets.some((et) => Math.abs(et - s.t) < 110);
+  if (!nearEvent && s.t - lastSampledAt < 1000) continue;
+  lastSampledAt = s.t;
+
+  const ev = events.find((e) => Math.abs(e.t - s.t) < 110);
+  const evMarker = ev ? `  ◀ ${ev.name}` : '';
+  const topicSuffix = s.topicId ? s.topicId.slice(-6) : '(none)';
+  const search = s.ind?.search ?? 0;
+  const crawl = s.ind?.crawl ?? 0;
+  console.log(
+    `  ${fmt.pad(s.t, 6)} ` +
+      `${fmt.pad(s.runOps, 6)}  ` +
+      `${fmt.pad(s.msgN, 4)}  ` +
+      `${fmt.pad(s.childN ?? 0, 5)} ` +
+      `${fmt.pad(s.cT ?? 0, 8)} ` +
+      `${fmt.pad(s.rT ?? 0, 9)} ` +
+      `${fmt.pad(s.toolT ?? 0, 5)} ` +
+      `${fmt.pad(s.domLen ?? 0, 7)} ` +
+      `${fmt.pad(search, 6)} ` +
+      `${fmt.pad(crawl, 5)}  ` +
+      `${topicSuffix.padEnd(8)}${evMarker}`,
+  );
+}
+
+console.log('\n=== REGRESSIONS (same topic, value dropped) ===');
+const regressions = [];
+for (let i = 1; i < samples.length; i++) {
+  const prev = samples[i - 1];
+  const cur = samples[i];
+  if (!cur.topicId || prev.topicId !== cur.topicId) continue;
+
+  const drops = [];
+  if (cur.msgN < prev.msgN) drops.push(`msgN: ${prev.msgN}→${cur.msgN}`);
+  if ((cur.childN ?? 0) < (prev.childN ?? 0)) drops.push(`childN: ${prev.childN}→${cur.childN}`);
+  if ((cur.cT ?? 0) < (prev.cT ?? 0)) drops.push(`content: ${prev.cT}→${cur.cT}`);
+  if ((cur.rT ?? 0) < (prev.rT ?? 0)) drops.push(`reasoning: ${prev.rT}→${cur.rT}`);
+  if ((cur.toolT ?? 0) < (prev.toolT ?? 0)) drops.push(`tools: ${prev.toolT}→${cur.toolT}`);
+  // domLen jitters by a few chars from counter labels — only flag big drops.
+  if ((cur.domLen ?? 0) < (prev.domLen ?? 0) - 100) {
+    drops.push(`domLen: ${prev.domLen}→${cur.domLen}`);
+  }
+  if (drops.length === 0) continue;
+
+  const nearbyEv = events.filter((e) => Math.abs(e.t - cur.t) < 600).map((e) => e.name);
+  regressions.push({ t: cur.t, topic: cur.topicId.slice(-6), drops, nearbyEv });
+}
+
+if (regressions.length === 0) {
+  console.log('  (none)');
+} else {
+  for (const r of regressions) {
+    const evStr = r.nearbyEv.length ? `  near:[${r.nearbyEv.join(',')}]` : '';
+    console.log(`  t=${fmt.pad(r.t, 7)}  topic=${r.topic}  ${r.drops.join(' | ')}${evStr}`);
+  }
+}
+
+console.log(`\n=== SUMMARY ===`);
+console.log(`  samples: ${samples.length}`);
+console.log(`  events:  ${events.length}`);
+console.log(`  regressions: ${regressions.length}`);
+if (samples.length) {
+  const last = samples.at(-1);
+  console.log(
+    `  final: msgN=${last.msgN} childN=${last.childN ?? 0} content=${last.cT ?? 0} ` +
+      `reasoning=${last.rT ?? 0} tools=${last.toolT ?? 0} runOps=${last.runOps}`,
+  );
+}

.agents/skills/agent-testing/scripts/agent-gateway/probe-dump.js

@@ -0,0 +1,17 @@
+// Stop the probe and serialize collected data.
+//
+//   agent-browser --cdp 9222 eval --stdin < probe-dump.js > /tmp/probe.json
+//
+// The whole thing is wrapped in a JSON.stringify so agent-browser returns it
+// as a single quoted string — the analyzer double-parses to handle that.
+
+(function () {
+  if (window.__PROBE_TIMER) {
+    clearInterval(window.__PROBE_TIMER);
+    window.__PROBE_TIMER = null;
+  }
+  return JSON.stringify({
+    events: window.__PROBE_EVENTS || [],
+    samples: window.__PROBE_SAMPLES || [],
+  });
+})();

.agents/skills/agent-testing/scripts/agent-gateway/probe-dump.ts

@@ -0,0 +1,37 @@
+// Stops the events-probe timeline timer and stashes the full capture as a
+// JSON string on `window.__PROBE_LAST_DUMP_JSON`. `run.ts` wraps the bundle
+// in an IIFE that returns that global, which `agent-browser eval` prints to
+// stdout — the runner then persists it under `.agent-gateway/`.
+
+import type { ProbeDump } from './types';
+
+declare global {
+  interface Window {
+    __PROBE_LAST_DUMP_JSON?: string;
+  }
+}
+
+const w = window;
+
+if (w.__PROBE_TIMELINE_TIMER) {
+  clearInterval(w.__PROBE_TIMELINE_TIMER);
+  w.__PROBE_TIMELINE_TIMER = null;
+}
+
+const mutations = w.__PROBE_MUTATIONS ?? [];
+
+const dump: ProbeDump & { mutations: typeof mutations } = {
+  meta: {
+    t0: w.__PROBE_T0 ?? 0,
+    collectedAt: Date.now(),
+    sampleCount: (w.__PROBE_MSG_TIMELINE ?? []).length,
+    eventCount: (w.__PROBE_STREAM_EVENTS ?? []).length,
+    callCount: (w.__PROBE_ACTION_CALLS ?? []).length,
+  },
+  streamEvents: w.__PROBE_STREAM_EVENTS ?? [],
+  actionCalls: w.__PROBE_ACTION_CALLS ?? [],
+  timeline: w.__PROBE_MSG_TIMELINE ?? [],
+  mutations,
+};
+
+w.__PROBE_LAST_DUMP_JSON = JSON.stringify(dump);

.agents/skills/agent-testing/scripts/agent-gateway/probe-events.ts

@@ -0,0 +1,637 @@
+// LobeHub gateway raw-event-stream probe.
+//
+// Gateway-mode chats subscribe via WebSocket — NOT via the `/api/agent/stream`
+// SSE endpoint (that one belongs to the direct/client durable-agent runtime).
+// `AgentStreamClient` (`packages/agent-gateway-client/src/client.ts`) opens
+// `new WebSocket('wss://.../ws?operationId=...')`, then parses JSON frames in
+// its `onmessage` handler and re-emits `agent_event.event` objects to the
+// chat store.
+//
+// To capture the RAW gateway events before the store touches them, we wrap
+// `window.WebSocket` so that for any socket whose URL contains `operationId=`
+// we intercept the `onmessage` handler / `addEventListener('message')` and
+// log every `agent_event` frame.
+//
+// We *also* keep the `window.fetch` hook for `/api/agent/stream` so this
+// probe still works for direct-mode runs — but gateway-mode events come
+// through the WebSocket path.
+//
+// Buffers (read via `dump`):
+//   __PROBE_STREAM_EVENTS  — raw events parsed off the wire
+//   __PROBE_ACTION_CALLS   — replaceMessages / refreshMessages calls (best-effort)
+//   __PROBE_MSG_TIMELINE   — 200ms snapshots of every messagesMap key
+
+import type {
+  ProbeActionCall,
+  ProbeMessageSummary,
+  ProbeStreamEvent,
+  ProbeTimelineSample,
+} from './types';
+
+// Bundled by esbuild as an IIFE. Top-level code runs once on injection.
+
+const w = window;
+
+// ── Buffers ─────────────────────────────────────────────────────────
+
+declare global {
+  interface Window {
+    __PROBE_MUTATIONS?: Array<{
+      t: number;
+      key: string;
+      n: number;
+      last?: { id: string; role: string; cLen: number; rLen: number; updatedAt?: unknown };
+      prevLast?: { id: string; role: string; cLen: number; rLen: number };
+      delta?: string;
+    }>;
+    __PROBE_STORE_UNSUB?: () => void;
+  }
+}
+
+const events: ProbeStreamEvent[] = (w.__PROBE_STREAM_EVENTS ??= []);
+const calls: ProbeActionCall[] = (w.__PROBE_ACTION_CALLS ??= []);
+const timeline: ProbeTimelineSample[] = (w.__PROBE_MSG_TIMELINE ??= []);
+const mutations = (w.__PROBE_MUTATIONS ??= []);
+events.length = 0;
+calls.length = 0;
+timeline.length = 0;
+mutations.length = 0;
+
+const t0 = Date.now();
+w.__PROBE_T0 = t0;
+const now = (): number => Date.now() - t0;
+
+// ── Helpers ─────────────────────────────────────────────────────────
+
+function summarizeData(data: unknown): Record<string, unknown> | unknown {
+  if (!data || typeof data !== 'object') return data;
+  const src = data as Record<string, unknown>;
+  const out: Record<string, unknown> = {};
+  for (const k of Object.keys(src)) {
+    const v = src[k];
+    if (v == null) {
+      out[k] = v;
+    } else if (Array.isArray(v)) {
+      out[k] = `Array(${v.length})`;
+      if (k === 'uiMessages') {
+        out.uiMessagesPreview = v.slice(0, 5).map((m: any) => ({
+          id: (m.id ?? '').slice(-8),
+          role: m.role,
+          cLen: (m.content ?? '').length,
+          children: (m.children ?? []).length,
+          tools: (m.tools ?? []).length,
+          reasoning: (m.reasoning?.content ?? '').length,
+        }));
+        out.uiMessagesTotal = v.length;
+      }
+    } else if (typeof v === 'object') {
+      const obj = v as Record<string, unknown>;
+      out[k] =
+        'Object{' +
+        Object.keys(obj)
+          .slice(0, 6)
+          .map((kk) => kk + (typeof obj[kk] === 'string' ? `=${(obj[kk] as string).length}ch` : ''))
+          .join(',') +
+        '}';
+    } else if (typeof v === 'string') {
+      out[k] = v.length > 100 ? v.slice(0, 100) + `…(${v.length})` : v;
+    } else {
+      out[k] = v;
+    }
+  }
+  return out;
+}
+
+function summarizeMessages(msgs: any[]): ProbeMessageSummary[] {
+  return (msgs ?? []).slice(0, 80).map((m) => ({
+    id: (m.id ?? '').slice(-8),
+    role: m.role,
+    cLen: (m.content ?? '').length,
+    rLen: (m.reasoning?.content ?? '').length,
+    tools: (m.tools ?? []).length,
+    chN: (m.children ?? []).length,
+  }));
+}
+
+function shortStack(): string {
+  const raw = new Error('probe-stack').stack ?? '';
+  return raw
+    .split('\n')
+    .slice(3)
+    .filter((l) => !l.includes('probe-events') && !l.includes('node_modules'))
+    .map((l) => l.trim().replace(/^at\s+/, ''))
+    .slice(0, 6)
+    .join(' ← ');
+}
+
+function recordAgentEvent(args: {
+  transport: 'ws' | 'sse';
+  opId: string | null;
+  agentEvent: any;
+  eventId?: string | null;
+  rawLen?: number;
+}): void {
+  const { transport, opId, agentEvent, eventId, rawLen } = args;
+  if (!agentEvent || typeof agentEvent !== 'object') return;
+  events.push({
+    t: now(),
+    transport,
+    opIdTail: (opId ?? '').slice(-10),
+    eventId: eventId ?? null,
+    type: agentEvent.type,
+    stepIndex: agentEvent.stepIndex,
+    dataKeys: agentEvent.data ? Object.keys(agentEvent.data) : [],
+    data: summarizeData(agentEvent.data) as Record<string, unknown>,
+    rawLen,
+  });
+}
+
+// ── 1. Patch window.WebSocket for gateway WS events ────────────────
+
+if (!w.__PROBE_ORIG_WEBSOCKET) w.__PROBE_ORIG_WEBSOCKET = w.WebSocket;
+const OrigWS = w.__PROBE_ORIG_WEBSOCKET;
+
+function extractOpIdFromWsUrl(url: string | URL): string | null {
+  const m = String(url ?? '').match(/operationId=([^&]+)/);
+  return m ? decodeURIComponent(m[1]) : null;
+}
+
+function isGatewayWs(url: string | URL): boolean {
+  return String(url ?? '').includes('operationId=');
+}
+
+function handleWsFrame(rawData: unknown, opId: string | null): void {
+  const rawLen = typeof rawData === 'string' ? rawData.length : -1;
+  let parsed: any;
+  try {
+    parsed = typeof rawData === 'string' ? JSON.parse(rawData) : null;
+  } catch {
+    events.push({
+      t: now(),
+      transport: 'ws',
+      opIdTail: (opId ?? '').slice(-10),
+      type: '_PARSE_ERROR_',
+      raw: typeof rawData === 'string' && rawData.length < 400 ? rawData : '(non-string or large)',
+    });
+    return;
+  }
+  if (!parsed) return;
+
+  if (parsed.type === 'agent_event') {
+    recordAgentEvent({
+      transport: 'ws',
+      opId,
+      agentEvent: parsed.event,
+      eventId: parsed.id,
+      rawLen,
+    });
+  } else {
+    events.push({
+      t: now(),
+      transport: 'ws',
+      opIdTail: (opId ?? '').slice(-10),
+      type: '_SERVER_MSG_',
+      serverType: parsed.type,
+      rawLen,
+    });
+  }
+}
+
+// Wrap the constructor. Instance `constructor` will still reflect OrigWS
+// (we share prototypes), so use the `_WS_OPEN_` sentinel events to confirm
+// the patch is firing.
+function PatchedWebSocket(this: WebSocket, url: string | URL, protocols?: string | string[]) {
+  const ws: WebSocket = protocols == null ? new OrigWS(url) : new OrigWS(url, protocols);
+  const opId = extractOpIdFromWsUrl(url);
+  if (!isGatewayWs(url)) return ws;
+
+  events.push({
+    t: now(),
+    transport: 'ws',
+    opIdTail: (opId ?? '').slice(-10),
+    type: '_WS_OPEN_',
+    url: String(url),
+  });
+
+  // One observer listener that always fires, regardless of how the consumer
+  // (AgentStreamClient uses `ws.onmessage = …`) subscribes.
+  ws.addEventListener('message', (e) => {
+    try {
+      handleWsFrame((e as MessageEvent).data, opId);
+    } catch {
+      /* swallow */
+    }
+  });
+
+  ws.addEventListener('close', () => {
+    events.push({
+      t: now(),
+      transport: 'ws',
+      opIdTail: (opId ?? '').slice(-10),
+      type: '_WS_CLOSE_',
+    });
+  });
+
+  return ws;
+}
+
+// Preserve prototype + static fields so `instanceof WebSocket` and
+// `WebSocket.OPEN` constants still work.
+(PatchedWebSocket as unknown as { prototype: WebSocket }).prototype = OrigWS.prototype;
+for (const k of Object.keys(OrigWS) as Array<keyof typeof OrigWS>) {
+  try {
+    (PatchedWebSocket as any)[k] = (OrigWS as any)[k];
+  } catch {
+    /* readonly */
+  }
+}
+(['CONNECTING', 'OPEN', 'CLOSING', 'CLOSED'] as const).forEach((k) => {
+  (PatchedWebSocket as any)[k] = (OrigWS as any)[k];
+});
+w.WebSocket = PatchedWebSocket as unknown as typeof WebSocket;
+
+// ── 2. Patch window.fetch for `/api/agent/stream` (direct-mode SSE) ─
+
+if (!w.__PROBE_ORIG_FETCH) w.__PROBE_ORIG_FETCH = w.fetch.bind(w);
+const origFetch = w.__PROBE_ORIG_FETCH;
+
+function isAgentStreamUrl(input: RequestInfo | URL): boolean {
+  let url = '';
+  if (typeof input === 'string') url = input;
+  else if (input instanceof URL) url = input.toString();
+  else if (input && typeof (input as Request).url === 'string') url = (input as Request).url;
+  return url.includes('/api/agent/stream');
+}
+
+function extractOpIdFromHttpUrl(input: RequestInfo | URL): string | null {
+  const url = typeof input === 'string' ? input : (input as Request | URL).toString();
+  const m = url.match(/operationId=([^&]+)/);
+  return m ? decodeURIComponent(m[1]) : null;
+}
+
+function pushFromSSEFrame(rawFrame: string, opId: string | null): void {
+  const lines = rawFrame.split('\n');
+  let dataJson = '';
+  let evtName = 'message';
+  for (const line of lines) {
+    if (line.startsWith('event:')) evtName = line.slice(6).trim();
+    else if (line.startsWith('data:')) dataJson += line.slice(5).trim();
+  }
+  if (!dataJson) return;
+  let parsed: any;
+  try {
+    parsed = JSON.parse(dataJson);
+  } catch {
+    events.push({
+      t: now(),
+      transport: 'sse',
+      opIdTail: (opId ?? '').slice(-10),
+      type: '_PARSE_ERROR_',
+      sseEvent: evtName,
+      raw: dataJson.length > 400 ? dataJson.slice(0, 400) + '…' : dataJson,
+    });
+    return;
+  }
+  recordAgentEvent({
+    transport: 'sse',
+    opId,
+    agentEvent: parsed,
+    eventId: null,
+    rawLen: dataJson.length,
+  });
+}
+
+async function teeAndDrain(response: Response, opId: string | null): Promise<Response> {
+  if (!response.body) return response;
+  const [a, b] = response.body.tee();
+
+  void (async () => {
+    const reader = b.getReader();
+    const decoder = new TextDecoder();
+    let buf = '';
+    try {
+      while (true) {
+        const { value, done } = await reader.read();
+        if (done) break;
+        buf += decoder.decode(value, { stream: true });
+        let idx: number;
+
+        while ((idx = buf.indexOf('\n\n')) !== -1) {
+          const frame = buf.slice(0, idx);
+          buf = buf.slice(idx + 2);
+          if (frame.trim()) pushFromSSEFrame(frame, opId);
+        }
+      }
+      if (buf.trim()) pushFromSSEFrame(buf, opId);
+    } catch (e: any) {
+      events.push({
+        t: now(),
+        transport: 'sse',
+        opIdTail: (opId ?? '').slice(-10),
+        type: '_TEE_ERROR_',
+        message: String(e?.message ?? e),
+      });
+    }
+  })();
+
+  return new Response(a, {
+    headers: response.headers,
+    status: response.status,
+    statusText: response.statusText,
+  });
+}
+
+w.fetch = async function patchedFetch(input: RequestInfo | URL, init?: RequestInit) {
+  const response = await origFetch(input as any, init);
+  if (!isAgentStreamUrl(input)) return response;
+  const opId = extractOpIdFromHttpUrl(input);
+  const url =
+    typeof input === 'string'
+      ? input.split('?')[0]
+      : (input as Request | URL).toString().split('?')[0];
+  events.push({
+    t: now(),
+    transport: 'sse',
+    opIdTail: (opId ?? '').slice(-10),
+    type: '_CONNECTED_',
+    url,
+    status: response.status,
+  });
+  return teeAndDrain(response, opId);
+} as typeof fetch;
+
+// ── 3. Wrap store actions (best-effort for "who called replace") ────
+
+// Side-global stash for the original chat-store actions. Re-installs ALWAYS
+// rewrap from the originals so updates to the probe body take effect
+// without a page reload — using only a `__probeWrapped` flag on the chat
+// state object would freeze the first-installed wrapper across re-installs.
+declare global {
+  interface Window {
+    __PROBE_ORIG_REFRESH_MESSAGES?: any;
+    __PROBE_ORIG_REPLACE_MESSAGES?: any;
+  }
+}
+
+try {
+  const chat = w.__LOBE_STORES?.chat?.();
+  if (chat) {
+    // First-time install: cache the originals. Re-install: restore from
+    // the cached originals before wrapping again.
+    if (!w.__PROBE_ORIG_REFRESH_MESSAGES) w.__PROBE_ORIG_REFRESH_MESSAGES = chat.refreshMessages;
+    if (!w.__PROBE_ORIG_REPLACE_MESSAGES) w.__PROBE_ORIG_REPLACE_MESSAGES = chat.replaceMessages;
+    const origRefresh = w.__PROBE_ORIG_REFRESH_MESSAGES;
+    const origReplace = w.__PROBE_ORIG_REPLACE_MESSAGES;
+    chat.refreshMessages = origRefresh;
+    chat.replaceMessages = origReplace;
+
+    chat.refreshMessages = async function probeRefresh(this: unknown, ...args: any[]) {
+      calls.push({
+        t: now(),
+        name: 'refreshMessages',
+        args: { context: args[0] ?? null },
+        stack: shortStack(),
+      });
+      return origRefresh.apply(this, args);
+    };
+    chat.replaceMessages = function probeReplace(this: unknown, ...args: any[]) {
+      const msgs = (args[0] as any[]) ?? [];
+      const snapshot = msgs.slice(-2).map((m) => ({
+        id: (m.id ?? '').slice(-8),
+        role: m.role,
+        cLen: (m.content ?? '').length,
+        rLen: (m.reasoning?.content ?? '').length,
+        updatedAt: m.updatedAt,
+      }));
+      calls.push({
+        t: now(),
+        name: 'replaceMessages',
+        args: { count: msgs.length, params: args[1] ?? null, snapshot } as any,
+        stack: shortStack(),
+      });
+
+      // Pair the call with a mutation row so the analyzer can build a
+      // single ordered timeline across replaceMessages + dispatchMessage.
+      const stackTop = shortStack().split(' ← ')[0]?.slice(0, 80);
+      const last = msgs.at(-1);
+      const lastSum = last
+        ? {
+            id: (last.id ?? '').slice(-8),
+            role: last.role,
+            cLen: (last.content ?? '').length,
+            rLen: (last.reasoning?.content ?? '').length,
+            updatedAt: last.updatedAt,
+          }
+        : undefined;
+      const params: any = args[1] ?? {};
+      const ctxKey = params.context
+        ? `main_${params.context.agentId ?? '?'}_${
+            params.context.topicId ? 'tpc_' + params.context.topicId : 'new'
+          }`.replace('main_tpc_', 'main_') // crude key inference
+        : '(no-ctx)';
+      mutations.push({
+        t: now(),
+        key: ctxKey,
+        n: msgs.length,
+        last: lastSum,
+        delta: `replaceMessages(action=${params.action ?? '-'})  src=${stackTop ?? '-'}`,
+      });
+
+      return origReplace.apply(this, args);
+    };
+  }
+} catch (e: any) {
+  calls.push({ t: now(), name: '_WRAP_ERROR_', error: String(e?.message ?? e) });
+}
+
+// ── 3.5. Mutation log — wrap the TWO ChatStore writers (replaceMessages,
+// internal_dispatchMessage) to record EVERY dbMessagesMap[key] reference
+// change with a one-line "before/after last assistant message" delta. This
+// reveals dispatchMessage-driven collapses that the replaceMessages wrap
+// alone cannot see.
+
+declare global {
+  interface Window {
+    __PROBE_ORIG_DISPATCH_MESSAGE?: any;
+  }
+}
+
+try {
+  const chat = w.__LOBE_STORES?.chat?.();
+  if (chat?.internal_dispatchMessage) {
+    if (!w.__PROBE_ORIG_DISPATCH_MESSAGE)
+      w.__PROBE_ORIG_DISPATCH_MESSAGE = chat.internal_dispatchMessage;
+    const origDispatch = w.__PROBE_ORIG_DISPATCH_MESSAGE;
+    chat.internal_dispatchMessage = origDispatch;
+
+    chat.internal_dispatchMessage = function probeDispatch(this: unknown, payload: any, ctx?: any) {
+      // Snapshot BEFORE — read the would-be target key + last message.
+      const before = (() => {
+        try {
+          const state = w.__LOBE_STORES?.chat?.();
+          if (!state) return null;
+          // Replicate state.internal_getConversationContext logic enough to
+          // resolve a key — but most callers pass operationId on ctx, and
+          // operationId-keyed lookup needs store internals. Easiest: snapshot
+          // ALL keys' last-assistant cLen and compare BEFORE vs AFTER below.
+          const map = state.dbMessagesMap ?? {};
+          const out: Record<string, any> = {};
+          for (const k of Object.keys(map)) {
+            const last = (map[k] ?? []).at(-1);
+            out[k] = last
+              ? {
+                  id: (last.id ?? '').slice(-8),
+                  cLen: (last.content ?? '').length,
+                  rLen: (last.reasoning?.content ?? '').length,
+                  n: map[k].length,
+                }
+              : { n: 0 };
+          }
+          return out;
+        } catch {
+          return null;
+        }
+      })();
+
+      const result = origDispatch.apply(this, [payload, ctx]);
+
+      // Snapshot AFTER — find which key(s) actually changed.
+      try {
+        const state = w.__LOBE_STORES?.chat?.();
+        if (state && before) {
+          const map = state.dbMessagesMap ?? {};
+          for (const k of Object.keys(map)) {
+            const last = (map[k] ?? []).at(-1);
+            const beforeSnap = before[k];
+            const afterSnap = last
+              ? {
+                  id: (last.id ?? '').slice(-8),
+                  cLen: (last.content ?? '').length,
+                  rLen: (last.reasoning?.content ?? '').length,
+                  n: map[k].length,
+                }
+              : { n: 0 };
+            const changed =
+              !beforeSnap ||
+              beforeSnap.n !== afterSnap.n ||
+              beforeSnap.id !== (afterSnap as any).id ||
+              beforeSnap.cLen !== (afterSnap as any).cLen ||
+              beforeSnap.rLen !== (afterSnap as any).rLen;
+            if (!changed) continue;
+            let delta = '';
+            if (beforeSnap?.id !== undefined && beforeSnap.id !== (afterSnap as any).id)
+              delta += `id:${beforeSnap.id}→${(afterSnap as any).id};`;
+            if (
+              beforeSnap?.cLen !== undefined &&
+              (afterSnap as any).cLen !== undefined &&
+              (afterSnap as any).cLen < beforeSnap.cLen
+            )
+              delta += `cLen↓${beforeSnap.cLen}→${(afterSnap as any).cLen};`;
+            if (
+              beforeSnap?.rLen !== undefined &&
+              (afterSnap as any).rLen !== undefined &&
+              (afterSnap as any).rLen < beforeSnap.rLen
+            )
+              delta += `rLen↓${beforeSnap.rLen}→${(afterSnap as any).rLen};`;
+            if (beforeSnap?.n !== undefined && afterSnap.n < beforeSnap.n)
+              delta += `n↓${beforeSnap.n}→${afterSnap.n};`;
+            mutations.push({
+              t: now(),
+              key: k,
+              n: afterSnap.n,
+              last: (afterSnap as any).id ? (afterSnap as any) : undefined,
+              prevLast: beforeSnap?.id ? beforeSnap : undefined,
+              delta: delta || `dispatch:${payload?.type}`,
+            });
+          }
+        }
+      } catch (e: any) {
+        mutations.push({
+          t: now(),
+          key: '_DISPATCH_PROBE_ERROR_',
+          n: -1,
+          delta: String(e?.message ?? e),
+        });
+      }
+      return result;
+    };
+  }
+} catch (e: any) {
+  calls.push({ t: now(), name: '_DISPATCH_WRAP_ERROR_', error: String(e?.message ?? e) });
+}
+
+// ── 4. Periodic per-key timeline snapshots ─────────────────────────
+
+function captureTimeline(): void {
+  try {
+    const c = w.__LOBE_STORES?.chat?.();
+    if (!c) return;
+    const msgsMap = (c.messagesMap ?? {}) as Record<string, any[]>;
+    const dbMap = (c.dbMessagesMap ?? {}) as Record<string, any[]>;
+    const byKey: ProbeTimelineSample['byKey'] = {};
+    for (const k of Object.keys(msgsMap)) {
+      const display = msgsMap[k] ?? [];
+      const db = dbMap[k] ?? [];
+      if (display.length === 0 && db.length === 0) continue;
+      byKey[k] = {
+        n: display.length,
+        dbN: db.length,
+        msgs: summarizeMessages(display),
+      };
+    }
+    const ops = Object.values((c.operations ?? {}) as Record<string, any>);
+    timeline.push({
+      t: now(),
+      activeTopic: ((c.activeTopicId as string | null) ?? '').slice(-10) || null,
+      keys: Object.keys(byKey),
+      byKey,
+      runOps: ops.filter((o: any) => o.status === 'running').length,
+    });
+  } catch (e: any) {
+    timeline.push({
+      t: now(),
+      activeTopic: null,
+      keys: [],
+      byKey: {},
+      runOps: 0,
+      err: e?.message ?? String(e),
+    });
+  }
+}
+captureTimeline();
+if (w.__PROBE_TIMELINE_TIMER) clearInterval(w.__PROBE_TIMELINE_TIMER);
+w.__PROBE_TIMELINE_TIMER = setInterval(captureTimeline, 200);
+
+// ── 5. Tab-switch helpers ──────────────────────────────────────────
+
+function listTopBarTabs(): HTMLElement[] {
+  return Array.from(
+    document.querySelectorAll<HTMLElement>(
+      '[data-insp-path*="TabItem.tsx"][data-contextmenu-trigger]',
+    ),
+  ).filter((t) => t.getBoundingClientRect().top < 30);
+}
+
+w.__listTabs = () =>
+  listTopBarTabs().map((t, i) => ({
+    i,
+    key: t.getAttribute('data-contextmenu-trigger'),
+    active: t.getAttribute('data-active') === 'true',
+    title: (t.innerText ?? '').slice(0, 60),
+  }));
+
+w.__clickTabByKey = (key: string) => {
+  const tab = listTopBarTabs().find((t) => t.getAttribute('data-contextmenu-trigger') === key);
+  if (!tab) return 'not found: ' + key;
+  if (tab.getAttribute('data-active') === 'true') return 'already active: ' + key;
+  tab.click();
+  return 'clicked key=' + key;
+};
+
+w.__PROBE_EVENT = (name: string) => {
+  calls.push({ t: now(), name: 'MARK:' + name });
+};
+
+// `run.ts` wraps the bundle in an IIFE and appends a `return <confirmation>`
+// after the bundle body — agent-browser then prints the confirmation back to
+// the operator. Nothing to do here at the end of the module body.

.agents/skills/agent-testing/scripts/agent-gateway/probe.js

@@ -0,0 +1,204 @@
+// LobeHub chat streaming time-series probe.
+//
+// Inject into the renderer (via agent-browser eval) to record store + DOM
+// snapshots every 200ms during a streaming session. Designed to surface
+// "UI rolled back to an earlier state" symptoms — especially around
+// gateway-mode tab switches that happen while the assistant is still writing.
+//
+// Usage:
+//   agent-browser --cdp 9222 eval --stdin < probe.js
+//   # ...do test interactions, call window.__PROBE_EVENT('LABEL') to mark moments...
+//   agent-browser --cdp 9222 eval --stdin < probe-dump.js > /tmp/probe.json
+//   node analyze.mjs /tmp/probe.json
+//
+// What it captures per sample:
+//   - activeTopicId
+//   - msgN: top-level messages in chat.messagesMap for this topic
+//   - childN: total assistantGroup.children blocks across all msgs (THIS is
+//     where streaming content actually lives — top-level assistantGroup stays empty)
+//   - cT / rT / toolT: totals across messages AND their children
+//                       (content, reasoning, tool-call count)
+//   - perMsg: per-message breakdown so regressions can be located precisely
+//   - runOps: number of running operations (execServerAgentRuntime etc.)
+//   - domLen: total innerText length of the rendered chat list area
+//   - ind: visible UI indicators (Search pages, Crawled pages, Deeply Thought, Sending)
+//
+// Event markers: window.__PROBE_EVENT('NAME') records {t, name} into
+// __PROBE_EVENTS, used by the analyzer to align state changes with
+// user-driven actions (SENT, AWAY_1, BACK_1, ...).
+
+(function () {
+  if (window.__PROBE_TIMER) clearInterval(window.__PROBE_TIMER);
+  window.__PROBE_SAMPLES = [];
+  window.__PROBE_EVENTS = [];
+  const t0 = Date.now();
+
+  function snapshot() {
+    try {
+      const chat = window.__LOBE_STORES.chat();
+      const topicId = chat.activeTopicId;
+      const idTail = topicId ? topicId.replace('tpc_', '') : null;
+      const keys = Object.keys(chat.messagesMap || {});
+
+      // Collect messages for the active topic. Before a topic is committed,
+      // optimistic messages live under the `<agentScope>_new` key — fall
+      // back to those when no topic is active yet.
+      let msgs = [];
+      if (idTail) {
+        keys.forEach((k) => {
+          if (k.includes(idTail)) msgs = msgs.concat(chat.messagesMap[k] || []);
+        });
+      } else {
+        keys
+          .filter((k) => k.endsWith('_new'))
+          .forEach((k) => {
+            msgs = msgs.concat(chat.messagesMap[k] || []);
+          });
+      }
+
+      // Walk top-level + assistantGroup.children. children carry the actual
+      // streamed content / reasoning / tool calls; the parent assistantGroup
+      // remains a placeholder (cLen=0, rLen=0) for its whole lifetime.
+      let totalContent = 0;
+      let totalReason = 0;
+      let totalTools = 0;
+      let childCount = 0;
+      const perMsg = msgs.map((m) => {
+        const cLen = (m.content || '').length;
+        const rLen = ((m.reasoning && m.reasoning.content) || '').length;
+        const tools = (m.tools || []).length;
+        totalContent += cLen;
+        totalReason += rLen;
+        totalTools += tools;
+
+        const children = m.children || [];
+        let chC = 0;
+        let chR = 0;
+        let chT = 0;
+        children.forEach((c) => {
+          chC += (c.content || '').length;
+          chR += ((c.reasoning && c.reasoning.content) || '').length;
+          chT += (c.tools || []).length;
+        });
+        totalContent += chC;
+        totalReason += chR;
+        totalTools += chT;
+        childCount += children.length;
+
+        return {
+          id: (m.id || '').slice(-8),
+          role: m.role,
+          cLen,
+          rLen,
+          tools,
+          chCount: children.length,
+          chC,
+          chR,
+          chT,
+        };
+      });
+
+      const ops = Object.values(chat.operations || {});
+      const runningOps = ops.filter((o) => o.status === 'running');
+
+      // DOM probe: total rendered text in the chat scroll area (proxy for
+      // "how much is actually visible to the user").
+      const convScroll =
+        document.querySelector(
+          '[data-chat-list], [class*="ChatList"], [class*="ConversationList"]',
+        ) ||
+        document.querySelector('main [class*="scroll"]') ||
+        document.querySelector('main');
+      const domTxt = convScroll ? convScroll.innerText || '' : '';
+
+      const bodyTxt = document.body.innerText || '';
+      const searchMatches = (bodyTxt.match(/Search pages?:|Searched the web/g) || []).length;
+      const crawlMatches = (bodyTxt.match(/Crawl(ed|ing) pages?/g) || []).length;
+
+      window.__PROBE_SAMPLES.push({
+        t: Date.now() - t0,
+        topicId,
+        msgN: msgs.length,
+        childN: childCount,
+        cT: totalContent,
+        rT: totalReason,
+        toolT: totalTools,
+        perMsg,
+        runOps: runningOps.length,
+        runOpTypes: runningOps.map((o) => o.type),
+        domLen: domTxt.length,
+        ind: {
+          search: searchMatches,
+          crawl: crawlMatches,
+          sending: bodyTxt.includes('Sending message'),
+          deeplyThinking: bodyTxt.includes('Deeply Thinking'),
+          deeplyThought: bodyTxt.includes('Deeply Thought'),
+        },
+      });
+    } catch (e) {
+      window.__PROBE_SAMPLES.push({ t: Date.now() - t0, err: e.message });
+    }
+  }
+
+  snapshot();
+  window.__PROBE_TIMER = setInterval(snapshot, 200);
+  window.__PROBE_EVENT = function (name) {
+    window.__PROBE_EVENTS.push({ t: Date.now() - t0, name });
+  };
+
+  // Tab-switch helpers installed alongside the probe.
+  //
+  // The Electron tab bar mounts each tab as a div with data-insp-path
+  // ending in `TabItem.tsx:...`. The active tab is marked with
+  // data-active="true". DO NOT search by innerText — the active tab's text
+  // includes a ` · <agent name>` suffix that produces false matches when
+  // your search string happens to overlap with the agent name.
+  function listTabs() {
+    return Array.from(
+      document.querySelectorAll('[data-insp-path*="TabItem.tsx"][data-contextmenu-trigger]'),
+    ).filter((t) => t.getBoundingClientRect().top < 30);
+  }
+  function tabKey(el) {
+    // Stable for the tab's lifetime; survives focus changes.
+    return el.getAttribute('data-contextmenu-trigger');
+  }
+  function findActiveTab() {
+    return listTabs().find((t) => t.getAttribute('data-active') === 'true') || null;
+  }
+
+  // Click by stable key captured earlier (preferred for round-trips).
+  window.__clickTabByKey = function (key) {
+    const tab = listTabs().find((t) => tabKey(t) === key);
+    if (!tab) return 'not found: key=' + key;
+    if (tab.getAttribute('data-active') === 'true') return 'already active: ' + key;
+    tab.click();
+    return 'clicked key=' + key;
+  };
+
+  // Click by index in the tab strip (0-based, left-to-right).
+  window.__clickTabByIndex = function (i) {
+    const tabs = listTabs();
+    if (i < 0 || i >= tabs.length) return 'index out of range: ' + i + '/' + tabs.length;
+    const t = tabs[i];
+    if (t.getAttribute('data-active') === 'true') return 'already active: i=' + i;
+    t.click();
+    return 'clicked i=' + i + ' key=' + tabKey(t);
+  };
+
+  // Snapshot all tabs in order: [{key, active, title (first 60 chars of innerText)}]
+  window.__listTabs = function () {
+    return listTabs().map((t, i) => ({
+      i,
+      key: tabKey(t),
+      active: t.getAttribute('data-active') === 'true',
+      title: (t.innerText || '').slice(0, 60),
+    }));
+  };
+
+  window.__activeTabKey = function () {
+    const a = findActiveTab();
+    return a ? tabKey(a) : null;
+  };
+
+  return 'probe installed';
+})();

.agents/skills/agent-testing/scripts/agent-gateway/run.ts

@@ -0,0 +1,211 @@
+// CLI for the agent-gateway probe.
+//
+// Bundles the TS probes with esbuild, pipes them into `agent-browser eval`,
+// and persists dumps under `.agent-gateway/` (gitignored) for later use as
+// streaming-replay test fixtures.
+//
+// Commands:
+//   bun run .agents/skills/agent-testing/scripts/agent-gateway/run.ts install
+//       Bundle probe-events.ts and inject into the CDP-attached browser.
+//       Re-installing clears all buffers and re-patches WebSocket / fetch.
+//
+//   bun run .agents/skills/agent-testing/scripts/agent-gateway/run.ts dump [name]
+//       Stop the timeline timer, fetch the capture as JSON, write it to
+//       `.agent-gateway/<name>-<YYYYMMDD-HHmmss>.json`. `name` defaults to
+//       `dump`. Prints the absolute path written.
+//
+//   bun run .agents/skills/agent-testing/scripts/agent-gateway/run.ts analyze [path]
+//       Run analyze-events.ts on the dump. `path` defaults to the most
+//       recently modified file in `.agent-gateway/`.
+//
+// Optional flags:
+//   --cdp <port>     CDP port (default 9222)
+//   --browser <bin>  agent-browser binary (default 'agent-browser')
+
+import { spawn } from 'node:child_process';
+import { mkdirSync, readdirSync, statSync, writeFileSync } from 'node:fs';
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const SCRIPT_DIR = path.dirname(fileURLToPath(import.meta.url));
+// .agents/skills/agent-testing/scripts/agent-gateway/ → 5 levels up
+const PROJECT_ROOT = path.resolve(SCRIPT_DIR, '../../../../..');
+const DUMP_DIR = path.join(PROJECT_ROOT, '.agent-gateway');
+
+interface Flags {
+  browser: string;
+  cdp: string;
+  positional: string[];
+}
+
+function parseFlags(argv: string[]): Flags {
+  const out: Flags = { cdp: '9222', browser: 'agent-browser', positional: [] };
+  for (let i = 0; i < argv.length; i++) {
+    const a = argv[i];
+    if (a === '--cdp') out.cdp = argv[++i] ?? out.cdp;
+    else if (a === '--browser') out.browser = argv[++i] ?? out.browser;
+    else out.positional.push(a);
+  }
+  return out;
+}
+
+async function bundle(entry: string): Promise<string> {
+  // Bun.build is built into the Bun runtime — no external dep needed.
+  const r = await Bun.build({
+    entrypoints: [path.join(SCRIPT_DIR, entry)],
+    target: 'browser',
+    format: 'esm',
+    minify: false,
+  });
+  if (!r.success) {
+    const msgs = r.logs.map((l) => `${l.level}: ${l.message}`).join('\n');
+    throw new Error(`bundle failed for ${entry}:\n${msgs}`);
+  }
+  return await r.outputs[0].text();
+}
+
+function wrapIife(body: string, returnExpr: string): string {
+  // Wrap as an IIFE that swallows the bundled top-level (top-level `const`
+  // declarations get scoped to the IIFE, so re-injection doesn't conflict)
+  // and returns the configured expression — which `agent-browser eval`
+  // captures and prints to stdout.
+  return `(() => {\n${body}\n;return ${returnExpr};\n})()`;
+}
+
+function runAgentBrowserEval(flags: Flags, script: string): Promise<string> {
+  return new Promise((resolveP, rejectP) => {
+    const child = spawn(flags.browser, ['--cdp', flags.cdp, 'eval', '--stdin'], {
+      stdio: ['pipe', 'pipe', 'inherit'],
+    });
+    let stdout = '';
+    child.stdout.on('data', (chunk: Buffer) => {
+      stdout += chunk.toString('utf8');
+    });
+    child.on('error', rejectP);
+    child.on('close', (code) => {
+      if (code === 0) resolveP(stdout);
+      else rejectP(new Error(`agent-browser exited ${code}`));
+    });
+    child.stdin.write(script);
+    child.stdin.end();
+  });
+}
+
+// agent-browser prints eval results as JSON (string values are quoted).
+function unquoteAgentBrowserResult(raw: string): string {
+  const trimmed = raw.trim();
+  if (trimmed.startsWith('"') && trimmed.endsWith('"')) {
+    try {
+      return JSON.parse(trimmed) as string;
+    } catch {
+      /* fall through */
+    }
+  }
+  return trimmed;
+}
+
+function isoStamp(): string {
+  const d = new Date();
+  const yyyy = d.getFullYear();
+  const mm = String(d.getMonth() + 1).padStart(2, '0');
+  const dd = String(d.getDate()).padStart(2, '0');
+  const hh = String(d.getHours()).padStart(2, '0');
+  const mi = String(d.getMinutes()).padStart(2, '0');
+  const ss = String(d.getSeconds()).padStart(2, '0');
+  return `${yyyy}${mm}${dd}-${hh}${mi}${ss}`;
+}
+
+function ensureDumpDir(): void {
+  mkdirSync(DUMP_DIR, { recursive: true });
+}
+
+function latestDump(): string | null {
+  ensureDumpDir();
+  const entries = readdirSync(DUMP_DIR)
+    .filter((f) => f.endsWith('.json'))
+    .map((f) => ({ f, mtime: statSync(path.join(DUMP_DIR, f)).mtimeMs }))
+    .sort((a, b) => b.mtime - a.mtime);
+  return entries[0] ? path.join(DUMP_DIR, entries[0].f) : null;
+}
+
+// ── Commands ────────────────────────────────────────────────────────
+
+async function cmdInstall(flags: Flags): Promise<void> {
+  const body = await bundle('probe-events.ts');
+  const installMsg = JSON.stringify(
+    'events probe installed: WebSocket+fetch interception. ' +
+      'WS captures operationId= sockets (gateway), fetch captures /api/agent/stream (direct).',
+  );
+  const script = wrapIife(body, installMsg);
+  const out = await runAgentBrowserEval(flags, script);
+  console.log(unquoteAgentBrowserResult(out));
+}
+
+async function cmdDump(flags: Flags): Promise<void> {
+  const name = flags.positional[1] ?? 'dump';
+  const body = await bundle('probe-dump.ts');
+  const script = wrapIife(body, 'window.__PROBE_LAST_DUMP_JSON');
+  const raw = await runAgentBrowserEval(flags, script);
+  const json = unquoteAgentBrowserResult(raw);
+  ensureDumpDir();
+  const filename = `${name}-${isoStamp()}.json`;
+  const dumpPath = path.join(DUMP_DIR, filename);
+  writeFileSync(dumpPath, json, 'utf8');
+  // Validate by parsing the meta header so we error early on bad capture
+  try {
+    const parsed = JSON.parse(json) as {
+      meta?: { eventCount?: number; callCount?: number; sampleCount?: number };
+    };
+    const meta = parsed.meta ?? {};
+    console.log(
+      `wrote ${dumpPath}  (${json.length} bytes  events=${meta.eventCount ?? '?'}  ` +
+        `calls=${meta.callCount ?? '?'}  samples=${meta.sampleCount ?? '?'})`,
+    );
+  } catch {
+    console.log(`wrote ${dumpPath}  (${json.length} bytes — JSON.parse failed; see file)`);
+  }
+}
+
+async function cmdAnalyze(flags: Flags): Promise<void> {
+  const target = flags.positional[1] ?? latestDump();
+  if (!target) {
+    console.error('no dump file found. run `dump` first or pass a path.');
+    process.exit(1);
+  }
+  const child = spawn('bun', ['run', path.join(SCRIPT_DIR, 'analyze-events.ts'), target], {
+    stdio: 'inherit',
+  });
+  await new Promise<void>((resolveP, rejectP) => {
+    child.on('error', rejectP);
+    child.on('close', (code) => (code === 0 ? resolveP() : rejectP(new Error(`exit ${code}`))));
+  });
+}
+
+// ── Entry point ─────────────────────────────────────────────────────
+
+const flags = parseFlags(process.argv.slice(2));
+const cmd = flags.positional[0];
+
+const usage = `usage:
+  bun run run.ts install [--cdp 9222]
+  bun run run.ts dump [name] [--cdp 9222]
+  bun run run.ts analyze [path]
+`;
+
+if (!cmd) {
+  console.error(usage);
+  process.exit(1);
+}
+
+try {
+  if (cmd === 'install') await cmdInstall(flags);
+  else if (cmd === 'dump') await cmdDump(flags);
+  else if (cmd === 'analyze') await cmdAnalyze(flags);
+  else {
+    console.error(`unknown command: ${cmd}\n\n${usage}`);
+    process.exit(1);
+  }
+} catch (e: any) {
+  console.error(e?.stack ?? e);
+  process.exit(1);
+}

.agents/skills/agent-testing/scripts/agent-gateway/tab-switch.js

@@ -0,0 +1,72 @@
+// Run N round-trip tab switches with event markers timed against the probe.
+//
+//   agent-browser --cdp 9222 eval --stdin < tab-switch.js
+//
+// Captures the currently-active tab as the BACK target and the rightmost
+// inactive tab as the AWAY target. Both are addressed by their stable
+// data-contextmenu-trigger key (NOT by visible title — the active tab's
+// innerText embeds a ` · <agent name>` suffix that breaks text matching).
+//
+// Fires the loop in the background and returns immediately so the
+// agent-browser eval doesn't have to await the full ROUND_TRIPS × DWELL_MS
+// duration. Wait on the `SWITCH_LOOP_DONE` event before dumping.
+//
+// Refuses to launch if a previous loop is still in flight.
+//
+// Requires probe.js to have been installed first (provides
+// window.__PROBE_EVENT / __listTabs / __clickTabByKey / __activeTabKey).
+
+(function () {
+  const ROUND_TRIPS = 4;
+  const DWELL_MS = 10_000;
+
+  if (!window.__PROBE_EVENT || !window.__listTabs || !window.__clickTabByKey) {
+    return 'probe not installed — eval probe.js first';
+  }
+  if (window.__SWITCH_LOOP_RUNNING) {
+    return 'switch loop already running — wait for SWITCH_LOOP_DONE first';
+  }
+
+  const tabs = window.__listTabs();
+  const activeTab = tabs.find((t) => t.active);
+  if (!activeTab) return 'no active tab — abort';
+
+  // Pick the first inactive tab as AWAY target. With multiple inactive tabs
+  // you'll usually want the one that's stable across the test — feel free
+  // to swap to tabs[tabs.length-1] if you want the rightmost.
+  const inactives = tabs.filter((t) => !t.active);
+  if (inactives.length === 0) return 'no inactive tab to switch to — abort';
+  const awayTab = inactives.at(-1); // rightmost inactive
+
+  const BACK_KEY = activeTab.key;
+  const AWAY_KEY = awayTab.key;
+
+  window.__SWITCH_LOOP_RUNNING = true;
+  window.__PROBE_EVENT('SWITCH_LOOP_CONFIG:back=' + BACK_KEY + ',away=' + AWAY_KEY);
+
+  (async function () {
+    function sleep(ms) {
+      return new Promise((r) => setTimeout(r, ms));
+    }
+
+    try {
+      window.__PROBE_EVENT('SWITCH_LOOP_START');
+      for (let i = 1; i <= ROUND_TRIPS; i++) {
+        window.__PROBE_EVENT('AWAY_' + i);
+        const awayResult = window.__clickTabByKey(AWAY_KEY);
+        window.__PROBE_EVENT('AWAY_' + i + '_RES:' + awayResult.slice(0, 50));
+        await sleep(DWELL_MS);
+
+        window.__PROBE_EVENT('BACK_' + i);
+        const backResult = window.__clickTabByKey(BACK_KEY);
+        window.__PROBE_EVENT('BACK_' + i + '_RES:' + backResult.slice(0, 50));
+        await sleep(DWELL_MS);
+      }
+      window.__PROBE_EVENT('SWITCH_LOOP_DONE');
+    } finally {
+      window.__SWITCH_LOOP_RUNNING = false;
+    }
+  })();
+
+  return 'switch loop kicked off (BACK=' + BACK_KEY + ', AWAY=' + AWAY_KEY + ')';
+})();

.agents/skills/agent-testing/scripts/agent-gateway/types.ts

@@ -0,0 +1,113 @@
+// Shared types between the in-browser probe and the Node-side analyzer.
+// Kept tiny on purpose — anything the analyzer can re-derive is left off.
+
+export interface ProbeStreamEvent {
+  /** Summarized payload — long strings truncated, arrays printed as Array(N) */
+  data?: Record<string, unknown>;
+  /** Keys present on the event's `data` payload — useful at a glance */
+  dataKeys?: string[];
+  /** ServerMessage.id — gateway WS frames carry an event-id we may resume from */
+  eventId?: string | null;
+  message?: string;
+  /** Last 10 chars of the operationId (full id is excessively long) */
+  opIdTail: string;
+  raw?: string;
+  /** Raw frame byte length, when applicable */
+  rawLen?: number;
+  /** For non-agent_event server frames (auth_success, heartbeat_ack, …) */
+  serverType?: string;
+  sseEvent?: string;
+  status?: number;
+  stepIndex?: number;
+  /** Milliseconds since the probe's t0 (install time). */
+  t: number;
+  /** 'ws' for gateway WebSocket frames, 'sse' for direct /api/agent/stream */
+  transport: 'ws' | 'sse';
+  /** Either the AgentStreamEvent.type, or a probe sentinel like `_WS_OPEN_` */
+  type: string;
+  url?: string;
+}
+
+export interface ProbeActionCall {
+  args?: {
+    count?: number;
+    context?: unknown;
+    params?: unknown;
+  };
+  error?: string;
+  /** `replaceMessages` / `refreshMessages` / `MARK:<label>` / `_WRAP_ERROR_` */
+  name: string;
+  stack?: string;
+  t: number;
+}
+
+export interface ProbeMessageSummary {
+  /** children.length */
+  chN: number;
+  /** content.length */
+  cLen: number;
+  /** Last 8 chars of the message id */
+  id: string;
+  /** reasoning.content.length */
+  rLen: number;
+  role: string;
+  /** tools.length */
+  tools: number;
+}
+
+export interface ProbeTimelineSample {
+  /** Last 10 chars of activeTopicId, or null */
+  activeTopic: string | null;
+  /** Per-key breakdown: display count, db count, message summaries */
+  byKey: Record<
+    string,
+    {
+      n: number;
+      dbN: number;
+      msgs: ProbeMessageSummary[];
+    }
+  >;
+  err?: string;
+  /** All messagesMap keys that have content at this moment */
+  keys: string[];
+  /** Number of operations in 'running' status */
+  runOps: number;
+  t: number;
+}
+
+export interface ProbeDumpMeta {
+  callCount: number;
+  /** Date.now() at dump call */
+  collectedAt: number;
+  eventCount: number;
+  sampleCount: number;
+  /** Date.now() at probe install */
+  t0: number;
+}
+
+export interface ProbeDump {
+  actionCalls: ProbeActionCall[];
+  meta: ProbeDumpMeta;
+  streamEvents: ProbeStreamEvent[];
+  timeline: ProbeTimelineSample[];
+}
+
+/**
+ * Globals the probe attaches to `window`. Keeps `as any` casts at the boundary
+ * instead of sprinkling them through the probe body.
+ */
+declare global {
+  interface Window {
+    __clickTabByKey?: (key: string) => string;
+    __listTabs?: () => Array<{ i: number; key: string | null; active: boolean; title: string }>;
+    __LOBE_STORES?: Record<string, () => any>;
+    __PROBE_ACTION_CALLS?: ProbeActionCall[];
+    __PROBE_EVENT?: (label: string) => void;
+    __PROBE_MSG_TIMELINE?: ProbeTimelineSample[];
+    __PROBE_ORIG_FETCH?: typeof fetch;
+    __PROBE_ORIG_WEBSOCKET?: typeof WebSocket;
+    __PROBE_STREAM_EVENTS?: ProbeStreamEvent[];
+    __PROBE_T0?: number;
+    __PROBE_TIMELINE_TIMER?: ReturnType<typeof setInterval> | null;
+  }
+}

.agents/skills/agent-testing/scripts/app-probe.sh

@@ -0,0 +1,95 @@
+#!/usr/bin/env bash
+# app-probe.sh — standardized probes for a running LobeHub app (Electron via
+# CDP, or a web agent-browser session). Use these instead of hand-rolling
+# `window.__LOBE_STORES` eval snippets — especially the auth check.
+#
+# Usage:
+#   app-probe.sh auth              # { isSignedIn, userId } from the user store
+#   app-probe.sh route             # current SPA route
+#   app-probe.sh ops               # running chat operations (type / status / startTime)
+#   app-probe.sh goto <path>       # navigate the SPA to a route (full reload), e.g. goto /agent/agt_xxx
+#   app-probe.sh errors-install    # install a console.error interceptor
+#   app-probe.sh errors            # dump errors captured since errors-install
+#
+# Target selection (default: Electron over CDP 9222):
+#   AB_TARGET="--cdp 9222"             # Electron (default; CDP_PORT also honored)
+#   AB_TARGET="--session lobehub-dev"  # web agent-browser session
+#
+# Common routes (desktop SPA): /  /agent/<agentId>  /agent/<agentId>/<topicId>
+#   /task  /task/<taskId>  /page  /settings  /community
+
+set -euo pipefail
+
+AB_TARGET="${AB_TARGET:---cdp ${CDP_PORT:-9222}}"
+
+run_eval() {
+  # shellcheck disable=SC2086
+  agent-browser $AB_TARGET eval --stdin
+}
+
+case "${1:-}" in
+  auth)
+    run_eval << 'EVALEOF'
+(function () {
+  var stores = window.__LOBE_STORES;
+  if (!stores || !stores.user) return JSON.stringify({ ok: false, reason: 'no user store — app not loaded yet?' });
+  var u = stores.user();
+  return JSON.stringify({ ok: !!u.isSignedIn, isSignedIn: !!u.isSignedIn, userId: (u.user && u.user.id) || null });
+})()
+EVALEOF
+    ;;
+  route)
+    run_eval << 'EVALEOF'
+location.pathname + location.search + location.hash
+EVALEOF
+    ;;
+  ops)
+    run_eval << 'EVALEOF'
+(function () {
+  var stores = window.__LOBE_STORES;
+  if (!stores || !stores.chat) return JSON.stringify({ ok: false, reason: 'no chat store — open a conversation first' });
+  var ops = Object.values(stores.chat().operations || {});
+  var running = ops.filter(function (o) { return o.status === 'running'; });
+  return JSON.stringify({
+    ok: true,
+    running: running.map(function (o) { return { startTime: o.metadata && o.metadata.startTime, type: o.type }; }),
+    runningCount: running.length,
+    total: ops.length,
+  });
+})()
+EVALEOF
+    ;;
+  goto)
+    TARGET_PATH="${2:?Usage: app-probe.sh goto <path>}"
+    # shellcheck disable=SC2086
+    agent-browser $AB_TARGET eval "location.href = '$TARGET_PATH'" > /dev/null
+    sleep 2
+    bash "${BASH_SOURCE[0]}" route
+    ;;
+  errors-install)
+    run_eval << 'EVALEOF'
+(function () {
+  window.__CAPTURED_ERRORS = [];
+  var orig = console.error;
+  console.error = function () {
+    var msg = Array.from(arguments).map(function (a) {
+      if (a instanceof Error) return a.message;
+      return typeof a === 'object' ? JSON.stringify(a) : String(a);
+    }).join(' ');
+    window.__CAPTURED_ERRORS.push(msg);
+    orig.apply(console, arguments);
+  };
+  return 'installed';
+})()
+EVALEOF
+    ;;
+  errors)
+    run_eval << 'EVALEOF'
+JSON.stringify(window.__CAPTURED_ERRORS || 'interceptor not installed — run errors-install first')
+EVALEOF
+    ;;
+  *)
+    echo "Usage: $0 {auth|route|ops|goto <path>|errors-install|errors}" >&2
+    exit 2
+    ;;
+esac

.agents/skills/agent-testing/scripts/electron-dev.sh

@@ -76,7 +76,9 @@ find_project_pids() {
   port_pid=$(lsof -ti tcp:"$CDP_PORT" -sTCP:LISTEN 2>/dev/null || true)
   pids="$pids $port_pid"
 
-  echo "$pids" | tr ' ' '\n' | sort -u | grep -v '^$' | tr '\n' ' '
+  # `|| true` because `grep -v '^$'` exits 1 when input has no non-empty
+  # lines, which (with pipefail + set -e) silently kills the caller.
+  echo "$pids" | tr ' ' '\n' | sort -u | grep -v '^$' | tr '\n' ' ' || true
 }
 
 # Wait for the CDP HTTP endpoint to respond, with a deadline + early bail-out
@@ -146,7 +148,7 @@ do_stop() {
   for pid in $seed_pids; do
     all_pids="$all_pids $(expand_descendants "$pid")"
   done
-  all_pids=$(echo "$all_pids" | tr ' ' '\n' | sort -u | grep -v '^$' | tr '\n' ' ')
+  all_pids=$(echo "$all_pids" | tr ' ' '\n' | sort -u | grep -v '^$' | tr '\n' ' ' || true)
 
   if [ -z "$all_pids" ]; then
     echo "[electron-dev] No project Electron/vite processes found."
@@ -270,10 +272,17 @@ do_start() {
   # Launch in a new session (setsid) so the whole process tree shares a PGID
   # we can later signal in one shot. `setsid bash -c '... exec ...' &` keeps
   # the bash shell as the session leader; its PID is what we save.
-  setsid bash -c "
+  # macOS doesn't ship setsid by default — fall back to plain bash; cleanup
+  # still works via `expand_descendants` walking the process tree.
+  local launch_cmd="
     cd '$PROJECT_ROOT/apps/desktop'
     exec npx electron-vite dev -- --remote-debugging-port=$CDP_PORT
-  " >> "$ELECTRON_LOG" 2>&1 < /dev/null &
+  "
+  if command -v setsid >/dev/null 2>&1; then
+    setsid bash -c "$launch_cmd" >> "$ELECTRON_LOG" 2>&1 < /dev/null &
+  else
+    bash -c "$launch_cmd" >> "$ELECTRON_LOG" 2>&1 < /dev/null &
+  fi
   local launcher_pid=$!
   echo "$launcher_pid" > "$PIDFILE"
   echo "[electron-dev] Launcher PID (session leader): $launcher_pid"

.agents/skills/agent-testing/scripts/init-dev-env.sh

@@ -0,0 +1,407 @@
+#!/usr/bin/env bash
+# init-dev-env.sh — self-contained local dev env for agent testing.
+#
+# This script initializes the env needed to run LobeHub's normal local dev
+# server without depending on a root .env file. It follows the same shape as
+# the e2e bootstrap (Postgres + migrations + auth/key-vault/S3 test env), but
+# starts the repo's dev server, not the standalone e2e server.
+#
+# Guardrail: if repo-root .env exists, every non-help command exits immediately.
+# Existing local config always wins.
+#
+# Usage:
+#   init-dev-env.sh env              # print shell exports
+#   init-dev-env.sh write [file]     # write a source-able env file
+#   init-dev-env.sh setup-db         # start local Postgres and run migrations
+#   init-dev-env.sh migrate          # run DB migrations against the configured DB
+#   init-dev-env.sh seed-user        # seed the baseline test user + CLI API key
+#   init-dev-env.sh qstash           # run local Upstash QStash dev server
+#   init-dev-env.sh dev-next         # exec `pnpm run dev:next` with this env
+#   init-dev-env.sh dev              # exec `bun run dev` with this env
+#   init-dev-env.sh clean-db         # remove the managed Postgres container
+#
+# Overrides:
+#   SERVER_PORT=3010 DB_PORT=5433 DB_CONTAINER=lobehub-agent-testing-postgres QSTASH_DEV_PORT=8080
+
+set -euo pipefail
+
+REPO_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/../../../.." && pwd)"
+ROOT_ENV_FILE="$REPO_ROOT/.env"
+
+SERVER_PORT="${SERVER_PORT:-3010}"
+DB_PORT="${DB_PORT:-5433}"
+DB_CONTAINER="${DB_CONTAINER:-lobehub-agent-testing-postgres}"
+DATABASE_URL="${DATABASE_URL:-postgresql://postgres:postgres@localhost:${DB_PORT}/postgres}"
+ENV_FILE_DEFAULT="$REPO_ROOT/.records/env/agent-testing-dev.env"
+CLI_ENV_FILE_DEFAULT="$REPO_ROOT/.records/env/agent-testing-cli.env"
+AGENT_TESTING_API_KEY="${AGENT_TESTING_API_KEY:-sk-lh-agenttesting0001}"
+QSTASH_DEV_PORT="${QSTASH_DEV_PORT:-8080}"
+QSTASH_LOCAL_TOKEN="${QSTASH_LOCAL_TOKEN:-eyJVc2VySUQiOiJkZWZhdWx0VXNlciIsIlBhc3N3b3JkIjoiZGVmYXVsdFBhc3N3b3JkIn0=}"
+QSTASH_LOCAL_CURRENT_SIGNING_KEY="${QSTASH_LOCAL_CURRENT_SIGNING_KEY:-sig_7kYjw48mhY7kAjqNGcy6cr29RJ6r}"
+QSTASH_LOCAL_NEXT_SIGNING_KEY="${QSTASH_LOCAL_NEXT_SIGNING_KEY:-sig_5ZB6DVzB1wjE8S6rZ7eenA8Pdnhs}"
+
+ok() { printf '  \033[32m✔\033[0m %s\n' "$1"; }
+bad() { printf '  \033[31m✘\033[0m %s\n' "$1"; }
+note() { printf '      %s\n' "$1"; }
+
+guard_no_root_env() {
+  if [[ -f "$ROOT_ENV_FILE" ]]; then
+    bad "root .env exists: $ROOT_ENV_FILE"
+    note "Use the existing local configuration instead of init-dev-env.sh."
+    note "Start normally from repo root, e.g. pnpm run dev:next or bun run dev."
+    exit 1
+  fi
+}
+
+apply_env() {
+  export APP_URL="${APP_URL:-http://localhost:${SERVER_PORT}}"
+  export AUTH_EMAIL_VERIFICATION="${AUTH_EMAIL_VERIFICATION:-0}"
+  export AUTH_SECRET="${AUTH_SECRET:-agent-testing-local-auth-secret-32chars}"
+  export DATABASE_DRIVER="${DATABASE_DRIVER:-node}"
+  export DATABASE_URL
+  export FEATURE_FLAGS="${FEATURE_FLAGS:--agent_self_iteration}"
+  export KEY_VAULTS_SECRET="${KEY_VAULTS_SECRET:-r2gbBPKyJ8ZRKCLKt+I3DImfcL+wGxaQyRC56xtm9Uk=}"
+  export NEXT_PUBLIC_AUTH_EMAIL_VERIFICATION="${NEXT_PUBLIC_AUTH_EMAIL_VERIFICATION:-0}"
+  export NODE_OPTIONS="${NODE_OPTIONS:---max-old-space-size=6144}"
+  export PORT="${PORT:-$SERVER_PORT}"
+  export QSTASH_CURRENT_SIGNING_KEY="${QSTASH_CURRENT_SIGNING_KEY:-$QSTASH_LOCAL_CURRENT_SIGNING_KEY}"
+  export QSTASH_DEV_PORT
+  export QSTASH_NEXT_SIGNING_KEY="${QSTASH_NEXT_SIGNING_KEY:-$QSTASH_LOCAL_NEXT_SIGNING_KEY}"
+  export QSTASH_TOKEN="${QSTASH_TOKEN:-$QSTASH_LOCAL_TOKEN}"
+  export QSTASH_URL="${QSTASH_URL:-http://127.0.0.1:${QSTASH_DEV_PORT}}"
+  export S3_ACCESS_KEY_ID="${S3_ACCESS_KEY_ID:-agent-testing-access-key}"
+  export S3_BUCKET="${S3_BUCKET:-agent-testing-bucket}"
+  export S3_ENDPOINT="${S3_ENDPOINT:-https://agent-testing-s3.localhost}"
+  export S3_SECRET_ACCESS_KEY="${S3_SECRET_ACCESS_KEY:-agent-testing-secret-key}"
+}
+
+env_keys() {
+  printf '%s\n' \
+    APP_URL \
+    AUTH_EMAIL_VERIFICATION \
+    AUTH_SECRET \
+    DATABASE_DRIVER \
+    DATABASE_URL \
+    FEATURE_FLAGS \
+    KEY_VAULTS_SECRET \
+    NEXT_PUBLIC_AUTH_EMAIL_VERIFICATION \
+    NODE_OPTIONS \
+    PORT \
+    QSTASH_CURRENT_SIGNING_KEY \
+    QSTASH_DEV_PORT \
+    QSTASH_NEXT_SIGNING_KEY \
+    QSTASH_TOKEN \
+    QSTASH_URL \
+    S3_ACCESS_KEY_ID \
+    S3_BUCKET \
+    S3_ENDPOINT \
+    S3_SECRET_ACCESS_KEY
+}
+
+print_env() {
+  apply_env
+  while IFS= read -r key; do
+    printf 'export %s=%q\n' "$key" "${!key}"
+  done < <(env_keys)
+}
+
+write_env() {
+  local file="${1:-$ENV_FILE_DEFAULT}"
+  apply_env
+  mkdir -p "$(dirname "$file")"
+  {
+    printf '# Source this file before starting LobeHub local dev server.\n'
+    printf '# Generated by %s\n' "$0"
+    while IFS= read -r key; do
+      printf 'export %s=%q\n' "$key" "${!key}"
+    done < <(env_keys)
+  } > "$file"
+  ok "wrote env file: $file"
+  note "source it with: source $file"
+}
+
+require_docker() {
+  if ! command -v docker > /dev/null 2>&1; then
+    bad "docker CLI is not available"
+    note "Install/start Docker Desktop, or provide DATABASE_URL for an existing Postgres."
+    return 1
+  fi
+}
+
+wait_for_db() {
+  printf '      waiting for Postgres'
+  until docker exec "$DB_CONTAINER" pg_isready -U postgres > /dev/null 2>&1; do
+    printf '.'
+    sleep 2
+  done
+  printf '\n'
+}
+
+start_db() {
+  require_docker
+
+  if docker ps --format '{{.Names}}' | grep -Fxq "$DB_CONTAINER"; then
+    ok "Postgres container already running: $DB_CONTAINER"
+  elif docker ps -a --format '{{.Names}}' | grep -Fxq "$DB_CONTAINER"; then
+    docker start "$DB_CONTAINER" > /dev/null
+    ok "started existing Postgres container: $DB_CONTAINER"
+  else
+    docker run -d \
+      --name "$DB_CONTAINER" \
+      -e POSTGRES_PASSWORD=postgres \
+      -p "${DB_PORT}:5432" \
+      paradedb/paradedb:latest > /dev/null
+    ok "created Postgres container: $DB_CONTAINER"
+  fi
+
+  wait_for_db
+}
+
+migrate_db() {
+  apply_env
+  cd "$REPO_ROOT"
+  bun run db:migrate
+}
+
+seed_user() {
+  apply_env
+  export AGENT_TESTING_API_KEY
+  export AGENT_TESTING_CLI_ENV_FILE="${AGENT_TESTING_CLI_ENV_FILE:-$CLI_ENV_FILE_DEFAULT}"
+  cd "$REPO_ROOT"
+  node <<'NODE'
+const bcrypt = require('bcryptjs');
+const crypto = require('node:crypto');
+const fs = require('node:fs');
+const path = require('node:path');
+const pg = require('pg');
+
+const databaseUrl = process.env.DATABASE_URL;
+if (!databaseUrl) {
+  throw new Error('DATABASE_URL is required to seed the baseline test user.');
+}
+
+const TEST_USER = {
+  email: 'agent-testing@lobehub.com',
+  fullName: 'Agent Testing User',
+  id: 'user_agent_testing_001',
+  password: 'TestPassword123!',
+  username: 'agent_testing_user',
+};
+
+const TEST_API_KEY = {
+  id: 'api_key_agent_testing_001',
+  key: process.env.AGENT_TESTING_API_KEY || 'sk-lh-agenttesting0001',
+  name: 'Agent Testing CLI API Key',
+};
+
+const validateApiKeyFormat = (apiKey) => /^sk-lh-[\da-z]{16}$/.test(apiKey);
+
+const hashApiKey = (apiKey) => {
+  const secret = process.env.KEY_VAULTS_SECRET;
+  if (!secret) throw new Error('KEY_VAULTS_SECRET is required to seed the baseline API key.');
+
+  return crypto.createHmac('sha256', secret).update(apiKey).digest('hex');
+};
+
+const encryptWithKeyVaultsSecret = (plaintext) => {
+  const secret = process.env.KEY_VAULTS_SECRET;
+  if (!secret) throw new Error('KEY_VAULTS_SECRET is required to seed the baseline API key.');
+
+  const rawKey = Buffer.from(secret, 'base64');
+  if (![16, 24, 32].includes(rawKey.length)) {
+    throw new Error(
+      `KEY_VAULTS_SECRET must decode to 16, 24, or 32 bytes, got ${rawKey.length} bytes.`,
+    );
+  }
+
+  const iv = crypto.randomBytes(12);
+  const cipher = crypto.createCipheriv(`aes-${rawKey.length * 8}-gcm`, rawKey, iv);
+  const encrypted = Buffer.concat([cipher.update(plaintext, 'utf8'), cipher.final()]);
+  const authTag = cipher.getAuthTag();
+
+  return `${iv.toString('hex')}:${authTag.toString('hex')}:${encrypted.toString('hex')}`;
+};
+
+const writeCliEnvFile = () => {
+  const file = process.env.AGENT_TESTING_CLI_ENV_FILE || '.records/env/agent-testing-cli.env';
+  fs.mkdirSync(path.dirname(file), { recursive: true });
+  fs.writeFileSync(
+    file,
+    [
+      '# Source this file before running LobeHub CLI agent tests.',
+      '# Generated by init-dev-env.sh seed-user',
+      `export LOBE_API_KEY=${TEST_API_KEY.key}`,
+      `export LOBEHUB_CLI_API_KEY="${'${LOBE_API_KEY}'}"`,
+      `export LOBEHUB_SERVER=${process.env.APP_URL}`,
+      'export LOBEHUB_CLI_HOME=.lobehub-dev',
+      '',
+    ].join('\n'),
+  );
+
+  return file;
+};
+
+const client = new pg.Client({ connectionString: databaseUrl });
+
+(async () => {
+  if (!validateApiKeyFormat(TEST_API_KEY.key)) {
+    throw new Error(`Invalid AGENT_TESTING_API_KEY format: ${TEST_API_KEY.key}`);
+  }
+
+  await client.connect();
+  const now = new Date().toISOString();
+  const onboarding = JSON.stringify({ finishedAt: now, version: 1 });
+  const passwordHash = await bcrypt.hash(TEST_USER.password, 10);
+  const encryptedApiKey = encryptWithKeyVaultsSecret(TEST_API_KEY.key);
+  const apiKeyHash = hashApiKey(TEST_API_KEY.key);
+
+  await client.query(
+    `INSERT INTO users (id, email, normalized_email, username, full_name, email_verified, onboarding, created_at, updated_at, last_active_at)
+     VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $8, $8)
+     ON CONFLICT (id) DO UPDATE SET onboarding = $7, updated_at = $8`,
+    [
+      TEST_USER.id,
+      TEST_USER.email,
+      TEST_USER.email.toLowerCase(),
+      TEST_USER.username,
+      TEST_USER.fullName,
+      true,
+      onboarding,
+      now,
+    ],
+  );
+
+  await client.query(
+    `INSERT INTO accounts (id, user_id, account_id, provider_id, password, created_at, updated_at)
+     VALUES ($1, $2, $3, $4, $5, $6, $6)
+     ON CONFLICT DO NOTHING`,
+    [
+      'agent_testing_account_001',
+      TEST_USER.id,
+      TEST_USER.email,
+      'credential',
+      passwordHash,
+      now,
+    ],
+  );
+
+  await client.query(
+    `INSERT INTO api_keys (id, name, key, key_hash, enabled, expires_at, user_id, workspace_id, created_at, updated_at)
+     VALUES ($1, $2, $3, $4, $5, NULL, $6, NULL, $7, $7)
+     ON CONFLICT (id) DO UPDATE
+     SET name = EXCLUDED.name,
+         key = EXCLUDED.key,
+         key_hash = EXCLUDED.key_hash,
+         enabled = EXCLUDED.enabled,
+         expires_at = NULL,
+         updated_at = EXCLUDED.updated_at`,
+    [
+      TEST_API_KEY.id,
+      TEST_API_KEY.name,
+      encryptedApiKey,
+      apiKeyHash,
+      true,
+      TEST_USER.id,
+      now,
+    ],
+  );
+
+  const cliEnvFile = writeCliEnvFile();
+
+  console.log('seeded baseline user:');
+  console.log(`  email: ${TEST_USER.email}`);
+  console.log(`  password: ${TEST_USER.password}`);
+  console.log('seeded baseline API key:');
+  console.log(`  LOBE_API_KEY: ${TEST_API_KEY.key}`);
+  console.log(`  CLI env: ${cliEnvFile}`);
+})()
+  .finally(() => client.end())
+  .catch((error) => {
+    console.error(error);
+    process.exit(1);
+  });
+NODE
+}
+
+cmd_status() {
+  apply_env
+  echo "agent-testing local dev env:"
+  note "APP_URL=$APP_URL"
+  note "DATABASE_URL=$DATABASE_URL"
+  note "PORT=$PORT"
+  note "QSTASH_URL=$QSTASH_URL"
+  if command -v docker > /dev/null 2>&1; then
+    ok "docker CLI available"
+    if docker ps --format '{{.Names}}' | grep -Fxq "$DB_CONTAINER"; then
+      ok "managed Postgres running: $DB_CONTAINER"
+    else
+      note "managed Postgres is not running: $DB_CONTAINER"
+    fi
+  else
+    bad "docker CLI is not available"
+  fi
+}
+
+cmd_qstash() {
+  apply_env
+  cd "$REPO_ROOT"
+  note "starting local QStash dev server at $QSTASH_URL"
+  note "keep this process running while testing workflow paths"
+  exec pnpm run qstash -- -port "$QSTASH_DEV_PORT"
+}
+
+cmd_dev_next() {
+  apply_env
+  cd "$REPO_ROOT"
+  exec pnpm run dev:next
+}
+
+cmd_dev() {
+  apply_env
+  cd "$REPO_ROOT"
+  exec bun run dev
+}
+
+cmd_clean_db() {
+  require_docker
+  if docker ps --format '{{.Names}}' | grep -Fxq "$DB_CONTAINER"; then
+    docker stop "$DB_CONTAINER" > /dev/null
+  fi
+  if docker ps -a --format '{{.Names}}' | grep -Fxq "$DB_CONTAINER"; then
+    docker rm "$DB_CONTAINER" > /dev/null
+    ok "removed Postgres container: $DB_CONTAINER"
+  else
+    note "Postgres container not found: $DB_CONTAINER"
+  fi
+}
+
+usage() {
+  sed -n '3,24p' "$0" >&2
+}
+
+COMMAND="${1:-status}"
+
+case "$COMMAND" in
+  help|-h|--help) usage; exit 0 ;;
+  *) guard_no_root_env ;;
+esac
+
+case "$COMMAND" in
+  env) print_env ;;
+  write) shift; write_env "${1:-}" ;;
+  setup-db)
+    start_db
+    migrate_db
+    ;;
+  migrate) migrate_db ;;
+  seed-user) seed_user ;;
+  qstash) cmd_qstash ;;
+  dev-next) cmd_dev_next ;;
+  dev) cmd_dev ;;
+  clean-db) cmd_clean_db ;;
+  status) cmd_status ;;
+  *)
+    usage
+    exit 2
+    ;;
+esac

.agents/skills/agent-testing/scripts/record-gif.sh

@@ -0,0 +1,61 @@
+#!/usr/bin/env bash
+# record-gif.sh — capture a frame sequence via agent-browser (CDP) and
+# synthesize a GIF for embedding in a test report.
+#
+# Use this whenever the asserted behavior is about CHANGE OVER TIME —
+# streaming output, a ticking timer, loading states, animations. A static
+# screenshot cannot prove those; a GIF can. Cloud-portable: frames come from
+# CDP rendering, no OS-level screen capture.
+#
+# Usage:
+#   record-gif.sh <output.gif> <duration_seconds> [fps]
+#
+#   AB_TARGET="--cdp 9222"             # Electron (default; CDP_PORT honored)
+#   AB_TARGET="--session lobehub-dev"  # web agent-browser session
+#   GIF_WIDTH=960                      # output width (px), default 960
+#
+# Requires ffmpeg (`brew install ffmpeg`). Effective fps is capped by
+# screenshot latency (~0.3-0.5s per frame); 1-2 fps is the realistic range.
+#
+# Example — record a 12s run and embed it in the report:
+#   ./record-gif.sh "$DIR/assets/case2-tray-running.gif" 12 2 &
+#   GIF_PID=$!
+#   # ... trigger the streaming behavior ...
+#   wait $GIF_PID
+
+set -euo pipefail
+
+OUT="${1:?Usage: record-gif.sh <output.gif> <duration_seconds> [fps]}"
+DUR="${2:?Usage: record-gif.sh <output.gif> <duration_seconds> [fps]}"
+FPS="${3:-2}"
+AB_TARGET="${AB_TARGET:---cdp ${CDP_PORT:-9222}}"
+GIF_WIDTH="${GIF_WIDTH:-960}"
+
+command -v ffmpeg > /dev/null || {
+  echo "ffmpeg not found — install with: brew install ffmpeg" >&2
+  exit 1
+}
+
+TMP=$(mktemp -d)
+trap 'rm -rf "$TMP"' EXIT
+
+FRAMES=$((DUR * FPS))
+INTERVAL=$(python3 -c "print(1 / $FPS)")
+
+for i in $(seq -f '%04g' 1 "$FRAMES"); do
+  # shellcheck disable=SC2086
+  agent-browser $AB_TARGET screenshot "$TMP/frame-$i.png" > /dev/null 2>&1 || true
+  sleep "$INTERVAL"
+done
+
+CAPTURED=$(find "$TMP" -name 'frame-*.png' | wc -l | tr -d ' ')
+[ "$CAPTURED" -gt 0 ] || {
+  echo "no frames captured — is the app reachable via $AB_TARGET?" >&2
+  exit 1
+}
+
+ffmpeg -y -loglevel error -framerate "$FPS" -pattern_type glob -i "$TMP/frame-*.png" \
+  -vf "fps=$FPS,scale=$GIF_WIDTH:-1:flags=lanczos,split[s0][s1];[s0]palettegen[p];[s1][p]paletteuse" \
+  "$OUT"
+
+echo "$OUT ($CAPTURED frames @ ${FPS}fps)"

.agents/skills/agent-testing/scripts/report-init.sh

@@ -0,0 +1,88 @@
+#!/usr/bin/env bash
+# report-init.sh — scaffold a structured test report under .records/reports/.
+#
+# Format spec and evidence rules: ../references/report.md
+#
+# Usage:
+#   report-init.sh <slug> [title]
+#
+# Prints the report directory path (capture it: DIR=$(report-init.sh my-test)).
+
+set -euo pipefail
+
+SLUG="${1:?Usage: report-init.sh <slug> [title]}"
+TITLE="${2:-$SLUG}"
+
+REPO_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/../../../.." && pwd)"
+TS="$(date +%Y%m%d-%H%M%S)"
+DIR="$REPO_ROOT/.records/reports/$TS-$SLUG"
+mkdir -p "$DIR/assets"
+
+BRANCH=$(git -C "$REPO_ROOT" branch --show-current 2> /dev/null || echo "unknown")
+COMMIT=$(git -C "$REPO_ROOT" rev-parse --short HEAD 2> /dev/null || echo "unknown")
+DATE_HUMAN=$(date '+%Y-%m-%d %H:%M')
+DATE_ISO=$(date '+%Y-%m-%dT%H:%M:%S%z')
+
+cat > "$DIR/report.md" << EOF
+# 测试报告：$TITLE
+
+## 范围
+
+<!-- 测试目标 / 变更范围 / 重点风险 -->
+
+- 分支：\`$BRANCH\`
+- 当前提交：\`$COMMIT\`
+- 日期：$DATE_HUMAN
+- 表面：<!-- CLI / Electron + CDP / Web / Bot:<platform> -->
+- 测试页 / 入口：<!-- e.g. /settings or http://localhost:3010 -->
+- 重点：<!-- 本轮最关心的体验、功能或回归点 -->
+
+## 用例
+
+| # | 用例 | 结果 | 关键现象 | 证据 |
+| - | ---- | ---- | -------- | ---- |
+| 1 |      | 待测 |          | ![用例 1](assets/case1.png) |
+
+## 结论
+
+整体结论：\`pending\`。
+
+<!-- 用 1-2 段概括用户最需要知道的结果；失败和阻塞必须明确说明影响。 -->
+
+仍需处理 / 跟进：
+
+- <!-- TODO -->
+
+## 本轮验证
+
+<!-- 如有自动化或命令行验证，保留精简命令与结果；没有则写“未运行额外自动化验证”。 -->
+
+\`\`\`bash
+# command
+\`\`\`
+
+结果：
+
+- <!-- TODO -->
+
+## 评分
+
+- 通过：0
+- 失败：0
+- 阻塞：0
+- 评分：— / 100
+EOF
+
+cat > "$DIR/result.json" << EOF
+{
+  "title": "$TITLE",
+  "createdAt": "$DATE_ISO",
+  "branch": "$BRANCH",
+  "commit": "$COMMIT",
+  "surfaces": [],
+  "cases": [],
+  "summary": { "total": 0, "passed": 0, "failed": 0, "blocked": 0, "verdict": "pending" }
+}
+EOF
+
+echo "$DIR"

.agents/skills/agent-testing/scripts/setup-auth.sh

@@ -0,0 +1,553 @@
+#!/usr/bin/env bash
+# setup-auth.sh — one-stop auth setup & check for local agent testing.
+#
+# Auth is the gate for all automated testing: prepare it BEFORE writing any
+# test step. Background and failure modes: ../references/auth.md
+#
+# Usage:
+#   setup-auth.sh status        # check server + CLI + web + Electron readiness
+#   setup-auth.sh status --surface web  # check only the Web surface gate
+#   setup-auth.sh cli-seed      # configure CLI API-key auth from seeded local env
+#   setup-auth.sh cli           # interactive CLI device-code login (run by a human)
+#   setup-auth.sh open-chrome   # open SERVER_URL in Chrome and show DevTools
+#   setup-auth.sh web-seed      # sign in seeded user and inject cookies automatically
+#   setup-auth.sh web           # stdin = Cookie header -> inject into agent-browser session
+#   setup-auth.sh web-verify    # live-check the agent-browser session is authenticated
+#
+# Env:
+#   SERVER_URL  (default from test-env.sh)        dev server under test
+#   SESSION     (default lobehub-dev)             agent-browser session name
+#   AUTH_DIR    (default ~/.lobehub-agent-testing) where web state is persisted
+#   SEED_EMAIL / SEED_PASSWORD                    seeded better-auth login
+
+set -euo pipefail
+
+REPO_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/../../../.." && pwd)"
+
+workspace_root_for_port() {
+  local root="$REPO_ROOT"
+  local name
+  name="$(basename "$root")"
+
+  if [[ "$name" == "lobehub" ]]; then
+    local parent
+    parent="$(cd "$root/.." && pwd)"
+    local parent_name
+    parent_name="$(basename "$parent")"
+    if [[ "$parent_name" == lobehub-cloud* ]]; then
+      root="$parent"
+    fi
+  fi
+
+  printf '%s\n' "$root"
+}
+
+default_server_url() {
+  local env_resolver resolved
+  env_resolver="$(dirname "${BASH_SOURCE[0]}")/test-env.sh"
+  if [[ -x "$env_resolver" ]]; then
+    resolved="$("$env_resolver" --value SERVER_URL 2> /dev/null || true)"
+    if [[ -n "$resolved" ]]; then
+      printf '%s\n' "$resolved"
+      return 0
+    fi
+  fi
+
+  local root name suffix port
+  root="$(workspace_root_for_port)"
+  name="$(basename "$root")"
+
+  case "$name" in
+    lobehub-cloud)
+      port=3020
+      ;;
+    lobehub-cloud-*)
+      suffix="${name#lobehub-cloud-}"
+      if [[ "$suffix" =~ ^[0-9]+$ ]]; then
+        port=$((3020 + 10#$suffix))
+      else
+        port=3010
+      fi
+      ;;
+    *)
+      port=3010
+      ;;
+  esac
+
+  printf 'http://localhost:%s\n' "$port"
+}
+
+SERVER_URL="${SERVER_URL:-$(default_server_url)}"
+SESSION="${SESSION:-lobehub-dev}"
+AUTH_DIR="${AUTH_DIR:-$HOME/.lobehub-agent-testing}"
+STATE_FILE="$AUTH_DIR/web-state.json"
+CLI_HOME_NAME="${LOBEHUB_CLI_HOME:-.lobehub-dev}"
+CLI_HOME="$HOME/${CLI_HOME_NAME#/}"
+CLI_CREDENTIALS_FILE="$CLI_HOME/credentials.json"
+SEED_EMAIL="${SEED_EMAIL:-agent-testing@lobehub.com}"
+SEED_PASSWORD="${SEED_PASSWORD:-TestPassword123!}"
+SEED_API_KEY="${SEED_API_KEY:-${AGENT_TESTING_API_KEY:-sk-lh-agenttesting0001}}"
+CLI_ENV_FILE="${CLI_ENV_FILE:-$REPO_ROOT/.records/env/agent-testing-cli.env}"
+
+ok()   { printf '  \033[32m✔\033[0m %s\n' "$1"; }
+bad()  { printf '  \033[31m✘\033[0m %s\n' "$1"; }
+note() { printf '      %s\n' "$1"; }
+
+usage() {
+  cat << EOF
+Usage:
+  $0 status [--surface all|cli|web|electron]
+  $0 cli-seed
+  $0 cli
+  $0 open-chrome [--dry-run]
+  $0 web-seed
+  $0 web
+  $0 web-verify
+
+Env:
+  SERVER_URL=$SERVER_URL
+  SESSION=$SESSION
+  AUTH_DIR=$AUTH_DIR
+  SEED_EMAIL=$SEED_EMAIL
+  CLI_HOME=$CLI_HOME
+EOF
+}
+
+check_server() {
+  local code
+  code=$(curl -s -o /dev/null -w '%{http_code}' "$SERVER_URL/" 2> /dev/null || true)
+  if [[ "$code" =~ ^[23] ]]; then
+    ok "dev server reachable at $SERVER_URL"
+  else
+    bad "dev server NOT reachable at $SERVER_URL (http_code='$code')"
+    note "start it: pnpm run dev:next  (see references/dev-server.md)"
+    return 1
+  fi
+}
+
+check_cli() {
+  local api_key="${LOBEHUB_CLI_API_KEY:-${LOBE_API_KEY:-}}"
+  if [[ -n "$api_key" ]]; then
+    local body_file code
+    body_file="$(mktemp)"
+    code=$(curl -sS -o "$body_file" -w '%{http_code}' \
+      -H "Authorization: Bearer $api_key" \
+      "$SERVER_URL/api/v1/users/me?includeCount=0" 2> /dev/null || true)
+
+    if [[ "$code" =~ ^[23] ]]; then
+      rm -f "$body_file"
+      ok "CLI API-key auth valid for $SERVER_URL"
+      return 0
+    fi
+
+    bad "CLI API-key auth failed for $SERVER_URL (http_code='$code')"
+    note "seed the local API key first:"
+    note "./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user"
+    note "source $CLI_ENV_FILE"
+    rm -f "$body_file"
+    return 1
+  fi
+
+  if [[ -f "$CLI_HOME/settings.json" ]] && grep -q "$SERVER_URL" "$CLI_HOME/settings.json" && [[ -f "$CLI_CREDENTIALS_FILE" ]]; then
+    ok "CLI device-code credentials configured for $SERVER_URL (creds: $CLI_HOME)"
+  else
+    bad "CLI not logged in to $SERVER_URL"
+    note "automated path:"
+    note "./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user && source $CLI_ENV_FILE && $0 cli-seed"
+    note "interactive fallback:"
+    note "cd apps/cli && LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts login --server $SERVER_URL"
+    return 1
+  fi
+}
+
+check_web() {
+  if [[ -f "$STATE_FILE" ]]; then
+    ok "web auth state saved ($STATE_FILE)"
+  else
+    bad "no web auth state for agent-browser"
+    note "for the seeded local user, run: $0 web-seed"
+    note "or copy the Cookie header from Chrome DevTools (Network tab), then:"
+    note "pbpaste | $0 web   (see references/auth.md)"
+    return 1
+  fi
+  cmd_web_verify --skip-server-check
+}
+
+check_agent_browser() {
+  if command -v agent-browser > /dev/null 2>&1; then
+    ok "agent-browser available"
+  else
+    bad "agent-browser command not found"
+    note "install or expose agent-browser before Web/Electron UI testing"
+    return 1
+  fi
+}
+
+check_electron() {
+  local cdp_port="${CDP_PORT:-9222}"
+  if ! curl -s -o /dev/null --max-time 2 "http://localhost:$cdp_port/json/version" 2> /dev/null; then
+    note "electron: not running (CDP $cdp_port unreachable) — start with electron-dev.sh; check skipped"
+    return 0
+  fi
+  local probe result
+  probe="$(dirname "${BASH_SOURCE[0]}")/app-probe.sh"
+  result=$(bash "$probe" auth 2> /dev/null || true)
+  # agent-browser eval returns the JSON string with escaped quotes — normalize.
+  result="${result//\\/}"
+  if [[ "$result" == *'"isSignedIn":true'* ]]; then
+    ok "electron app signed in ($result)"
+  else
+    bad "electron app NOT signed in ($result)"
+    note "log in once manually inside the app (state persists across restarts)"
+    return 1
+  fi
+}
+
+cmd_status() {
+  local surface="all"
+  while [[ $# -gt 0 ]]; do
+    case "$1" in
+      --surface)
+        if [[ $# -lt 2 ]]; then
+          echo "--surface requires one of: all, cli, web, electron" >&2
+          return 2
+        fi
+        surface="${2:-}"
+        shift 2
+        ;;
+      --surface=*)
+        surface="${1#*=}"
+        shift
+        ;;
+      all|cli|web|electron)
+        surface="$1"
+        shift
+        ;;
+      -h|--help)
+        usage
+        return 0
+        ;;
+      *)
+        echo "unknown status option: $1" >&2
+        usage >&2
+        return 2
+        ;;
+    esac
+  done
+
+  case "$surface" in
+    all|cli|web|electron) ;;
+    "")
+      echo "--surface requires one of: all, cli, web, electron" >&2
+      return 2
+      ;;
+    *)
+      echo "unknown surface: $surface" >&2
+      usage >&2
+      return 2
+      ;;
+  esac
+
+  echo "agent-testing auth status (surface=$surface, SERVER_URL=$SERVER_URL):"
+  local rc=0
+  case "$surface" in
+    all)
+      check_server || rc=1
+      check_cli || rc=1
+      check_web || rc=1
+      check_electron || rc=1
+      ;;
+    cli)
+      check_server || rc=1
+      check_cli || rc=1
+      ;;
+    web)
+      check_server || rc=1
+      check_web || rc=1
+      ;;
+    electron)
+      check_electron || rc=1
+      ;;
+  esac
+  if [[ $rc -eq 0 ]]; then
+    echo "$surface auth green — safe to start automated testing on this surface."
+  else
+    echo "$surface auth NOT ready — fix the ✘ items before writing any test step."
+  fi
+  return $rc
+}
+
+cmd_cli() {
+  echo "Starting CLI device-code login against $SERVER_URL ..."
+  echo "(opens a browser authorization — must be run by a human in a terminal)"
+  cd "$REPO_ROOT/apps/cli"
+  LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts login --server "$SERVER_URL"
+}
+
+write_cli_seed_env() {
+  mkdir -p "$(dirname "$CLI_ENV_FILE")"
+  cat > "$CLI_ENV_FILE" << EOF
+# Source this file before running LobeHub CLI agent tests.
+# Generated by setup-auth.sh cli-seed
+export LOBE_API_KEY=$SEED_API_KEY
+export LOBEHUB_CLI_API_KEY="\${LOBE_API_KEY}"
+export LOBEHUB_SERVER=$SERVER_URL
+export LOBEHUB_CLI_HOME=.lobehub-dev
+EOF
+}
+
+write_cli_settings() {
+  mkdir -p "$CLI_HOME"
+  python3 - "$CLI_HOME/settings.json" "$SERVER_URL" << 'PY'
+import json
+import os
+import sys
+
+path, server_url = sys.argv[1], sys.argv[2]
+os.makedirs(os.path.dirname(path), exist_ok=True)
+with open(path, "w") as f:
+    json.dump({"serverUrl": server_url}, f, indent=2)
+    f.write("\n")
+os.chmod(path, 0o600)
+PY
+}
+
+cmd_cli_seed() {
+  check_server || return 1
+  write_cli_seed_env
+  write_cli_settings
+  ok "wrote CLI seed env: $CLI_ENV_FILE"
+  note "source it before CLI commands: source $CLI_ENV_FILE"
+  note "settings saved at: $CLI_HOME/settings.json"
+  LOBE_API_KEY="$SEED_API_KEY" LOBEHUB_CLI_API_KEY="$SEED_API_KEY" check_cli
+}
+
+cmd_open_chrome() {
+  local mode="${1:-}"
+  if [[ "$mode" != "" && "$mode" != "--dry-run" ]]; then
+    echo "unknown open-chrome option: $mode" >&2
+    usage >&2
+    return 2
+  fi
+
+  if [[ "$mode" == "--dry-run" ]]; then
+    echo "would open Google Chrome at $SERVER_URL/"
+    echo "would press Cmd+Option+I to open DevTools"
+    echo "would open DevTools command menu and run 'Show Network'"
+    return 0
+  fi
+
+  if [[ "$(uname -s)" != "Darwin" ]]; then
+    bad "open-chrome is macOS-only"
+    note "open $SERVER_URL/ in your browser and open DevTools manually"
+    return 1
+  fi
+
+  if ! command -v osascript > /dev/null 2>&1; then
+    bad "osascript not found"
+    note "open $SERVER_URL/ in Chrome and press Cmd+Option+I manually"
+    return 1
+  fi
+
+  SERVER_URL="$SERVER_URL" osascript << 'OSA'
+set targetUrl to (system attribute "SERVER_URL") & "/"
+
+tell application "Google Chrome"
+  activate
+  if (count of windows) = 0 then
+    make new window
+  end if
+  tell front window to make new tab with properties {URL:targetUrl}
+end tell
+
+delay 1
+
+tell application "System Events"
+  tell process "Google Chrome"
+    set frontmost to true
+    keystroke "i" using {command down, option down}
+    delay 1
+    keystroke "p" using {command down, shift down}
+    delay 0.2
+    keystroke "Show Network"
+    key code 36
+  end tell
+end tell
+OSA
+  ok "opened Chrome at $SERVER_URL/ and requested DevTools Network panel"
+}
+
+cookie_header_from_jar() {
+  local jar="$1"
+  awk '
+    BEGIN { first = 1 }
+    /^$/ { next }
+    /^#/ {
+      if ($0 !~ /^#HttpOnly_/) next
+      sub(/^#HttpOnly_/, "")
+    }
+    NF >= 7 {
+      if (!first) printf "; "
+      printf "%s=%s", $6, $7
+      first = 0
+    }
+    END {
+      if (!first) printf "\n"
+    }
+  ' "$jar"
+}
+
+# Build a Playwright storageState file from a raw Cookie header on stdin,
+# keeping only the better-auth cookies. See references/auth.md for why the
+# header must come from a Network request (HttpOnly) and why httpOnly=false.
+cmd_web() {
+  mkdir -p "$AUTH_DIR"
+  local raw
+  raw="$(cat)"
+  COOKIE_INPUT="$raw" python3 - "$STATE_FILE" << 'PY'
+import json, os, sys, time
+
+raw = os.environ.get("COOKIE_INPUT", "").strip()
+cookie_lines = []
+for line in raw.splitlines():
+    stripped = line.strip()
+    if not stripped:
+        continue
+    if stripped.lower().startswith("cookie:"):
+        cookie_lines.append(stripped.split(":", 1)[1].strip())
+    else:
+        cookie_lines.append(stripped)
+
+raw = "; ".join(cookie_lines)
+
+WANTED = {"better-auth.session_token", "better-auth.session_data", "better-auth.state"}
+exp = int(time.time()) + 30 * 24 * 3600  # 30 days
+
+cookies = []
+for pair in raw.split(";"):
+    pair = pair.strip()
+    if "=" not in pair:
+        continue
+    name, _, value = pair.partition("=")
+    if name not in WANTED:
+        continue
+    cookies.append({
+        "name": name,
+        "value": value,
+        "domain": "localhost",
+        "path": "/",
+        "expires": exp,
+        "httpOnly": False,
+        "secure": False,
+        "sameSite": "Lax",
+    })
+
+if not cookies:
+    sys.stderr.write("no better-auth cookies found in input — paste the raw Cookie header from a Network request\n")
+    sys.exit(1)
+
+with open(sys.argv[1], "w") as f:
+    json.dump({"cookies": cookies, "origins": []}, f, indent=2)
+print(f"wrote {len(cookies)} cookie(s) to {sys.argv[1]}")
+PY
+  cmd_web_verify
+}
+
+cmd_web_seed() {
+  check_server || return 1
+  mkdir -p "$AUTH_DIR"
+
+  local cookie_jar="$AUTH_DIR/web-seed-cookie.jar"
+  local response_body="$AUTH_DIR/web-seed-response.json"
+  local payload code
+  payload="$(
+    SEED_EMAIL="$SEED_EMAIL" SEED_PASSWORD="$SEED_PASSWORD" python3 - << 'PY'
+import json
+import os
+
+print(json.dumps({
+    "callbackURL": "/",
+    "email": os.environ["SEED_EMAIL"],
+    "password": os.environ["SEED_PASSWORD"],
+}))
+PY
+  )"
+
+  code=$(curl -sS -o "$response_body" -w '%{http_code}' \
+    -c "$cookie_jar" \
+    -H 'Content-Type: application/json' \
+    -X POST "$SERVER_URL/api/auth/sign-in/email" \
+    --data "$payload" 2> /dev/null || true)
+
+  if [[ ! "$code" =~ ^[23] ]]; then
+    bad "seed user sign-in failed at $SERVER_URL/api/auth/sign-in/email (http_code='$code')"
+    note "make sure the seed user exists:"
+    note "./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user"
+    return 1
+  fi
+
+  local cookie_header
+  cookie_header="$(cookie_header_from_jar "$cookie_jar")"
+  if [[ -z "$cookie_header" ]]; then
+    bad "seed sign-in succeeded but no cookies were written to $cookie_jar"
+    return 1
+  fi
+
+  printf '%s\n' "$cookie_header" | cmd_web
+}
+
+cmd_web_verify() {
+  local skip_server_check="${1:-}"
+  if [[ "$skip_server_check" != "--skip-server-check" ]]; then
+    check_server || return 1
+  fi
+  if [[ ! -f "$STATE_FILE" ]]; then
+    bad "no web auth state for agent-browser"
+    note "for the seeded local user, run: $0 web-seed"
+    note "or copy the Cookie header from Chrome DevTools (Network tab), then:"
+    note "pbpaste | $0 web"
+    return 1
+  fi
+  check_agent_browser || return 1
+  if ! agent-browser --session "$SESSION" state load "$STATE_FILE" > /dev/null; then
+    bad "failed to load web auth state into agent-browser session '$SESSION'"
+    return 1
+  fi
+  if ! agent-browser --session "$SESSION" open "$SERVER_URL/" > /dev/null; then
+    bad "failed to open $SERVER_URL in agent-browser session '$SESSION'"
+    return 1
+  fi
+  local url
+  url=$(agent-browser --session "$SESSION" get url 2> /dev/null || true)
+  if [[ -z "$url" ]]; then
+    bad "agent-browser session '$SESSION' did not report a current URL"
+    return 1
+  fi
+  if [[ "$url" == *"/signin"* || "$url" == *"/login"* ]]; then
+    bad "agent-browser session '$SESSION' NOT authenticated (landed on $url)"
+    note "re-copy the Cookie header and re-run: pbpaste | $0 web"
+    return 1
+  fi
+  ok "agent-browser session '$SESSION' authenticated (at $url)"
+}
+
+case "${1:-status}" in
+  status)
+    shift || true
+    cmd_status "$@"
+    ;;
+  cli-seed) cmd_cli_seed ;;
+  cli) cmd_cli ;;
+  open-chrome)
+    shift || true
+    cmd_open_chrome "$@"
+    ;;
+  web-seed) cmd_web_seed ;;
+  web) cmd_web ;;
+  web-verify) cmd_web_verify ;;
+  -h|--help) usage ;;
+  *)
+    echo "Usage: $0 {status|cli-seed|cli|open-chrome|web-seed|web|web-verify}" >&2
+    exit 2
+    ;;
+esac

.agents/skills/agent-testing/scripts/setup-auth.test.sh

@@ -0,0 +1,197 @@
+#!/usr/bin/env bash
+# Smoke tests for setup-auth.sh. Uses a temporary agent-browser stub and local
+# HTTP server, so it does not need real browser auth.
+
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SCRIPT="$SCRIPT_DIR/setup-auth.sh"
+
+fail() {
+  echo "FAIL: $*" >&2
+  exit 1
+}
+
+assert_contains() {
+  local file="$1"
+  local text="$2"
+  grep -Fq "$text" "$file" || fail "expected '$text' in $file"
+}
+
+tmp_dir="$(mktemp -d)"
+server_pid=""
+
+cleanup() {
+  if [[ -n "$server_pid" ]]; then
+    kill "$server_pid" > /dev/null 2>&1 || true
+    wait "$server_pid" > /dev/null 2>&1 || true
+  fi
+  rm -rf "$tmp_dir"
+}
+trap cleanup EXIT
+export HOME="$tmp_dir/home"
+
+port="$(python3 - << 'PY'
+import socket
+
+sock = socket.socket()
+sock.bind(("127.0.0.1", 0))
+print(sock.getsockname()[1])
+sock.close()
+PY
+)"
+
+python3 - "$port" << 'PY' > "$tmp_dir/http.log" 2>&1 &
+from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
+import sys
+
+
+class Handler(BaseHTTPRequestHandler):
+    def do_GET(self):
+        if self.path.startswith("/api/v1/users/me"):
+            if self.headers.get("authorization") != "Bearer sk-lh-agenttesting0001":
+                self.send_response(401)
+                self.end_headers()
+                self.wfile.write(b'{"success":false}')
+                return
+
+            self.send_response(200)
+            self.send_header("Content-Type", "application/json")
+            self.end_headers()
+            self.wfile.write(b'{"success":true,"data":{"id":"user_agent_testing_001"}}')
+            return
+
+        self.send_response(200)
+        self.end_headers()
+        self.wfile.write(b"ok")
+
+    def do_POST(self):
+        length = int(self.headers.get("content-length") or "0")
+        if length:
+            self.rfile.read(length)
+
+        if self.path != "/api/auth/sign-in/email":
+            self.send_response(404)
+            self.end_headers()
+            return
+
+        self.send_response(200)
+        self.send_header(
+            "Set-Cookie",
+            "better-auth.session_token=seed.token; Path=/; HttpOnly; SameSite=Lax",
+        )
+        self.send_header(
+            "Set-Cookie",
+            "better-auth.session_data=seed.data; Path=/; HttpOnly; SameSite=Lax",
+        )
+        self.send_header("Content-Type", "application/json")
+        self.end_headers()
+        self.wfile.write(b'{"ok":true}')
+
+    def log_message(self, format, *args):
+        return
+
+
+ThreadingHTTPServer(("localhost", int(sys.argv[1])), Handler).serve_forever()
+PY
+server_pid="$!"
+
+server_url="http://localhost:$port"
+for _ in {1..50}; do
+  if curl -s -o /dev/null "$server_url/"; then
+    break
+  fi
+  sleep 0.1
+done
+curl -s -o /dev/null "$server_url/" || fail "test HTTP server did not start"
+
+mkdir -p "$tmp_dir/bin" "$tmp_dir/auth"
+cat > "$tmp_dir/bin/agent-browser" << 'SH'
+#!/usr/bin/env bash
+set -euo pipefail
+
+if [[ "${1:-}" == "--session" ]]; then
+  shift 2
+fi
+
+case "${1:-}" in
+  state)
+    [[ "${2:-}" == "load" ]] || exit 2
+    [[ -f "${3:-}" ]] || exit 1
+    ;;
+  open)
+    printf '%s\n' "${2:-}" > "${AGENT_BROWSER_URL_FILE:?}"
+    ;;
+  get)
+    [[ "${2:-}" == "url" ]] || exit 2
+    cat "${AGENT_BROWSER_URL_FILE:?}"
+    ;;
+  *)
+    echo "unexpected agent-browser command: $*" >&2
+    exit 2
+    ;;
+esac
+SH
+chmod +x "$tmp_dir/bin/agent-browser"
+
+export PATH="$tmp_dir/bin:$PATH"
+export AUTH_DIR="$tmp_dir/auth"
+export SESSION="setup-auth-test"
+export SERVER_URL="$server_url"
+export AGENT_BROWSER_URL_FILE="$tmp_dir/current-url"
+
+cookie_header="Cookie: foo=bar; better-auth.session_token=test.token; better-auth.session_data=encoded%3D; theme=dark"
+printf '%s\n' "$cookie_header" | "$SCRIPT" web > "$tmp_dir/web.out"
+
+python3 - "$AUTH_DIR/web-state.json" << 'PY'
+import json, sys
+
+with open(sys.argv[1]) as f:
+    state = json.load(f)
+
+names = {cookie["name"] for cookie in state["cookies"]}
+expected = {"better-auth.session_token", "better-auth.session_data"}
+if names != expected:
+    raise SystemExit(f"unexpected cookies: {sorted(names)}")
+PY
+
+"$SCRIPT" web-seed > "$tmp_dir/web-seed.out"
+
+python3 - "$AUTH_DIR/web-state.json" << 'PY'
+import json, sys
+
+with open(sys.argv[1]) as f:
+    state = json.load(f)
+
+values = {cookie["name"]: cookie["value"] for cookie in state["cookies"]}
+expected = {
+    "better-auth.session_token": "seed.token",
+    "better-auth.session_data": "seed.data",
+}
+if values != expected:
+    raise SystemExit(f"unexpected seeded cookies: {values}")
+PY
+
+"$SCRIPT" status --surface web > "$tmp_dir/status.out"
+assert_contains "$tmp_dir/status.out" "surface=web"
+assert_contains "$tmp_dir/status.out" "web auth green"
+
+"$SCRIPT" cli-seed > "$tmp_dir/cli-seed.out"
+assert_contains "$tmp_dir/cli-seed.out" "CLI API-key auth valid"
+assert_contains "$tmp_dir/cli-seed.out" "settings saved at: $HOME/.lobehub-dev/settings.json"
+
+if "$SCRIPT" status --surface cli > "$tmp_dir/cli-no-env.out"; then
+  fail "cli status without API key unexpectedly passed"
+fi
+assert_contains "$tmp_dir/cli-no-env.out" "CLI not logged in"
+
+LOBEHUB_CLI_API_KEY=sk-lh-agenttesting0001 "$SCRIPT" status --surface cli > "$tmp_dir/cli-status.out"
+assert_contains "$tmp_dir/cli-status.out" "CLI API-key auth valid"
+assert_contains "$tmp_dir/cli-status.out" "cli auth green"
+
+if printf 'foo=bar\n' | "$SCRIPT" web > "$tmp_dir/invalid.out" 2> "$tmp_dir/invalid.err"; then
+  fail "invalid cookie unexpectedly passed"
+fi
+assert_contains "$tmp_dir/invalid.err" "no better-auth cookies found"
+
+echo "setup-auth tests passed"

.agents/skills/agent-testing/scripts/test-env.sh

@@ -0,0 +1,377 @@
+#!/usr/bin/env bash
+# Print the resolved local test environment for agent-testing.
+#
+# This is intentionally read-only. It mirrors scripts/runWithEnv.mts precedence:
+# .env -> .env.$NODE_ENV -> .env.local -> .env.$NODE_ENV.local, then shell env.
+
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "$SCRIPT_DIR/../../../.." && pwd)"
+NODE_ENV="${NODE_ENV:-development}"
+
+VALUE_APP_URL=""
+VALUE_PORT=""
+VALUE_SERVER_URL=""
+VALUE_AUTH_TRUSTED_ORIGINS=""
+VALUE_SPA_PORT=""
+VALUE_MOBILE_SPA_PORT=""
+VALUE_DESKTOP_PORT=""
+
+SOURCE_APP_URL=""
+SOURCE_PORT=""
+SOURCE_SERVER_URL=""
+SOURCE_AUTH_TRUSTED_ORIGINS=""
+SOURCE_SPA_PORT=""
+SOURCE_MOBILE_SPA_PORT=""
+SOURCE_DESKTOP_PORT=""
+
+LOADED_ENV_FILES=""
+
+keys() {
+  printf '%s\n' \
+    APP_URL \
+    PORT \
+    SERVER_URL \
+    AUTH_TRUSTED_ORIGINS \
+    SPA_PORT \
+    MOBILE_SPA_PORT \
+    DESKTOP_PORT
+}
+
+trim() {
+  local value="$1"
+  value="${value#"${value%%[![:space:]]*}"}"
+  value="${value%"${value##*[![:space:]]}"}"
+  printf '%s' "$value"
+}
+
+workspace_root() {
+  local root="$REPO_ROOT"
+  local name
+  name="$(basename "$root")"
+
+  if [[ "$name" == "lobehub" ]]; then
+    local parent parent_name
+    parent="$(cd "$root/.." && pwd)"
+    parent_name="$(basename "$parent")"
+    if [[ "$parent_name" == lobehub-cloud* ]]; then
+      root="$parent"
+    fi
+  fi
+
+  printf '%s\n' "$root"
+}
+
+workspace_offset() {
+  local name="$1"
+
+  case "$name" in
+    lobehub-cloud)
+      printf '0\n'
+      ;;
+    lobehub-cloud-*)
+      local suffix="${name#lobehub-cloud-}"
+      if [[ "$suffix" =~ ^[0-9]+$ ]]; then
+        printf '%s\n' "$((10#$suffix))"
+      else
+        printf '\n'
+      fi
+      ;;
+    *)
+      printf '\n'
+      ;;
+  esac
+}
+
+default_port() {
+  local base="$1"
+  local fallback="$2"
+  local root name offset
+  root="$(workspace_root)"
+  name="$(basename "$root")"
+  offset="$(workspace_offset "$name")"
+
+  if [[ -n "$offset" ]]; then
+    printf '%s\n' "$((base + offset))"
+  else
+    printf '%s\n' "$fallback"
+  fi
+}
+
+url_port() {
+  local url="$1"
+  local hostport
+  hostport="${url#*://}"
+  hostport="${hostport%%/*}"
+
+  if [[ "$hostport" == *:* ]]; then
+    local port="${hostport##*:}"
+    if [[ "$port" =~ ^[0-9]+$ ]]; then
+      printf '%s\n' "$port"
+      return 0
+    fi
+  fi
+
+  return 1
+}
+
+url_origin() {
+  local url="$1"
+  local scheme rest hostport
+  if [[ "$url" == *"://"* ]]; then
+    scheme="${url%%://*}"
+    rest="${url#*://}"
+    hostport="${rest%%/*}"
+    printf '%s://%s\n' "$scheme" "$hostport"
+  else
+    printf '%s\n' "$url"
+  fi
+}
+
+set_value() {
+  local key="$1"
+  local value="$2"
+  local source="$3"
+
+  case "$key" in
+    APP_URL) VALUE_APP_URL="$value"; SOURCE_APP_URL="$source" ;;
+    PORT) VALUE_PORT="$value"; SOURCE_PORT="$source" ;;
+    SERVER_URL) VALUE_SERVER_URL="$value"; SOURCE_SERVER_URL="$source" ;;
+    AUTH_TRUSTED_ORIGINS) VALUE_AUTH_TRUSTED_ORIGINS="$value"; SOURCE_AUTH_TRUSTED_ORIGINS="$source" ;;
+    SPA_PORT) VALUE_SPA_PORT="$value"; SOURCE_SPA_PORT="$source" ;;
+    MOBILE_SPA_PORT) VALUE_MOBILE_SPA_PORT="$value"; SOURCE_MOBILE_SPA_PORT="$source" ;;
+    DESKTOP_PORT) VALUE_DESKTOP_PORT="$value"; SOURCE_DESKTOP_PORT="$source" ;;
+  esac
+}
+
+value_for() {
+  case "$1" in
+    APP_URL) printf '%s\n' "$VALUE_APP_URL" ;;
+    PORT) printf '%s\n' "$VALUE_PORT" ;;
+    SERVER_URL) printf '%s\n' "$VALUE_SERVER_URL" ;;
+    AUTH_TRUSTED_ORIGINS) printf '%s\n' "$VALUE_AUTH_TRUSTED_ORIGINS" ;;
+    SPA_PORT) printf '%s\n' "$VALUE_SPA_PORT" ;;
+    MOBILE_SPA_PORT) printf '%s\n' "$VALUE_MOBILE_SPA_PORT" ;;
+    DESKTOP_PORT) printf '%s\n' "$VALUE_DESKTOP_PORT" ;;
+  esac
+}
+
+source_for() {
+  case "$1" in
+    APP_URL) printf '%s\n' "$SOURCE_APP_URL" ;;
+    PORT) printf '%s\n' "$SOURCE_PORT" ;;
+    SERVER_URL) printf '%s\n' "$SOURCE_SERVER_URL" ;;
+    AUTH_TRUSTED_ORIGINS) printf '%s\n' "$SOURCE_AUTH_TRUSTED_ORIGINS" ;;
+    SPA_PORT) printf '%s\n' "$SOURCE_SPA_PORT" ;;
+    MOBILE_SPA_PORT) printf '%s\n' "$SOURCE_MOBILE_SPA_PORT" ;;
+    DESKTOP_PORT) printf '%s\n' "$SOURCE_DESKTOP_PORT" ;;
+  esac
+}
+
+is_tracked_key() {
+  case "$1" in
+    APP_URL|PORT|SERVER_URL|AUTH_TRUSTED_ORIGINS|SPA_PORT|MOBILE_SPA_PORT|DESKTOP_PORT) return 0 ;;
+    *) return 1 ;;
+  esac
+}
+
+parse_env_file() {
+  local file="$1"
+  local root="$2"
+  local label="${file#$root/}"
+  local line key value
+
+  [[ -f "$file" ]] || return 0
+  if [[ -z "$LOADED_ENV_FILES" ]]; then
+    LOADED_ENV_FILES="$label"
+  else
+    LOADED_ENV_FILES="$LOADED_ENV_FILES, $label"
+  fi
+
+  while IFS= read -r line || [[ -n "$line" ]]; do
+    line="$(trim "$line")"
+    [[ -z "$line" || "$line" == \#* ]] && continue
+
+    if [[ "$line" == export[[:space:]]* ]]; then
+      line="$(trim "${line#export}")"
+    fi
+
+    [[ "$line" == *=* ]] || continue
+    key="$(trim "${line%%=*}")"
+    value="$(trim "${line#*=}")"
+    is_tracked_key "$key" || continue
+
+    if [[ "$value" == \"*\" && "$value" == *\" && ${#value} -ge 2 ]]; then
+      value="${value:1:${#value}-2}"
+    elif [[ "$value" == \'* && "$value" == *\' && ${#value} -ge 2 ]]; then
+      value="${value:1:${#value}-2}"
+    fi
+
+    set_value "$key" "$value" "$label"
+  done < "$file"
+}
+
+apply_env_files() {
+  local root="$1"
+  parse_env_file "$root/.env" "$root"
+  parse_env_file "$root/.env.$NODE_ENV" "$root"
+  parse_env_file "$root/.env.local" "$root"
+  parse_env_file "$root/.env.$NODE_ENV.local" "$root"
+}
+
+apply_shell_overrides() {
+  local key value
+  while IFS= read -r key; do
+    if [[ -n "${!key+x}" ]]; then
+      value="${!key}"
+      set_value "$key" "$value" "shell"
+    fi
+  done < <(keys)
+}
+
+resolve_defaults() {
+  local app_port spa_port mobile_spa_port desktop_port
+  app_port="$(default_port 3020 3010)"
+  spa_port="$(default_port 9800 9876)"
+  mobile_spa_port="$(default_port 3810 3012)"
+  desktop_port="$(default_port 3030 3015)"
+
+  if [[ -z "$VALUE_APP_URL" ]]; then
+    set_value APP_URL "http://localhost:$app_port" "inferred"
+  fi
+
+  if [[ -z "$VALUE_PORT" ]]; then
+    if app_port="$(url_port "$VALUE_APP_URL")"; then
+      set_value PORT "$app_port" "inferred from APP_URL"
+    else
+      set_value PORT "$(default_port 3020 3010)" "inferred"
+    fi
+  fi
+
+  if [[ -z "$VALUE_SERVER_URL" ]]; then
+    set_value SERVER_URL "$VALUE_APP_URL" "from APP_URL"
+  fi
+
+  if [[ -z "$VALUE_SPA_PORT" ]]; then
+    set_value SPA_PORT "$spa_port" "inferred"
+  fi
+
+  if [[ -z "$VALUE_MOBILE_SPA_PORT" ]]; then
+    set_value MOBILE_SPA_PORT "$mobile_spa_port" "inferred"
+  fi
+
+  if [[ -z "$VALUE_DESKTOP_PORT" ]]; then
+    set_value DESKTOP_PORT "$desktop_port" "inferred"
+  fi
+
+  if [[ -z "$VALUE_AUTH_TRUSTED_ORIGINS" ]]; then
+    set_value AUTH_TRUSTED_ORIGINS "$(url_origin "$VALUE_APP_URL"),http://localhost:$VALUE_SPA_PORT" "inferred"
+  fi
+}
+
+contains_origin() {
+  local list="$1"
+  local expected="$2"
+  local item
+  IFS=',' read -r -a items <<< "$list"
+  for item in "${items[@]}"; do
+    item="$(trim "$item")"
+    [[ "$item" == "$expected" ]] && return 0
+  done
+  return 1
+}
+
+print_exports() {
+  local key value
+  while IFS= read -r key; do
+    value="$(value_for "$key")"
+    printf 'export %s=%q\n' "$key" "$value"
+  done < <(keys)
+}
+
+print_value() {
+  local key="$1"
+  if ! is_tracked_key "$key"; then
+    echo "unknown key: $key" >&2
+    exit 2
+  fi
+  value_for "$key"
+}
+
+print_human() {
+  local root="$1"
+  local key value source
+
+  echo "agent-testing test env:"
+  printf '  workspace: %s\n' "$root"
+  printf '  NODE_ENV: %s\n' "$NODE_ENV"
+  printf '  env files: %s\n' "${LOADED_ENV_FILES:-none}"
+  echo
+  echo "resolved values:"
+  while IFS= read -r key; do
+    value="$(value_for "$key")"
+    source="$(source_for "$key")"
+    printf '  %-22s %s  (%s)\n' "$key=$value" "" "$source"
+  done < <(keys)
+  echo
+  echo "checks:"
+
+  local app_origin spa_origin app_port
+  app_origin="$(url_origin "$VALUE_APP_URL")"
+  spa_origin="http://localhost:$VALUE_SPA_PORT"
+  if app_port="$(url_port "$VALUE_APP_URL")" && [[ "$app_port" == "$VALUE_PORT" ]]; then
+    printf '  OK   PORT matches APP_URL (%s)\n' "$VALUE_PORT"
+  else
+    printf '  WARN PORT (%s) does not match APP_URL (%s)\n' "$VALUE_PORT" "$VALUE_APP_URL"
+  fi
+
+  if contains_origin "$VALUE_AUTH_TRUSTED_ORIGINS" "$app_origin"; then
+    printf '  OK   AUTH_TRUSTED_ORIGINS includes %s\n' "$app_origin"
+  else
+    printf '  WARN AUTH_TRUSTED_ORIGINS is missing %s\n' "$app_origin"
+  fi
+
+  if contains_origin "$VALUE_AUTH_TRUSTED_ORIGINS" "$spa_origin"; then
+    printf '  OK   AUTH_TRUSTED_ORIGINS includes %s\n' "$spa_origin"
+  else
+    printf '  WARN AUTH_TRUSTED_ORIGINS is missing %s\n' "$spa_origin"
+  fi
+}
+
+usage() {
+  cat << EOF
+Usage:
+  $0                 # print resolved test environment
+  $0 --exports       # print source-able export lines
+  $0 --value KEY     # print one resolved value
+
+Tracked keys:
+  APP_URL PORT SERVER_URL AUTH_TRUSTED_ORIGINS SPA_PORT MOBILE_SPA_PORT DESKTOP_PORT
+EOF
+}
+
+ROOT="$(workspace_root)"
+apply_env_files "$ROOT"
+apply_shell_overrides
+resolve_defaults
+
+case "${1:-}" in
+  "")
+    print_human "$ROOT"
+    ;;
+  --exports)
+    print_exports
+    ;;
+  --value)
+    print_value "${2:-}"
+    ;;
+  -h|--help)
+    usage
+    ;;
+  *)
+    echo "unknown option: $1" >&2
+    usage >&2
+    exit 2
+    ;;
+esac

.agents/skills/agent-testing/scripts/test-env.test.sh

@@ -0,0 +1,57 @@
+#!/usr/bin/env bash
+# Smoke tests for test-env.sh.
+
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+fail() {
+  echo "FAIL: $*" >&2
+  exit 1
+}
+
+assert_eq() {
+  local actual="$1"
+  local expected="$2"
+  [[ "$actual" == "$expected" ]] || fail "expected '$expected', got '$actual'"
+}
+
+assert_contains() {
+  local file="$1"
+  local text="$2"
+  grep -Fq "$text" "$file" || fail "expected '$text' in $file"
+}
+
+tmp_dir="$(mktemp -d)"
+trap 'rm -rf "$tmp_dir"' EXIT
+
+mkdir -p "$tmp_dir/lobehub-cloud-1/.agents/skills" "$tmp_dir/lobehub/.agents/skills"
+ln -s "$SCRIPT_DIR/.." "$tmp_dir/lobehub-cloud-1/.agents/skills/agent-testing"
+ln -s "$SCRIPT_DIR/.." "$tmp_dir/lobehub/.agents/skills/agent-testing"
+
+cloud_script="$tmp_dir/lobehub-cloud-1/.agents/skills/agent-testing/scripts/test-env.sh"
+oss_script="$tmp_dir/lobehub/.agents/skills/agent-testing/scripts/test-env.sh"
+
+assert_eq "$("$cloud_script" --value SERVER_URL)" "http://localhost:3021"
+assert_eq "$("$cloud_script" --value SPA_PORT)" "9801"
+assert_eq "$("$cloud_script" --value MOBILE_SPA_PORT)" "3811"
+assert_eq "$("$cloud_script" --value DESKTOP_PORT)" "3031"
+assert_eq "$("$oss_script" --value SERVER_URL)" "http://localhost:3010"
+
+cat > "$tmp_dir/lobehub-cloud-1/.env" << 'EOF'
+APP_URL=http://localhost:4123
+PORT=4123
+AUTH_TRUSTED_ORIGINS=http://localhost:4123,http://localhost:9823
+SPA_PORT=9823
+MOBILE_SPA_PORT=3823
+DESKTOP_PORT=3043
+EOF
+
+assert_eq "$("$cloud_script" --value SERVER_URL)" "http://localhost:4123"
+assert_eq "$("$cloud_script" --value SPA_PORT)" "9823"
+"$cloud_script" --exports > "$tmp_dir/exports.out"
+assert_contains "$tmp_dir/exports.out" "export APP_URL=http://localhost:4123"
+assert_contains "$tmp_dir/exports.out" "export SERVER_URL=http://localhost:4123"
+assert_contains "$tmp_dir/exports.out" "export AUTH_TRUSTED_ORIGINS=http://localhost:4123\\,http://localhost:9823"
+
+echo "test-env tests passed"

.agents/skills/agent-testing/ui/electron.md

@@ -0,0 +1,154 @@
+# Electron (LobeHub Desktop) UI Testing
+
+Default surface for verifying **pure frontend changes** (components, store logic, styles, interactions) in the primary product shape. Drives the Electron renderer over CDP with `agent-browser` — see [../references/agent-browser.md](../references/agent-browser.md) for the full command reference.
+
+**Auth**: the Electron app keeps its own persistent login state — log in once manually in the app; sessions survive restarts. Run `../scripts/setup-auth.sh status` before testing (see [../references/auth.md](../references/auth.md)).
+
+**Linux / headless (cloud)**: Electron itself runs on Linux, but it has no true headless mode — it needs a display server. In a headless environment wrap the launch with `xvfb-run` (virtual framebuffer). Everything CDP-based keeps working under Xvfb: the `agent-browser --cdp 9222` connection, snapshots, eval, and `agent-browser screenshot` (captured from the renderer via CDP, not the OS screen). What does NOT work on Linux: `capture-app-window.sh` (macOS `screencapture`), osascript, and the ffmpeg recording scripts in their current form.
+
+### Setup / Teardown
+
+Use the `electron-dev.sh` script to manage the Electron dev environment. It handles process lifecycle, waits for SPA readiness, and reliably kills all child processes (main + helpers + vite).
+
+```bash
+SCRIPT=".agents/skills/agent-testing/scripts/electron-dev.sh"
+
+# Start Electron dev with CDP (idempotent — skips if already running)
+$SCRIPT start
+
+# Check if Electron is running and CDP is reachable
+$SCRIPT status
+
+# Kill all Electron-related processes (main + helper + vite)
+$SCRIPT stop
+
+# Force fresh restart
+$SCRIPT restart
+```
+
+After `start` succeeds, connect with: `agent-browser --cdp 9222 snapshot -i`
+
+**Always run `$SCRIPT stop` when done testing** — `pkill -f "Electron"` alone won't catch all helper processes.
+
+#### Environment Variables
+
+| Variable          | Default                 | Description                              |
+| ----------------- | ----------------------- | ---------------------------------------- |
+| `CDP_PORT`        | `9222`                  | Chrome DevTools Protocol port            |
+| `ELECTRON_LOG`    | `/tmp/electron-dev.log` | Electron process log                     |
+| `ELECTRON_WAIT_S` | `60`                    | Max seconds to wait for Electron process |
+| `RENDERER_WAIT_S` | `60`                    | Max seconds to wait for SPA to load      |
+
+### LobeHub Probes & Quick Navigation
+
+`scripts/app-probe.sh` is the standard fast path into app state — **use it
+instead of hand-rolling `__LOBE_STORES` eval snippets** for these common needs:
+
+```bash
+PROBE=".agents/skills/agent-testing/scripts/app-probe.sh"
+
+$PROBE auth              # login check (Step 0.3) → { isSignedIn, userId }
+$PROBE route             # current SPA route
+$PROBE ops               # running chat operations (type / startTime)
+$PROBE goto /settings    # jump the SPA straight to a route (full reload)
+$PROBE errors-install    # install console.error interceptor
+$PROBE errors            # dump captured errors
+```
+
+`goto` lets a test enter the state under test directly instead of clicking
+through the UI. Common desktop routes:
+
+| Route                         | Where it lands                       |
+| ----------------------------- | ------------------------------------ |
+| `/`                           | Home (has a chat input)              |
+| `/agent/<agentId>`            | Agent conversation (latest topic)    |
+| `/agent/<agentId>/<topicId>`  | Specific topic in a conversation     |
+| `/task` · `/task/<taskId>`    | Task list / task detail              |
+| `/page`                       | Documents (文稿)                     |
+| `/settings`                   | Settings                             |
+| `/community`                  | Discover / community                 |
+
+Targets default to Electron (`--cdp 9222`); set `AB_TARGET="--session <name>"`
+for web sessions. For deeper or one-off state inspection, fall back to raw
+eval below.
+
+### LobeHub-Specific Patterns
+
+#### Access Zustand Store State
+
+```bash
+agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
+(function() {
+  var chat = window.__LOBE_STORES.chat();
+  var ops = Object.values(chat.operations);
+  return JSON.stringify({
+    ops: ops.map(function(o) { return { type: o.type, status: o.status }; }),
+    activeAgent: chat.activeAgentId,
+    activeTopic: chat.activeTopicId,
+  });
+})()
+EVALEOF
+```
+
+#### Find and Use the Chat Input
+
+```bash
+# The chat input is contenteditable — must use -C flag
+agent-browser --cdp 9222 snapshot -i -C 2>&1 | grep "editable"
+
+agent-browser --cdp 9222 click @e48
+agent-browser --cdp 9222 type @e48 "Hello world"
+agent-browser --cdp 9222 press Enter
+```
+
+#### Wait for Agent to Complete
+
+```bash
+agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
+(function() {
+  var chat = window.__LOBE_STORES.chat();
+  var ops = Object.values(chat.operations);
+  var running = ops.filter(function(o) { return o.status === 'running'; });
+  return running.length === 0 ? 'done' : 'running: ' + running.length;
+})()
+EVALEOF
+```
+
+#### Install Error Interceptor
+
+```bash
+agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
+(function() {
+  window.__CAPTURED_ERRORS = [];
+  var orig = console.error;
+  console.error = function() {
+    var msg = Array.from(arguments).map(function(a) {
+      if (a instanceof Error) return a.message;
+      return typeof a === 'object' ? JSON.stringify(a) : String(a);
+    }).join(' ');
+    window.__CAPTURED_ERRORS.push(msg);
+    orig.apply(console, arguments);
+  };
+  return 'installed';
+})()
+EVALEOF
+
+# Later, check captured errors:
+agent-browser --cdp 9222 eval "JSON.stringify(window.__CAPTURED_ERRORS)"
+```
+
+## Electron Gotchas
+
+- **Always use `electron-dev.sh stop` to clean up** — `pkill -f "Electron"` only kills the main process; helper processes (GPU, renderer, network) survive. The script finds and kills all of them via PID matching against the project's electron binary path.
+- **`npx electron-vite dev` must run from `apps/desktop/`** — running from project root fails silently. The `electron-dev.sh` script handles this automatically.
+- **Dev build auto-opens DevTools, which hijacks the CDP target** — `agent-browser --cdp 9222` may attach to the DevTools page (`devtools://…`) instead of the app (`app://renderer/`). Symptom: `get url` returns a `devtools://` URL. Fix: close the DevTools target and reconnect:
+
+  ```bash
+  DT_ID=$(curl -s http://localhost:9222/json/list | python3 -c "import json,sys; ts=json.load(sys.stdin); print(next(t['id'] for t in ts if t['type']=='page' and t['url'].startswith('devtools://')))")
+  curl -s "http://localhost:9222/json/close/$DT_ID" > /dev/null
+  agent-browser close --all && agent-browser --cdp 9222 get url   # expect app://renderer/
+  ```
+
+- **Don't resize the Electron window after load** — resizing triggers full SPA reload
+- **Store is at `window.__LOBE_STORES`** not `window.__ZUSTAND_STORES__`
+- **Streaming / ticking UI needs GIF evidence** — see `scripts/record-gif.sh`; a static screenshot cannot prove time-based behavior.

.agents/skills/agent-testing/ui/web.md

@@ -0,0 +1,78 @@
+# Web (Full-Stack) Testing
+
+Default surface for **full-stack changes** — a new/changed API plus the UI that
+consumes it. The browser is the one surface where network requests and UI state
+are observable together, so you can assert both sides of the contract in a
+single run.
+
+For pure-frontend changes prefer [electron.md](./electron.md); for
+backend-only changes prefer [../cli/index.md](../cli/index.md).
+
+## Prerequisites
+
+- Complete [Step 0.0](../SKILL.md#00-resolve-the-current-test-environment) (resolve ports) and [Step -1](../SKILL.md#step--1--plan-approval-for-non-trivial-tests) (plan approval) first.
+- Local dev server running — [../references/dev-server.md](../references/dev-server.md)
+- Web auth verified in agent-browser — prefer `setup-auth.sh web-seed`, see [auth decision flow](../references/auth.md#web--decision-flow).
+
+## Option A — agent-browser with seeded auth (recommended)
+
+```bash
+./.agents/skills/agent-testing/scripts/init-dev-env.sh seed-user
+./.agents/skills/agent-testing/scripts/setup-auth.sh web-seed
+```
+
+Then drive the verified session:
+
+```bash
+SESSION=lobehub-dev
+
+agent-browser --session $SESSION open "$SERVER_URL/"
+agent-browser --session $SESSION snapshot -i
+# interact via refs — full command reference: ../references/agent-browser.md
+```
+
+Use this session as the evidence source. Do not use ordinary Chrome screenshots
+or Chrome Network records as proof for Web tests; ordinary Chrome is only a
+fallback source for copying cookies into agent-browser when the seeded login is
+not available.
+
+### Watch the API while driving the UI
+
+```bash
+# After triggering the UI action under test:
+agent-browser --session $SESSION network requests --type xhr,fetch
+agent-browser --session $SESSION network requests --method POST
+
+# Record a full HAR for the report
+agent-browser --session $SESSION network har start
+# ... drive the scenario ...
+agent-browser --session $SESSION network har stop ./capture.har
+```
+
+Assert both layers: the request/response shape (network) and the rendered
+result (snapshot/screenshot). Both belong in the report as evidence.
+
+## Option B — real Chrome with remote debugging
+
+For flows that need a real, visible browser (e.g. exercising the login UI
+itself):
+
+```bash
+/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome \
+  --remote-debugging-port=9222 \
+  --user-data-dir=/tmp/chrome-test-profile \
+  "<URL>" &
+sleep 5
+agent-browser --cdp 9222 snapshot -i
+
+# Or auto-discover running Chrome with remote debugging
+agent-browser --auto-connect snapshot -i
+```
+
+## Option C — Debug Proxy (local frontend, production backend)
+
+`bun run dev:spa` prints a **Debug Proxy** URL
+(`https://app.lobehub.com/_dangerous_local_dev_proxy?debug-host=…`) that loads
+your local Vite SPA inside the online environment — HMR against real server
+config. Useful for verifying frontend behavior against production data, **not**
+for testing backend changes (the backend is production, not your branch).

.agents/skills/agent-tracing/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: agent-tracing
-description: "Agent tracing CLI for inspecting agent execution snapshots. Use when user mentions 'agent-tracing', 'trace', 'snapshot', wants to debug agent execution, inspect LLM calls, view context engine data, or analyze agent steps. Triggers on agent debugging, trace inspection, or execution analysis tasks."
+description: 'Agent tracing CLI for execution snapshots. Use for agent-tracing, traces, snapshots, LLM call inspection, context engine data, agent step analysis, or execution debugging.'
 user-invocable: false
 ---
 
@@ -14,7 +14,7 @@ In `NODE_ENV=development`, `AgentRuntimeService.executeStep()` automatically rec
 
 **Data flow**: executeStep loop -> build `StepPresentationData` -> write partial snapshot to disk -> on completion, finalize to `.agent-tracing/{timestamp}_{traceId}.json`
 
-**Context engine capture**: In `RuntimeExecutors.ts`, the `call_llm` executor emits a `context_engine_result` event after `serverMessagesEngine()` processes messages. This event carries the full `contextEngineInput` (DB messages, systemRole, model, knowledge, tools, userMemory, etc.) and the processed `output` messages (the final LLM payload).
+**Context engine capture**: In `RuntimeExecutors.ts`, the `call_llm` executor calls `ctx.tracingContextEngine(input, output)` after `serverMessagesEngine()` processes messages. `AgentRuntimeService.executeStep` buffers the call per step and forwards it to `OperationTraceRecorder.appendStep` as the typed `contextEngine` field. CE flows through this side channel rather than the `events` array so its heavy payload (agentDocuments, systemRole, …) never enters the Redis state pipeline (LOBE-9110).
 
 ## Package Location
 
@@ -199,9 +199,10 @@ interface StepSnapshot {
   messages?: any[]; // DB messages before step
   context?: { phase: string; payload?: unknown; stepContext?: unknown };
   events?: Array<{ type: string; [key: string]: unknown }>;
-  // context_engine_result event contains:
-  //   input: full contextEngineInput (messages, systemRole, model, knowledge, tools, userMemory, ...)
-  //   output: processed messages array (final LLM payload)
+  contextEngine?: {
+    input?: unknown; // contextEngineInput minus messages + toolsConfig (reconstructible from baseline)
+    output?: unknown; // processed messages array (final LLM payload)
+  };
 }
 ```
 
@@ -215,6 +216,6 @@ When using `--messages`, the output shows three sections (if context engine data
 
 ## Integration Points
 
-- **Recording**: `src/server/services/agentRuntime/AgentRuntimeService.ts` — in the `executeStep()` method, after building `stepPresentationData`, writes partial snapshot in dev mode
-- **Context engine event**: `src/server/modules/AgentRuntime/RuntimeExecutors.ts` — in `call_llm` executor, after `serverMessagesEngine()` returns, emits `context_engine_result` event
+- **Recording**: `apps/server/src/services/agentRuntime/AgentRuntimeService.ts` — in the `executeStep()` method, after building `stepPresentationData`, writes partial snapshot in dev mode
+- **Context engine capture**: `apps/server/src/modules/AgentRuntime/RuntimeExecutors.ts` — in `call_llm` executor, after `serverMessagesEngine()` returns, calls `ctx.tracingContextEngine(input, output)`. `AgentRuntimeService.executeStep` buffers it per step and passes it to `traceRecorder.appendStep` as the typed `contextEngine` field (kept off the `events` array to stay out of Redis state).
 - **Store**: `FileSnapshotStore` reads/writes to `.agent-tracing/` relative to `process.cwd()`

.agents/skills/builtin-tool/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: builtin-tool
-description: Build a new builtin tool package under `packages/builtin-tool-<name>/`. Use when adding a new agent-callable toolset, designing its API surface (manifest / ApiName / Params / State), implementing the Executor + ExecutionRuntime, building the Inspector / Render / Placeholder / Streaming / Intervention / Portal UI, or wiring a tool into the central registries (`packages/builtin-tools/src/{index,identifiers,inspectors,renders,placeholders,streamings,interventions,portals}.ts` and `src/store/tool/slices/builtin/executors/index.ts`). Triggers on "new builtin tool", "add a tool", "tool inspector", "tool render", "tool placeholder", "tool streaming", "tool intervention", "BuiltinToolManifest", "BaseExecutor", "ExecutionRuntime".
+description: 'Build LobeHub builtin tool packages. Use when adding agent-callable tools, manifests, executors, runtimes, inspectors, renders, placeholders, streaming, interventions, portals, or tool registries.'
 ---
 
 # Builtin Tool Authoring Guide
@@ -19,11 +19,11 @@ A builtin tool is a package the agent runtime can call. It ships **five faces**:
 
 ## Read These First
 
-| Question                                                                             | Doc                                |
-| ------------------------------------------------------------------------------------ | ---------------------------------- |
-| Where do files live? What does each face do? Wiring?                                 | [architecture.md](architecture.md) |
-| How do I name the tool, design APIs, write the manifest, executor, ExecutionRuntime? | [tool-design.md](tool-design.md)   |
-| How do I build Inspector / Render / Placeholder / Streaming / Intervention / Portal? | [ui.md](ui.md)                     |
+| Question                                                                             | Doc                                           |
+| ------------------------------------------------------------------------------------ | --------------------------------------------- |
+| Where do files live? What does each face do? Wiring?                                 | [architecture.md](references/architecture.md) |
+| How do I name the tool, design APIs, write the manifest, executor, ExecutionRuntime? | [tool-design.md](references/tool-design.md)   |
+| How do I build Inspector / Render / Placeholder / Streaming / Intervention / Portal? | [ui/](references/ui/README.md)                |
 
 ---
 
@@ -109,7 +109,7 @@ Before opening the PR:
 - [ ] Placeholder added if the API has a perceivable execution lag (search, list, crawl).
 - [ ] Streaming added for APIs that emit incremental output (run command, write file, code execution).
 - [ ] Intervention added if `humanIntervention` is set in the manifest.
-- [ ] All registry files updated (see [architecture.md → Registry wiring](architecture.md#registry-wiring)).
+- [ ] All registry files updated (see [architecture.md → Registry wiring](references/architecture.md#registry-wiring)).
 - [ ] i18n keys in `src/locales/default/plugin.ts` plus dev seeds in `en-US`/`zh-CN`.
 - [ ] `bunx vitest run --silent='passed-only' 'packages/builtin-tool-<name>'` passes.
 - [ ] `bun run type-check` passes.

.agents/skills/builtin-tool/references/architecture.md

@@ -213,7 +213,7 @@ The runtime hands every executor method an optional `BuiltinToolContext` as the
 | `operationId`                 | Operation lineage (use for cancellation, tracing)              |
 | `scope`                       | `'task' \| 'agent' \| …` — toggles default behaviors           |
 | `signal: AbortSignal`         | Honor for long-running ops                                     |
-| `stepContext`                 | Cross-message runtime state (GTD todos, etc.)                  |
+| `stepContext`                 | Cross-message runtime state (lobe-agent todos, etc.)           |
 | `registerAfterCompletion(cb)` | Defer side-effects past message-update race                    |
 | `groupOrchestration`          | Group orchestration callbacks                                  |
 
@@ -271,7 +271,7 @@ Lists in the same file you may need to touch:
 
 - `defaultToolIds` — added to the agent's tool list by default
 - `alwaysOnToolIds` — forced on regardless of user selection (use sparingly)
-- `runtimeManagedToolIds` — enable state controlled by runtime, not user UI; **must mirror the rules map** in `src/server/modules/Mecha/AgentToolsEngine/index.ts` and `src/helpers/toolEngineering/index.ts`
+- `runtimeManagedToolIds` — enable state controlled by runtime, not user UI; **must mirror the rules map** in `apps/server/src/modules/Mecha/AgentToolsEngine/index.ts` and `src/helpers/toolEngineering/index.ts`
 
 ---
 

.agents/skills/builtin-tool/references/tool-design.md

@@ -2,7 +2,7 @@
 
 This doc covers everything that **isn't UI**: the tool's identifier, API surface, manifest, types, system prompt, ExecutionRuntime, and the executor that wires it into the frontend.
 
-For UI surfaces (Inspector / Render / Placeholder / Streaming / Intervention / Portal), see [ui.md](ui.md).
+For UI surfaces (Inspector / Render / Placeholder / Streaming / Intervention / Portal), see [ui/](ui/README.md).
 For where files live and how registries work, see [architecture.md](architecture.md).
 
 ---
@@ -156,7 +156,7 @@ export const TaskManifest: BuiltinToolManifest = {
   executors: ['client', 'server'],
 
   /* Default human intervention policy for all APIs that don't specify one.
-     Pair with an Intervention component (see ui.md). */
+     Pair with an Intervention component (see ui/intervention.md). */
   humanIntervention: 'never' | 'always' | { /* extended config */ },
 }
 ```

.agents/skills/builtin-tool/references/ui/README.md

@@ -0,0 +1,36 @@
+# Tool UI Surfaces
+
+A builtin tool can ship up to **six client-side surfaces**, each with a different role in the chat UI. Only `Inspector` is required; the other five are added on demand and registered in their own central files.
+
+| Surface      | Required? | When the chat shows it                                                | Registered in                                 |
+| ------------ | --------- | --------------------------------------------------------------------- | --------------------------------------------- |
+| Inspector    | ✅ Always | Header strip of every tool call (one-line chip)                       | `packages/builtin-tools/src/inspectors.ts`    |
+| Render       | Optional  | Rich result card below the header, after the call returns             | `packages/builtin-tools/src/renders.ts`       |
+| Placeholder  | Optional  | Skeleton between "args streaming complete" and "result arrives"       | `packages/builtin-tools/src/placeholders.ts`  |
+| Streaming    | Optional  | Live output during execution (e.g. command stdout)                    | `packages/builtin-tools/src/streamings.ts`    |
+| Intervention | Optional  | Approval / edit-before-run dialog (when `humanIntervention` triggers) | `packages/builtin-tools/src/interventions.ts` |
+| Portal       | Optional  | Full-screen detail view (right-side or modal)                         | `packages/builtin-tools/src/portals.ts`       |
+
+The two reference tools to read end-to-end:
+
+- **`builtin-tool-web-browsing/src/client/`** — Inspector + Render + Placeholder + Portal (no Intervention/Streaming).
+- **`builtin-tool-local-system/src/client/`** — all six surfaces, including `components/` for shared building blocks.
+
+---
+
+## Files in this folder
+
+Read **principles** and **shared-rules** first — they apply to every surface. Then jump to the surface you're building.
+
+| File                               | What it covers                                                          |
+| ---------------------------------- | ----------------------------------------------------------------------- |
+| [principles.md](principles.md)     | Design principles — when each surface exists and how far to take it     |
+| [shared-rules.md](shared-rules.md) | Cross-surface rules: component skeleton, styling, single-layer surfaces |
+| [inspector.md](inspector.md)       | Inspector — header chip (required)                                      |
+| [render.md](render.md)             | Render — rich result card                                               |
+| [placeholder.md](placeholder.md)   | Placeholder — skeleton between args and result                          |
+| [streaming.md](streaming.md)       | Streaming — live output during execution                                |
+| [intervention.md](intervention.md) | Intervention — approval / edit-before-run                               |
+| [portal.md](portal.md)             | Portal — full-screen detail view                                        |
+| [composition.md](composition.md)   | Shared subcomponents (`client/components/`) + package public API        |
+| [diagnostics.md](diagnostics.md)   | Symptom → surface quick-lookup                                          |

.agents/skills/builtin-tool/references/ui/composition.md

@@ -0,0 +1,51 @@
+# Composition — Shared Components & Package API
+
+## `client/components/` — Shared Subcomponents
+
+Cross-cutting building blocks used by multiple surfaces live here, not duplicated in each surface folder.
+
+Examples from `web-browsing/src/client/components/`:
+
+- `CategoryAvatar.tsx` — search category icon
+- `EngineAvatar.tsx` — search engine logo (used in Inspector chip + Render list + Portal header)
+- `SearchBar.tsx` — editable query bar (used in Render and Portal)
+
+Examples from `local-system/src/client/components/`:
+
+- `FileItem.tsx` — single file row (used in ListFiles Render, SearchFiles Render, MoveLocalFiles Render)
+- `FilePathDisplay.tsx` — path with truncation (used everywhere)
+
+### Rules
+
+- Live under `client/components/`, exported via `client/components/index.ts`.
+- Re-export from `client/index.ts` only if other packages need them; otherwise keep internal.
+- Keep them dumb — props in, JSX out, no store reads. The store reads belong in the surface that composes them.
+
+---
+
+## `client/index.ts` — Package Public API
+
+Re-exports everything the registries need plus useful types/manifest:
+
+```ts
+// Inspector — required
+export { TaskInspectors } from './Inspector';
+
+// Render — only if any API has one
+export { TaskRenders, CreateTaskRender, RunTasksRender } from './Render';
+
+// Placeholder / Streaming / Intervention — only if used
+export { LocalSystemListFilesPlaceholder, LocalSystemSearchFilesPlaceholder } from './Placeholder';
+export { LocalSystemStreamings } from './Streaming';
+export { LocalSystemInterventions } from './Intervention';
+
+// Portal — single export per tool
+export { default as WebBrowsingPortal } from './Portal';
+
+// Reusable components if other packages need them
+export { CategoryAvatar, EngineAvatar, SearchBar } from './components';
+
+// Re-export manifest, identifier, types for convenience
+export { TaskManifest, TaskIdentifier } from '../manifest';
+export * from '../types';
+```

.agents/skills/builtin-tool/references/ui/diagnostics.md

@@ -0,0 +1,15 @@
+# Diagnostic Quick-Lookup
+
+| Symptom                                         | Surface to check                                                                                                                      |
+| ----------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------- |
+| No header at all on the tool call               | Inspector missing from `client/Inspector/index.ts` registry                                                                           |
+| Header shows the API name but no chips          | Inspector missing `args?.X \|\| partialArgs?.X` fallback                                                                              |
+| Header doesn't pulse during loading             | Missing `shinyTextStyles.shinyText` on `isArgumentsStreaming \|\| isLoading`                                                          |
+| Empty result card under header                  | Render returned `<div />` instead of `null` when no data                                                                              |
+| Render looks "complex" / card-in-card           | Filled container (`colorFillQuaternary`) wrapping more filled boxes — flatten to single-layer, see [shared-rules.md](shared-rules.md) |
+| Layout jump when result arrives                 | Placeholder dimensions don't match Render dimensions                                                                                  |
+| Approval dialog never appears                   | Manifest missing `humanIntervention`, or Intervention not in registry                                                                 |
+| Approval click doesn't wait for inline edit     | Missing `registerBeforeApprove(id, flushFn)`                                                                                          |
+| Portal opens but blank                          | Switch in `Portal/index.tsx` doesn't cover the apiName                                                                                |
+| Strings show as `builtins.lobe-foo.apiName.bar` | Missing i18n key in `src/locales/default/plugin.ts` (or not seeded in dev locale files)                                               |
+| Wrong color shade on `<Text type="secondary">`  | `type='secondary'` is lighter than `colorTextSecondary` — pass via `style={{ color: cssVar.colorTextSecondary }}`                     |

.agents/skills/builtin-tool/references/ui/inspector.md

@@ -0,0 +1,118 @@
+# Inspector — Header Chip (required)
+
+**Lifecycle:** Inspector renders for **every phase** of a tool call: while args are streaming in, while the executor is running, and after results come back. It's the only surface that's always visible.
+
+**Goal:** keep it to a single line. Show what's happening with as much context as is currently available.
+
+## Props (`BuiltinInspectorProps<Args, State>`)
+
+```ts
+interface BuiltinInspectorProps<Arguments = any, State = any> {
+  apiName: string;
+  args: Arguments; // final args (only after the assistant stops streaming)
+  identifier: string;
+  isArgumentsStreaming?: boolean; // args still arriving
+  isLoading?: boolean; // args complete, executor running
+  partialArgs?: Arguments; // partial JSON during streaming
+  pluginState?: State; // executor's `state` after success
+  result?: { content: string | null; error?: any };
+}
+```
+
+## State machine
+
+| Phase                               | What's available                                           | What to show                                               |
+| ----------------------------------- | ---------------------------------------------------------- | ---------------------------------------------------------- |
+| Args streaming, no useful field yet | `isArgumentsStreaming === true`, `partialArgs.X` undefined | Just the API title with `shinyTextStyles.shinyText`        |
+| Args streaming, key field arrived   | `partialArgs.X` populated                                  | Title + key field chip, still pulse-animated               |
+| Args complete, executor running     | `args` populated, `isLoading === true`                     | Same as above, still pulse-animated                        |
+| Result arrived                      | `pluginState` populated, `isLoading === false`             | Title + chips + result summary (count, identifier, status) |
+
+## Canonical example — Search
+
+`packages/builtin-tool-web-browsing/src/client/Inspector/Search/index.tsx`:
+
+```tsx
+'use client';
+
+import type { BuiltinInspectorProps, SearchQuery, UniformSearchResponse } from '@lobechat/types';
+import { Text } from '@lobehub/ui';
+import { cssVar, cx } from 'antd-style';
+import { memo } from 'react';
+import { useTranslation } from 'react-i18next';
+
+import { highlightTextStyles, inspectorTextStyles, shinyTextStyles } from '@/styles';
+
+export const SearchInspector = memo<BuiltinInspectorProps<SearchQuery, UniformSearchResponse>>(
+  ({ args, partialArgs, isArgumentsStreaming, isLoading, pluginState }) => {
+    const { t } = useTranslation('plugin');
+
+    const query = args?.query || partialArgs?.query || '';
+    const resultCount = pluginState?.results?.length ?? 0;
+    const hasResults = resultCount > 0;
+
+    if (isArgumentsStreaming && !query) {
+      return (
+        <div className={cx(inspectorTextStyles.root, shinyTextStyles.shinyText)}>
+          <span>{t('builtins.lobe-web-browsing.apiName.search')}</span>
+        </div>
+      );
+    }
+
+    return (
+      <div
+        className={cx(
+          inspectorTextStyles.root,
+          (isArgumentsStreaming || isLoading) && shinyTextStyles.shinyText,
+        )}
+      >
+        <span>{t('builtins.lobe-web-browsing.apiName.search')}:&nbsp;</span>
+        {query && <span className={highlightTextStyles.primary}>{query}</span>}
+        {!isLoading &&
+          !isArgumentsStreaming &&
+          pluginState?.results &&
+          (hasResults ? (
+            <span style={{ marginInlineStart: 4 }}>({resultCount})</span>
+          ) : (
+            <Text as="span" color={cssVar.colorTextDescription} fontSize={12}>
+              ({t('builtins.lobe-web-browsing.inspector.noResults')})
+            </Text>
+          ))}
+      </div>
+    );
+  },
+);
+SearchInspector.displayName = 'SearchInspector';
+export default SearchInspector;
+```
+
+## Inspector rules
+
+- Wrap the whole row with `inspectorTextStyles.root` (provides correct flex / line-height baseline).
+- Pulse with `shinyTextStyles.shinyText` whenever `isArgumentsStreaming || isLoading`.
+- Show the i18n title first so the row is non-empty during the earliest streaming phase.
+- Read both `args?.X` and `partialArgs?.X` together — `args` is final, `partialArgs` is in-stream.
+- Use chips/tags for distinct facets (identifier, name, parent, status, count). Each chip should clip with `text-overflow: ellipsis` and have a `max-width` so long values don't blow out the chat bubble.
+- Append `pluginState`-derived suffixes only **after** loading finishes — count or "(no results)" should not appear while still searching.
+- **Switch copy by phase.** If the verb implies an ongoing action ("Creating", "Searching", "Listing"), define `<api>.loading` and `<api>.completed` keys and select via `isArgumentsStreaming || isLoading ? loadingKey : completedKey`. Inspector chips persist in chat history — leaving "Creating task" frozen on a finished call reads as if the tool is still running. Read-only labels that are already noun-form ("View task") can keep a single key. See `CallSubAgentInspector` for the canonical two-key pattern.
+
+## Inspector registry — `client/Inspector/index.ts`
+
+```ts
+import type { BuiltinInspector } from '@lobechat/types';
+
+import { TaskApiName } from '../../types';
+import { CreateTaskInspector } from './CreateTask';
+import { ListTasksInspector } from './ListTasks';
+/* … */
+
+export const TaskInspectors: Record<string, BuiltinInspector> = {
+  [TaskApiName.createTask]: CreateTaskInspector as BuiltinInspector,
+  [TaskApiName.listTasks]: ListTasksInspector as BuiltinInspector,
+  /* one entry per ApiName */
+};
+
+export { CreateTaskInspector } from './CreateTask';
+export { ListTasksInspector } from './ListTasks';
+/* re-export each */
+```

.agents/skills/builtin-tool/references/ui/intervention.md

@@ -0,0 +1,88 @@
+# Intervention — Approval / Edit-Before-Run (optional)
+
+**Lifecycle:** rendered **before the executor runs** for APIs whose manifest sets `humanIntervention`. The user sees a preview of the args, can edit them, then approves or skips/cancels.
+
+**Add for** destructive or sensitive ops: shell commands, file writes, file moves, payments, message broadcasts.
+
+## Props (`BuiltinInterventionProps<Args>`)
+
+```ts
+interface BuiltinInterventionProps<Arguments = any> {
+  apiName?: string;
+  args: Arguments;
+  identifier?: string;
+  interactionMode?: 'approval' | 'custom';
+  messageId: string;
+
+  /** Called when the user edits the args; the approve action awaits this. */
+  onArgsChange?: (args: Arguments) => void | Promise<void>;
+
+  /** Called on approve / skip / cancel. */
+  onInteractionAction?: (
+    action:
+      | { type: 'submit'; payload: Record<string, unknown> }
+      | { type: 'skip'; payload?: Record<string, unknown>; reason?: string }
+      | { type: 'cancel'; payload?: Record<string, unknown> },
+  ) => Promise<void>;
+
+  /** Register a callback to flush pending saves before approval. Returns cleanup. */
+  registerBeforeApprove?: (id: string, callback: () => void | Promise<void>) => () => void;
+}
+```
+
+## Canonical example — RunCommand Intervention
+
+`packages/builtin-tool-local-system/src/client/Intervention/RunCommand/index.tsx`:
+
+```tsx
+import type { RunCommandParams } from '@lobechat/electron-client-ipc';
+import type { BuiltinInterventionProps } from '@lobechat/types';
+import { Flexbox, Highlighter, Text } from '@lobehub/ui';
+import { memo } from 'react';
+
+const RunCommand = memo<BuiltinInterventionProps<RunCommandParams>>(({ args }) => {
+  const { description, command, timeout } = args;
+  return (
+    <Flexbox gap={8}>
+      <Flexbox horizontal justify="space-between">
+        {description && <Text>{description}</Text>}
+        {timeout && (
+          <Text style={{ fontSize: 12 }} type="secondary">
+            timeout: {formatTimeout(timeout)}
+          </Text>
+        )}
+      </Flexbox>
+      {command && (
+        <Highlighter wrap language="sh" showLanguage={false} variant="outlined">
+          {command}
+        </Highlighter>
+      )}
+    </Flexbox>
+  );
+});
+export default RunCommand;
+```
+
+## Intervention rules
+
+- **Show a preview, not a form by default.** Editing UI is opt-in via `onArgsChange` and is usually inline (click to edit a code block, etc.).
+- For args with debounced edit state (text fields), use `registerBeforeApprove(id, flushFn)` so the approve action waits for the debounce to flush. Always return the cleanup function.
+- Call `onInteractionAction({ type: 'submit', payload })` when the user approves; `'skip'` if they skip with a reason; `'cancel'` if they cancel the whole turn.
+- Add a corresponding `interventionAudit.ts` in the package root if the tool needs scope/path validation before approval (see `local-system/src/interventionAudit.ts`).
+
+## Intervention registry — `client/Intervention/index.ts`
+
+```ts
+import { LocalSystemApiName } from '../..';
+import EditLocalFile from './EditLocalFile';
+import RunCommand from './RunCommand';
+import WriteFile from './WriteFile';
+/* … */
+
+export const LocalSystemInterventions = {
+  [LocalSystemApiName.editLocalFile]: EditLocalFile,
+  [LocalSystemApiName.runCommand]: RunCommand,
+  [LocalSystemApiName.writeLocalFile]: WriteFile,
+  /* one entry per API that needs approval */
+};
+```

.agents/skills/builtin-tool/references/ui/placeholder.md

@@ -0,0 +1,93 @@
+# Placeholder — Skeleton Between Args and Result (optional)
+
+**Lifecycle:** rendered when the args have finished streaming but the executor hasn't returned yet. Disappears when `pluginState` arrives. Bridges the moment of perceived lag.
+
+**Add for** APIs with noticeable execution time: web search, network crawl, file list, large grep. **Skip for** instant ops (status flips, calculator).
+
+## Props (`BuiltinPlaceholderProps<Args>`)
+
+```ts
+interface BuiltinPlaceholderProps<T extends Record<string, any> = any> {
+  apiName: string;
+  args?: T;
+  identifier: string;
+}
+```
+
+No `pluginState` — Placeholder lives entirely in the "executing" gap.
+
+## Canonical example — Search Placeholder
+
+`packages/builtin-tool-web-browsing/src/client/Placeholder/Search.tsx`:
+
+```tsx
+import type { BuiltinPlaceholderProps, SearchQuery } from '@lobechat/types';
+import { Flexbox, Icon, Skeleton } from '@lobehub/ui';
+import { createStaticStyles, cx } from 'antd-style';
+import { SearchIcon } from 'lucide-react';
+import { memo } from 'react';
+
+import { useIsMobile } from '@/hooks/useIsMobile';
+import { shinyTextStyles } from '@/styles';
+
+const styles = createStaticStyles(({ css, cssVar }) => ({
+  query: cx(
+    css`
+      padding: 4px 8px;
+      border-radius: 8px;
+      font-size: 12px;
+      color: ${cssVar.colorTextSecondary};
+      &:hover {
+        background: ${cssVar.colorFillTertiary};
+      }
+    `,
+    shinyTextStyles.shinyText,
+  ),
+}));
+
+export const Search = memo<BuiltinPlaceholderProps<SearchQuery>>(({ args }) => {
+  const { query } = args || {};
+  const isMobile = useIsMobile();
+
+  return (
+    <Flexbox gap={8}>
+      <Flexbox horizontal={!isMobile} gap={isMobile ? 8 : 40}>
+        <Flexbox horizontal align="center" className={styles.query} gap={8}>
+          <Icon icon={SearchIcon} />
+          {query ? query : <Skeleton.Block active style={{ height: 20, width: 40 }} />}
+        </Flexbox>
+        <Skeleton.Block active style={{ height: 20, width: 40 }} />
+      </Flexbox>
+      <Flexbox horizontal gap={12}>
+        {[1, 2, 3, 4, 5].map((id) => (
+          <Skeleton.Button active key={id} style={{ borderRadius: 8, height: 80, width: 160 }} />
+        ))}
+      </Flexbox>
+    </Flexbox>
+  );
+});
+```
+
+## Placeholder rules
+
+- **Mirror the eventual Render's layout.** When the result arrives the Placeholder unmounts and the Render mounts; if they share dimensions, the chat doesn't jump.
+- Use `Skeleton.Block` / `Skeleton.Button` from `@lobehub/ui` for placeholder shapes.
+- Embed any args you have (e.g. the query text) — context helps the user know what's loading.
+- Pulse with `shinyTextStyles.shinyText` if the Placeholder includes literal text.
+
+## Placeholder registry — `client/Placeholder/index.ts`
+
+```ts
+import { WebBrowsingApiName } from '../../types';
+import CrawlMultiPages from './CrawlMultiPages';
+import CrawlSinglePage from './CrawlSinglePage';
+import { Search } from './Search';
+
+export const WebBrowsingPlaceholders = {
+  [WebBrowsingApiName.crawlMultiPages]: CrawlMultiPages,
+  [WebBrowsingApiName.crawlSinglePage]: CrawlSinglePage,
+  [WebBrowsingApiName.search]: Search,
+};
+
+export { CrawlMultiPages, CrawlSinglePage, Search };
+```

.agents/skills/builtin-tool/references/ui/portal.md

@@ -0,0 +1,71 @@
+# Portal — Full-Screen Detail View (optional)
+
+**Lifecycle:** rendered when the user opens the tool message in a side panel or full-screen modal. One Portal per **tool**, not per API — the Portal switches on `apiName` internally.
+
+**Add for** tools whose results deserve a deep-dive view: search results with editable filters, page content with reader mode, code interpreter sessions.
+
+## Props (`BuiltinPortalProps<Args, State>`)
+
+```ts
+interface BuiltinPortalProps<Arguments = Record<string, any>, State = any> {
+  apiName?: string;
+  arguments: Arguments;
+  identifier: string;
+  messageId: string;
+  state: State;
+}
+```
+
+## Canonical example — Web-Browsing Portal
+
+`packages/builtin-tool-web-browsing/src/client/Portal/index.tsx`:
+
+```tsx
+import type { BuiltinPortalProps, CrawlPluginState, SearchQuery } from '@lobechat/types';
+import { memo } from 'react';
+
+import { WebBrowsingApiName } from '../../types';
+import PageContent from './PageContent';
+import PageContents from './PageContents';
+import Search from './Search';
+
+const Portal = memo<BuiltinPortalProps>(({ arguments: args, messageId, state, apiName }) => {
+  switch (apiName) {
+    case WebBrowsingApiName.search:
+      return <Search messageId={messageId} query={args as SearchQuery} response={state} />;
+
+    case WebBrowsingApiName.crawlSinglePage: {
+      const result = (state as CrawlPluginState).results.find((r) => r.originalUrl === args.url);
+      return <PageContent messageId={messageId} result={result} />;
+    }
+
+    case WebBrowsingApiName.crawlMultiPages:
+      return (
+        <PageContents
+          messageId={messageId}
+          results={(state as CrawlPluginState).results}
+          urls={args.urls}
+        />
+      );
+  }
+  return null;
+});
+export default Portal;
+```
+
+## Portal rules
+
+- One Portal per tool — the file is the routing layer, subcomponents implement each API's view.
+- Portals can read the chat store directly to detect "still streaming" and render a Skeleton internally (see `Search/index.tsx:20-46`).
+- Layout assumes more space than the Render — use `Flexbox` with `height={'100%'}` and structure for a side panel viewport.
+
+## Portal registry — `packages/builtin-tools/src/portals.ts`
+
+```ts
+import { WebBrowsingManifest, WebBrowsingPortal } from '@lobechat/builtin-tool-web-browsing/client';
+import { type BuiltinPortal } from '@lobechat/types';
+
+export const BuiltinToolsPortals: Record<string, BuiltinPortal> = {
+  [WebBrowsingManifest.identifier]: WebBrowsingPortal as BuiltinPortal,
+};
+```

.agents/skills/builtin-tool/references/ui/principles.md

@@ -0,0 +1,19 @@
+# Tool Render 设计原则（中文草案）
+
+这些原则用于判断一个 builtin tool 的 Inspector / Render / Placeholder / Streaming / Intervention / Portal 应该做什么，以及做到什么程度。
+
+1. **先保证折叠态可读。** 每个 API 都必须有 Inspector；用户不展开也应该能看懂 “正在做什么 / 对什么做 / 当前结果是什么”。Inspector 不应该只展示函数名和原始参数。
+2. **Inspector 是一句话，不是详情页。** 优先表达动作、关键对象、数量、状态，例如 “分析图片 3 张”“搜索 12 个结果”“读取 config.json”。长文本、列表和结构化结果放到 Render 或 Portal。
+3. **Inspector 要覆盖执行生命周期。** `args` 还在 streaming、工具执行中、执行完成、执行失败时都应该有稳定展示；必要时同时读取 `args`、`partialArgs` 和 `pluginState`，避免出现空白、跳变或只显示半截参数。
+4. **文案要随状态切换时态。** 同一个动作在 loading 与 completed 两个阶段必须用不同的措辞：执行中用现在进行时（“正在创建任务 / Creating task / 正在搜索”），执行完成后切到完成态（“已创建任务 / Task created / 已找到 N 条”）。Inspector chip 会一直留在聊天记录里 —— 如果一直挂着 “正在 xxx”，几小时后回看历史时会读起来像还在跑。约定的 i18n 形式是 `<api>.loading` / `<api>.completed` 一对键（见 `lobe-agent.apiName.callSubAgent.{loading,completed}` 与 `lobe-claude-code.task.{create,list,update,get}.{loading,completed}`），渲染时按 `isArgumentsStreaming || isLoading` 决定取哪一个。只读 / 查询类（“查看任务” 这种本来就是名词性的）可以共用一个键。
+5. **只有结构化结果才需要 Render。** 如果工具结果只是自然语言总结，通常不需要 Render；如果结果包含列表、媒体、文件、表格、代码、diff、地图、时间线、权限请求等结构，就应该提供 Render。
+6. **Render 要帮助用户检查结果，而不是复述参数。** Render 的主体应该围绕工具产物组织：可预览、可比较、可筛选、可定位。参数只作为上下文辅助出现，不要把 Render 做成一块更大的 args dump。
+7. **参数和结果要一起参与渲染。** 好的 Tool UI 通常同时用 `args` 解释意图，用 `pluginState` 展示真实执行结果；但 `pluginState` 只放结果域数据，不要反向塞入可以从 `args` 推导出的内容。
+8. **慢操作要有 Placeholder。** 如果工具通常需要等待网络、文件系统、模型或外部进程，Placeholder 应该先占住最终 Render 的版式，让用户知道即将看到什么，而不是只显示一个泛化 loading。
+9. **Streaming 只用于连续产物。** 搜索列表、日志、长文本、文件分析、分阶段计划适合 Streaming；一次性小结果不需要强行做 Streaming。Streaming UI 要能渐进追加，并且完成后自然过渡到最终 Render。
+10. **有风险的动作必须 Intervention。** 写文件、删除、发送、安装、执行命令、外部可见操作、权限敏感操作，都应该在执行前给出可理解的确认界面；确认文案要说明影响范围，而不是只问 “是否继续”。
+11. **错误、空态和截断都是正式状态。** Render 不能在失败、无结果、超长结果时退化成空白。错误要说明发生在哪一步；空态要告诉用户没有产物；超长内容要明确 “展示前 N 项 / 还有 N 项”。
+12. **信息密度要克制。** 默认展示最有判断价值的部分：标题、来源、状态、摘要、少量关键字段。大对象、长列表、原文、调试数据放进可展开区域或 Portal，避免把聊天流撑成后台管理页。
+13. **视觉上融入聊天流。** Tool UI 应该使用 `@lobehub/ui` / base-ui、`Flexbox`、`createStaticStyles` 和 `cssVar.*`，遵循现有间距、圆角、颜色、字号；不要为单个工具发明一套独立视觉语言。具体的样式约定见 [shared-rules.md](shared-rules.md)。
+14. **Devtools fixture 是验收入口。** 新增或修改 Tool UI 时，应在 `/devtools` 里准备覆盖典型态、loading/streaming、空态、错误态、长内容态的 fixture；一个 API 如果在真实聊天里会出现，就不应该在 devtools 中缺席。
+15. **先做用户会看的 UI，再做调试 UI。** Raw JSON、trace、schema、内部 id 可以存在，但应默认收起或放到调试区；主界面先回答用户最关心的问题：工具做了什么，结果值不值得信任，下一步能做什么。

.agents/skills/builtin-tool/references/ui/render.md

@@ -0,0 +1,101 @@
+# Render — Rich Result Card (optional)
+
+**Lifecycle:** rendered **once the result arrives** (after Placeholder/Streaming hand off). Sits below the Inspector header.
+
+**Skip if** the API is read-only or the result is just text — the framework already shows the executor's `content` string. Add a Render only when there's a structured artifact worth seeing: a card, a chart, a diff, a list of files.
+
+## Props (`BuiltinRenderProps<Args, State, Content>`)
+
+```ts
+interface BuiltinRenderProps<Arguments = any, State = any, Content = any> {
+  apiName?: string;
+  args: Arguments; // final params from the LLM
+  content: Content; // executor's content string (or parsed)
+  identifier?: string;
+  messageId: string; // for store lookups
+  pluginError?: any; // from BuiltinToolResult.error
+  pluginState?: State; // executor's state
+  toolCallId?: string;
+}
+```
+
+## Two patterns
+
+**Pattern A — Single-file Render** (web-browsing CrawlSinglePage):
+
+```tsx
+// client/Render/CrawlSinglePage.tsx
+import type { BuiltinRenderProps, CrawlPluginState, CrawlSinglePageQuery } from '@lobechat/types';
+import { memo } from 'react';
+
+import PageContent from './PageContent';
+
+const CrawlSinglePage = memo<BuiltinRenderProps<CrawlSinglePageQuery, CrawlPluginState>>(
+  ({ messageId, pluginState, args }) => (
+    <PageContent messageId={messageId} results={pluginState?.results} urls={[args?.url]} />
+  ),
+);
+export default CrawlSinglePage;
+```
+
+**Pattern B — Folder with subcomponents** (web-browsing Search):
+
+```
+client/Render/Search/
+├── index.tsx           # composes the subcomponents, handles error states
+├── ConfigForm.tsx      # appears when pluginError.type === 'PluginSettingsInvalid'
+├── SearchQuery.tsx     # editable query header
+└── SearchResult.tsx    # result list
+```
+
+Use Pattern B when the Render has internal state (editing mode, expanded items), error variants, or is large enough to benefit from splitting.
+
+## Error handling in Render
+
+Renders are the canonical place to surface `pluginError` because the chat doesn't auto-render typed errors:
+
+```tsx
+if (pluginError) {
+  if (pluginError?.type === 'PluginSettingsInvalid') {
+    return <ConfigForm id={messageId} provider={pluginError.body?.provider} />;
+  }
+  return (
+    <Alert
+      title={pluginError?.message}
+      type="error"
+      extra={<Highlighter language="json">{JSON.stringify(pluginError.body, null, 2)}</Highlighter>}
+    />
+  );
+}
+```
+
+## Render rules
+
+- **Return `null`** if there's nothing useful to draw yet (avoids empty cards during stream).
+- Use `pluginState` for server-truth (ids, counts, server-assigned status) and `args` for what the LLM asked. **Combine — neither alone is enough.**
+- For lists, summarize with a header line and show top N items with a "+N more" tail rather than rendering everything.
+- **Keep the Render single-layer** — the tool card is already your surface, so don't open with your own filled container and then nest more filled boxes inside it. See [shared-rules.md](shared-rules.md) → "Stay single-layer".
+- For modals from a Render, use `@lobehub/ui/base-ui` (`createModal`, `useModalContext`, `confirmModal`) — see the **modal** skill.
+
+## Render registry — `client/Render/index.ts`
+
+```ts
+import type { BuiltinRender } from '@lobechat/types';
+
+import { TaskApiName } from '../../types';
+import CreateTaskRender from './CreateTask';
+import RunTasksRender from './RunTasks';
+
+export const TaskRenders: Record<string, BuiltinRender> = {
+  [TaskApiName.createTask]: CreateTaskRender as BuiltinRender,
+  [TaskApiName.runTasks]: RunTasksRender as BuiltinRender,
+  /* only the APIs with rich result UI — others fall back to text content */
+};
+
+export { default as CreateTaskRender } from './CreateTask';
+export { default as RunTasksRender } from './RunTasks';
+```
+
+## Render display control (rare)
+
+If the Render should hide for certain results (e.g. ClaudeCode's TodoWrite hides when the agent is mid-stream), add a `RenderDisplayControl` to `packages/builtin-tools/src/displayControls.ts`. See `ClaudeCodeRenderDisplayControls` for the pattern.

.agents/skills/builtin-tool/references/ui/shared-rules.md

@@ -0,0 +1,89 @@
+# Shared Style Rules
+
+These apply across every surface.
+
+## The component skeleton
+
+Every surface file is the same shape, so internalize it once instead of re-deriving it per rule. The skeleton below bakes in five mechanical conventions — copy it and fill the body:
+
+```tsx
+'use client'; // (a) leaves of the chat tree must not block server rendering
+
+import type { BuiltinInspectorProps, SearchQuery, UniformSearchResponse } from '@lobechat/types';
+import { memo } from 'react';
+import { useTranslation } from 'react-i18next';
+
+// (b) type with BuiltinXProps<Args, State> — never widen to `any`.
+//     Args = the JSON Schema params, State = the executor's `state` field;
+//     they should match <Name>Params / <Name>State from types.ts.
+export const SearchInspector = memo<BuiltinInspectorProps<SearchQuery, UniformSearchResponse>>(
+  ({ args, pluginState }) => {
+    const { t } = useTranslation('plugin'); // (c) all strings from the `plugin` namespace
+
+    // (d) cross-cutting state (loading, streaming buffer) comes from the store,
+    //     not props — props only carry args/state/messageId.
+    // const buffer = useChatStore((s) => chatToolSelectors.streamingBuffer(messageId)(s));
+
+    return <span>{t('builtins.<identifier>.apiName.search')}</span>;
+  },
+);
+SearchInspector.displayName = 'SearchInspector'; // (e) always memo + displayName
+export default SearchInspector;
+```
+
+- **(c)** Default an Inspector to `t('builtins.<identifier>.apiName.<api>')` so the row is non-empty while args stream in.
+- **(d)** Read the store via Zustand selectors inside the component; see [streaming.md](streaming.md) for the buffer selector.
+
+## Styling: `createStaticStyles + cssVar.*`, `@lobehub/ui` over `antd`
+
+Zero-runtime CSS-in-JS — styles compile once and read CSS variables at runtime:
+
+```tsx
+import { createStaticStyles, cssVar } from 'antd-style';
+
+const styles = createStaticStyles(({ css, cssVar }) => ({
+  chip: css`
+    padding-block: 2px;
+    padding-inline: 8px;
+    border-radius: 999px;
+    color: ${cssVar.colorText};
+    background: ${cssVar.colorFillTertiary};
+  `,
+}));
+```
+
+- Fall back to `createStyles + token` only when you need runtime token computation (rare). Inline `style={{ color: cssVar.colorTextSecondary }}` is fine for one-off dynamic values.
+- Components come from `@lobehub/ui` (`Block`, `Text`, `Flexbox`, `Highlighter`, `Alert`, `Tooltip`, `Skeleton`), not raw `antd`. Modals come from `@lobehub/ui/base-ui` (`createModal`, `useModalContext`, `confirmModal`) — see the **modal** skill.
+- Note: `<Text type='secondary'>` is a lighter shade than `colorTextSecondary`. For that exact token color, write `<Text style={{ color: cssVar.colorTextSecondary }}>`.
+
+## Stay single-layer — don't nest filled cards
+
+The framework already wraps every Render / Intervention in a tool card, so that card **is** your surface. A Render that opens with its own `background: ${cssVar.colorFillQuaternary}` container is already one card deep; put another filled box inside it (`colorBgContainer` / `colorFillTertiary`) and you get the card-in-card look that reads as "complex" — two or three stacked fills for what is really a flat list of fields.
+
+- **The outermost wrapper carries no fill.** Use a flat container with only `padding-block: 4px` for breathing room; let the tool card provide the card. (See `Agent/index.tsx`'s `container`.)
+- **At most one filled box, and only to delineate real content** — a Markdown preview, a diff, a code/result block. Labels, key–value fields, question/answer text, chips: render flat on the surface, separated by spacing or a hairline divider (`height: 1px; background: ${cssVar.colorFillSecondary}`), not by wrapping each in its own box.
+- **A box on a flat surface needs a visible fill.** Once the outer fill is gone, an inner `colorBgContainer` box can vanish against the tool card (same color). Use `colorFillTertiary` for the one content box so it still reads as delineated.
+- Don't wrap a single value in a box just to give it padding — that's the redundant-nesting smell (a `detailCard` around a `value` box around one string).
+
+```tsx
+// ❌ card-in-card: filled container wrapping a filled preview box
+container: css`
+  padding: 12px;
+  background: ${cssVar.colorFillQuaternary};
+`,
+previewBox: css`
+  background: ${cssVar.colorBgContainer};
+`,
+
+// ✅ single-layer: flat container, one visible content box
+container: css`
+  padding-block: 4px;
+`,
+previewBox: css`
+  background: ${cssVar.colorFillTertiary};
+`,
+```
+
+For the common "icon + file/title header, then one content box" shape, reuse `ToolResultCard` from `@lobechat/shared-tool-ui/components` instead of rebuilding it — it's already single-layer (flat wrapper, one `colorFillTertiary` content box) and is what CC `Read` / `Grep` / `Glob` / `Write` / `WebSearch` / `WebFetch` render through.
+
+The exception is a deliberate **panel** pattern — an `<Block variant="outlined">` with a header bar + list rows (CC `TodoWrite` / `Task`). There the single outlined block is the panel and the header fill is a header bar, not a nested card. One structured panel is fine; stacked decorative fills are not.

.agents/skills/builtin-tool/references/ui/streaming.md

@@ -0,0 +1,83 @@
+# Streaming — Live Output During Execution (optional)
+
+**Lifecycle:** rendered **while the executor is still running** for APIs that emit incremental output. The component is responsible for fetching the in-flight stream from the chat store and rendering it.
+
+**Add for** long-running ops with continuous output: shell command execution (stdout/stderr), file write progress, code interpreter cells.
+
+## Props (`BuiltinStreamingProps<Args>`)
+
+```ts
+interface BuiltinStreamingProps<Arguments = any> {
+  apiName: string;
+  args: Arguments;
+  identifier: string;
+  messageId: string; // use to fetch the streaming buffer from store
+  toolCallId: string;
+}
+```
+
+Note there's **no `state` or `result` prop** — the Streaming component is for the in-flight phase. It pulls the live buffer from the store itself (typically via `chatToolSelectors.streamingContent(messageId)` or similar).
+
+## Canonical example — RunCommandStreaming
+
+`packages/builtin-tool-local-system/src/client/Streaming/RunCommand/index.tsx`:
+
+```tsx
+'use client';
+
+import type { BuiltinStreamingProps } from '@lobechat/types';
+import { Highlighter } from '@lobehub/ui';
+import { memo } from 'react';
+
+interface RunCommandParams {
+  command?: string;
+  description?: string;
+  timeout?: number;
+}
+
+export const RunCommandStreaming = memo<BuiltinStreamingProps<RunCommandParams>>(({ args }) => {
+  const { command } = args || {};
+  if (!command) return null;
+
+  return (
+    <Highlighter
+      animated
+      wrap
+      language="sh"
+      showLanguage={false}
+      style={{ padding: '4px 8px' }}
+      variant="outlined"
+    >
+      {command}
+    </Highlighter>
+  );
+});
+RunCommandStreaming.displayName = 'RunCommandStreaming';
+```
+
+For real-time output beyond just the command (stderr/stdout streaming), pull from the chat store:
+
+```tsx
+const buffer = useChatStore((state) =>
+  chatToolSelectors.streamingBuffer(messageId, toolCallId)(state),
+);
+```
+
+## Streaming rules
+
+- Render `null` until you have something to display (avoids flash).
+- For terminal-style output, use `Highlighter` with `animated` to show typing-like effect.
+- The Streaming component must **unmount cleanly** when execution ends — typically the framework swaps it out for the Render automatically.
+
+## Streaming registry — `client/Streaming/index.ts`
+
+```ts
+import { LocalSystemApiName } from '../..';
+import { RunCommandStreaming } from './RunCommand';
+import { WriteFileStreaming } from './WriteFile';
+
+export const LocalSystemStreamings = {
+  [LocalSystemApiName.runCommand]: RunCommandStreaming,
+  [LocalSystemApiName.writeLocalFile]: WriteFileStreaming,
+};
+```

.agents/skills/builtin-tool/ui.md

@@ -1,721 +0,0 @@
-# Tool UI Surfaces
-
-A builtin tool can ship up to **six client-side surfaces**, each with a different role in the chat UI. Only `Inspector` is required; the other five are added on demand and registered in their own central files.
-
-| Surface      | Required? | When the chat shows it                                                | Registered in                                 |
-| ------------ | --------- | --------------------------------------------------------------------- | --------------------------------------------- |
-| Inspector    | ✅ Always | Header strip of every tool call (one-line chip)                       | `packages/builtin-tools/src/inspectors.ts`    |
-| Render       | Optional  | Rich result card below the header, after the call returns             | `packages/builtin-tools/src/renders.ts`       |
-| Placeholder  | Optional  | Skeleton between "args streaming complete" and "result arrives"       | `packages/builtin-tools/src/placeholders.ts`  |
-| Streaming    | Optional  | Live output during execution (e.g. command stdout)                    | `packages/builtin-tools/src/streamings.ts`    |
-| Intervention | Optional  | Approval / edit-before-run dialog (when `humanIntervention` triggers) | `packages/builtin-tools/src/interventions.ts` |
-| Portal       | Optional  | Full-screen detail view (right-side or modal)                         | `packages/builtin-tools/src/portals.ts`       |
-
-The two reference tools to read end-to-end:
-
-- **`builtin-tool-web-browsing/src/client/`** — Inspector + Render + Placeholder + Portal (no Intervention/Streaming).
-- **`builtin-tool-local-system/src/client/`** — all six surfaces, including `components/` for shared building blocks.
-
----
-
-## 0. Shared Style Rules
-
-These apply across every surface.
-
-### 0.1 Use `'use client'` at the top of every component file
-
-Tool surfaces are leaves in the chat tree and must not block server rendering.
-
-### 0.2 Prefer `createStaticStyles + cssVar.*`
-
-Zero-runtime CSS-in-JS — the styles compile once and read CSS variables at runtime.
-
-```tsx
-import { createStaticStyles, cssVar } from 'antd-style';
-
-const styles = createStaticStyles(({ css, cssVar }) => ({
-  chip: css`
-    padding-block: 2px;
-    padding-inline: 8px;
-    border-radius: 999px;
-    color: ${cssVar.colorText};
-    background: ${cssVar.colorFillTertiary};
-  `,
-}));
-```
-
-Fall back to `createStyles + token` only when you need runtime token computation (rare). Inline `style={{ color: cssVar.colorTextSecondary }}` is fine for one-off dynamic values.
-
-### 0.3 Use `@lobehub/ui`, not raw `antd`
-
-`Block`, `Text`, `Flexbox`, `Highlighter`, `Alert`, `Tooltip`, `Skeleton` all come from `@lobehub/ui`. Modals come from `@lobehub/ui/base-ui` (`createModal`, `useModalContext`, `confirmModal`) — see the **modal** skill.
-
-Memory note: `@lobehub/ui`'s `<Text type='secondary'>` is a lighter shade than `colorTextSecondary`. If you need that exact token color, write `<Text style={{ color: cssVar.colorTextSecondary }}>`.
-
-### 0.4 Always `memo` and set `displayName`
-
-```tsx
-export const SearchInspector = memo<BuiltinInspectorProps<SearchQuery, UniformSearchResponse>>(
-  ({ args /* … */ }) => {
-    /* … */
-  },
-);
-SearchInspector.displayName = 'SearchInspector';
-export default SearchInspector;
-```
-
-### 0.5 Always type with `BuiltinXProps<Args, State>` generics
-
-Don't widen to `any`. The Args generic is the JSON Schema params, the State generic is the executor's `state` field. The two should match `<Name>Params` and `<Name>State` from `types.ts`.
-
-### 0.6 Pull strings from `t('plugin')`
-
-```tsx
-const { t } = useTranslation('plugin');
-t('builtins.<identifier>.apiName.<api>');
-```
-
-Every Inspector should default to `t('builtins.<identifier>.apiName.<api>')` so it shows something while args stream in.
-
-### 0.7 Read store state from `@/store/chat`, not props
-
-Tool surfaces sometimes need cross-cutting state (loading, streaming buffer). Read it inside the component via Zustand selectors, not from props — props only carry args/state/messageId.
-
----
-
-## 1. Inspector — Header Chip (required)
-
-**Lifecycle:** Inspector renders for **every phase** of a tool call: while args are streaming in, while the executor is running, and after results come back. It's the only surface that's always visible.
-
-**Goal:** keep it to a single line. Show what's happening with as much context as is currently available.
-
-### Props (`BuiltinInspectorProps<Args, State>`)
-
-```ts
-interface BuiltinInspectorProps<Arguments = any, State = any> {
-  apiName: string;
-  args: Arguments; // final args (only after the assistant stops streaming)
-  identifier: string;
-  isArgumentsStreaming?: boolean; // args still arriving
-  isLoading?: boolean; // args complete, executor running
-  partialArgs?: Arguments; // partial JSON during streaming
-  pluginState?: State; // executor's `state` after success
-  result?: { content: string | null; error?: any };
-}
-```
-
-### State machine
-
-| Phase                               | What's available                                           | What to show                                               |
-| ----------------------------------- | ---------------------------------------------------------- | ---------------------------------------------------------- |
-| Args streaming, no useful field yet | `isArgumentsStreaming === true`, `partialArgs.X` undefined | Just the API title with `shinyTextStyles.shinyText`        |
-| Args streaming, key field arrived   | `partialArgs.X` populated                                  | Title + key field chip, still pulse-animated               |
-| Args complete, executor running     | `args` populated, `isLoading === true`                     | Same as above, still pulse-animated                        |
-| Result arrived                      | `pluginState` populated, `isLoading === false`             | Title + chips + result summary (count, identifier, status) |
-
-### Canonical example — Search
-
-`packages/builtin-tool-web-browsing/src/client/Inspector/Search/index.tsx`:
-
-```tsx
-'use client';
-
-import type { BuiltinInspectorProps, SearchQuery, UniformSearchResponse } from '@lobechat/types';
-import { Text } from '@lobehub/ui';
-import { cssVar, cx } from 'antd-style';
-import { memo } from 'react';
-import { useTranslation } from 'react-i18next';
-
-import { highlightTextStyles, inspectorTextStyles, shinyTextStyles } from '@/styles';
-
-export const SearchInspector = memo<BuiltinInspectorProps<SearchQuery, UniformSearchResponse>>(
-  ({ args, partialArgs, isArgumentsStreaming, isLoading, pluginState }) => {
-    const { t } = useTranslation('plugin');
-
-    const query = args?.query || partialArgs?.query || '';
-    const resultCount = pluginState?.results?.length ?? 0;
-    const hasResults = resultCount > 0;
-
-    if (isArgumentsStreaming && !query) {
-      return (
-        <div className={cx(inspectorTextStyles.root, shinyTextStyles.shinyText)}>
-          <span>{t('builtins.lobe-web-browsing.apiName.search')}</span>
-        </div>
-      );
-    }
-
-    return (
-      <div
-        className={cx(
-          inspectorTextStyles.root,
-          (isArgumentsStreaming || isLoading) && shinyTextStyles.shinyText,
-        )}
-      >
-        <span>{t('builtins.lobe-web-browsing.apiName.search')}:&nbsp;</span>
-        {query && <span className={highlightTextStyles.primary}>{query}</span>}
-        {!isLoading &&
-          !isArgumentsStreaming &&
-          pluginState?.results &&
-          (hasResults ? (
-            <span style={{ marginInlineStart: 4 }}>({resultCount})</span>
-          ) : (
-            <Text as="span" color={cssVar.colorTextDescription} fontSize={12}>
-              ({t('builtins.lobe-web-browsing.inspector.noResults')})
-            </Text>
-          ))}
-      </div>
-    );
-  },
-);
-SearchInspector.displayName = 'SearchInspector';
-export default SearchInspector;
-```
-
-### Inspector rules
-
-- Wrap the whole row with `inspectorTextStyles.root` (provides correct flex / line-height baseline).
-- Pulse with `shinyTextStyles.shinyText` whenever `isArgumentsStreaming || isLoading`.
-- Show the i18n title first so the row is non-empty during the earliest streaming phase.
-- Read both `args?.X` and `partialArgs?.X` together — `args` is final, `partialArgs` is in-stream.
-- Use chips/tags for distinct facets (identifier, name, parent, status, count). Each chip should clip with `text-overflow: ellipsis` and have a `max-width` so long values don't blow out the chat bubble.
-- Append `pluginState`-derived suffixes only **after** loading finishes — count or "(no results)" should not appear while still searching.
-
-### Inspector registry — `client/Inspector/index.ts`
-
-```ts
-import type { BuiltinInspector } from '@lobechat/types';
-
-import { TaskApiName } from '../../types';
-import { CreateTaskInspector } from './CreateTask';
-import { ListTasksInspector } from './ListTasks';
-/* … */
-
-export const TaskInspectors: Record<string, BuiltinInspector> = {
-  [TaskApiName.createTask]: CreateTaskInspector as BuiltinInspector,
-  [TaskApiName.listTasks]: ListTasksInspector as BuiltinInspector,
-  /* one entry per ApiName */
-};
-
-export { CreateTaskInspector } from './CreateTask';
-export { ListTasksInspector } from './ListTasks';
-/* re-export each */
-```
-
----
-
-## 2. Render — Rich Result Card (optional)
-
-**Lifecycle:** rendered **once the result arrives** (after Placeholder/Streaming hand off). Sits below the Inspector header.
-
-**Skip if** the API is read-only or the result is just text — the framework already shows the executor's `content` string. Add a Render only when there's a structured artifact worth seeing: a card, a chart, a diff, a list of files.
-
-### Props (`BuiltinRenderProps<Args, State, Content>`)
-
-```ts
-interface BuiltinRenderProps<Arguments = any, State = any, Content = any> {
-  apiName?: string;
-  args: Arguments; // final params from the LLM
-  content: Content; // executor's content string (or parsed)
-  identifier?: string;
-  messageId: string; // for store lookups
-  pluginError?: any; // from BuiltinToolResult.error
-  pluginState?: State; // executor's state
-  toolCallId?: string;
-}
-```
-
-### Two patterns
-
-**Pattern A — Single-file Render** (web-browsing CrawlSinglePage):
-
-```tsx
-// client/Render/CrawlSinglePage.tsx
-import type { BuiltinRenderProps, CrawlPluginState, CrawlSinglePageQuery } from '@lobechat/types';
-import { memo } from 'react';
-
-import PageContent from './PageContent';
-
-const CrawlSinglePage = memo<BuiltinRenderProps<CrawlSinglePageQuery, CrawlPluginState>>(
-  ({ messageId, pluginState, args }) => (
-    <PageContent messageId={messageId} results={pluginState?.results} urls={[args?.url]} />
-  ),
-);
-export default CrawlSinglePage;
-```
-
-**Pattern B — Folder with subcomponents** (web-browsing Search):
-
-```
-client/Render/Search/
-├── index.tsx           # composes the subcomponents, handles error states
-├── ConfigForm.tsx      # appears when pluginError.type === 'PluginSettingsInvalid'
-├── SearchQuery.tsx     # editable query header
-└── SearchResult.tsx    # result list
-```
-
-Use Pattern B when the Render has internal state (editing mode, expanded items), error variants, or is large enough to benefit from splitting.
-
-### Error handling in Render
-
-Renders are the canonical place to surface `pluginError` because the chat doesn't auto-render typed errors:
-
-```tsx
-if (pluginError) {
-  if (pluginError?.type === 'PluginSettingsInvalid') {
-    return <ConfigForm id={messageId} provider={pluginError.body?.provider} />;
-  }
-  return (
-    <Alert
-      title={pluginError?.message}
-      type="error"
-      extra={<Highlighter language="json">{JSON.stringify(pluginError.body, null, 2)}</Highlighter>}
-    />
-  );
-}
-```
-
-### Render rules
-
-- **Return `null`** if there's nothing useful to draw yet (avoids empty cards during stream).
-- Use `pluginState` for server-truth (ids, counts, server-assigned status) and `args` for what the LLM asked. **Combine — neither alone is enough.**
-- For lists, summarize with a header line and show top N items with a "+N more" tail rather than rendering everything.
-- For modals from a Render, use `@lobehub/ui/base-ui` (`createModal`, `useModalContext`, `confirmModal`) — see the **modal** skill.
-
-### Render registry — `client/Render/index.ts`
-
-```ts
-import type { BuiltinRender } from '@lobechat/types';
-
-import { TaskApiName } from '../../types';
-import CreateTaskRender from './CreateTask';
-import RunTasksRender from './RunTasks';
-
-export const TaskRenders: Record<string, BuiltinRender> = {
-  [TaskApiName.createTask]: CreateTaskRender as BuiltinRender,
-  [TaskApiName.runTasks]: RunTasksRender as BuiltinRender,
-  /* only the APIs with rich result UI — others fall back to text content */
-};
-
-export { default as CreateTaskRender } from './CreateTask';
-export { default as RunTasksRender } from './RunTasks';
-```
-
-### Render display control (rare)
-
-If the Render should hide for certain results (e.g. ClaudeCode's TodoWrite hides when the agent is mid-stream), add a `RenderDisplayControl` to `packages/builtin-tools/src/displayControls.ts`. See `ClaudeCodeRenderDisplayControls` for the pattern.
-
----
-
-## 3. Placeholder — Skeleton Between Args and Result (optional)
-
-**Lifecycle:** rendered when the args have finished streaming but the executor hasn't returned yet. Disappears when `pluginState` arrives. Bridges the moment of perceived lag.
-
-**Add for** APIs with noticeable execution time: web search, network crawl, file list, large grep. **Skip for** instant ops (status flips, calculator).
-
-### Props (`BuiltinPlaceholderProps<Args>`)
-
-```ts
-interface BuiltinPlaceholderProps<T extends Record<string, any> = any> {
-  apiName: string;
-  args?: T;
-  identifier: string;
-}
-```
-
-No `pluginState` — Placeholder lives entirely in the "executing" gap.
-
-### Canonical example — Search Placeholder
-
-`packages/builtin-tool-web-browsing/src/client/Placeholder/Search.tsx`:
-
-```tsx
-import type { BuiltinPlaceholderProps, SearchQuery } from '@lobechat/types';
-import { Flexbox, Icon, Skeleton } from '@lobehub/ui';
-import { createStaticStyles, cx } from 'antd-style';
-import { SearchIcon } from 'lucide-react';
-import { memo } from 'react';
-
-import { useIsMobile } from '@/hooks/useIsMobile';
-import { shinyTextStyles } from '@/styles';
-
-const styles = createStaticStyles(({ css, cssVar }) => ({
-  query: cx(
-    css`
-      padding: 4px 8px;
-      border-radius: 8px;
-      font-size: 12px;
-      color: ${cssVar.colorTextSecondary};
-      &:hover {
-        background: ${cssVar.colorFillTertiary};
-      }
-    `,
-    shinyTextStyles.shinyText,
-  ),
-}));
-
-export const Search = memo<BuiltinPlaceholderProps<SearchQuery>>(({ args }) => {
-  const { query } = args || {};
-  const isMobile = useIsMobile();
-
-  return (
-    <Flexbox gap={8}>
-      <Flexbox horizontal={!isMobile} gap={isMobile ? 8 : 40}>
-        <Flexbox horizontal align="center" className={styles.query} gap={8}>
-          <Icon icon={SearchIcon} />
-          {query ? query : <Skeleton.Block active style={{ height: 20, width: 40 }} />}
-        </Flexbox>
-        <Skeleton.Block active style={{ height: 20, width: 40 }} />
-      </Flexbox>
-      <Flexbox horizontal gap={12}>
-        {[1, 2, 3, 4, 5].map((id) => (
-          <Skeleton.Button active key={id} style={{ borderRadius: 8, height: 80, width: 160 }} />
-        ))}
-      </Flexbox>
-    </Flexbox>
-  );
-});
-```
-
-### Placeholder rules
-
-- **Mirror the eventual Render's layout.** When the result arrives the Placeholder unmounts and the Render mounts; if they share dimensions, the chat doesn't jump.
-- Use `Skeleton.Block` / `Skeleton.Button` from `@lobehub/ui` for placeholder shapes.
-- Embed any args you have (e.g. the query text) — context helps the user know what's loading.
-- Pulse with `shinyTextStyles.shinyText` if the Placeholder includes literal text.
-
-### Placeholder registry — `client/Placeholder/index.ts`
-
-```ts
-import { WebBrowsingApiName } from '../../types';
-import CrawlMultiPages from './CrawlMultiPages';
-import CrawlSinglePage from './CrawlSinglePage';
-import { Search } from './Search';
-
-export const WebBrowsingPlaceholders = {
-  [WebBrowsingApiName.crawlMultiPages]: CrawlMultiPages,
-  [WebBrowsingApiName.crawlSinglePage]: CrawlSinglePage,
-  [WebBrowsingApiName.search]: Search,
-};
-
-export { CrawlMultiPages, CrawlSinglePage, Search };
-```
-
----
-
-## 4. Streaming — Live Output During Execution (optional)
-
-**Lifecycle:** rendered **while the executor is still running** for APIs that emit incremental output. The component is responsible for fetching the in-flight stream from the chat store and rendering it.
-
-**Add for** long-running ops with continuous output: shell command execution (stdout/stderr), file write progress, code interpreter cells.
-
-### Props (`BuiltinStreamingProps<Args>`)
-
-```ts
-interface BuiltinStreamingProps<Arguments = any> {
-  apiName: string;
-  args: Arguments;
-  identifier: string;
-  messageId: string; // use to fetch the streaming buffer from store
-  toolCallId: string;
-}
-```
-
-Note there's **no `state` or `result` prop** — the Streaming component is for the in-flight phase. It pulls the live buffer from the store itself (typically via `chatToolSelectors.streamingContent(messageId)` or similar).
-
-### Canonical example — RunCommandStreaming
-
-`packages/builtin-tool-local-system/src/client/Streaming/RunCommand/index.tsx`:
-
-```tsx
-'use client';
-
-import type { BuiltinStreamingProps } from '@lobechat/types';
-import { Highlighter } from '@lobehub/ui';
-import { memo } from 'react';
-
-interface RunCommandParams {
-  command?: string;
-  description?: string;
-  timeout?: number;
-}
-
-export const RunCommandStreaming = memo<BuiltinStreamingProps<RunCommandParams>>(({ args }) => {
-  const { command } = args || {};
-  if (!command) return null;
-
-  return (
-    <Highlighter
-      animated
-      wrap
-      language="sh"
-      showLanguage={false}
-      style={{ padding: '4px 8px' }}
-      variant="outlined"
-    >
-      {command}
-    </Highlighter>
-  );
-});
-RunCommandStreaming.displayName = 'RunCommandStreaming';
-```
-
-For real-time output beyond just the command (stderr/stdout streaming), pull from the chat store:
-
-```tsx
-const buffer = useChatStore((state) =>
-  chatToolSelectors.streamingBuffer(messageId, toolCallId)(state),
-);
-```
-
-### Streaming rules
-
-- Render `null` until you have something to display (avoids flash).
-- For terminal-style output, use `Highlighter` with `animated` to show typing-like effect.
-- The Streaming component must **unmount cleanly** when execution ends — typically the framework swaps it out for the Render automatically.
-
-### Streaming registry — `client/Streaming/index.ts`
-
-```ts
-import { LocalSystemApiName } from '../..';
-import { RunCommandStreaming } from './RunCommand';
-import { WriteFileStreaming } from './WriteFile';
-
-export const LocalSystemStreamings = {
-  [LocalSystemApiName.runCommand]: RunCommandStreaming,
-  [LocalSystemApiName.writeLocalFile]: WriteFileStreaming,
-};
-```
-
----
-
-## 5. Intervention — Approval / Edit-Before-Run (optional)
-
-**Lifecycle:** rendered **before the executor runs** for APIs whose manifest sets `humanIntervention`. The user sees a preview of the args, can edit them, then approves or skips/cancels.
-
-**Add for** destructive or sensitive ops: shell commands, file writes, file moves, payments, message broadcasts.
-
-### Props (`BuiltinInterventionProps<Args>`)
-
-```ts
-interface BuiltinInterventionProps<Arguments = any> {
-  apiName?: string;
-  args: Arguments;
-  identifier?: string;
-  interactionMode?: 'approval' | 'custom';
-  messageId: string;
-
-  /** Called when the user edits the args; the approve action awaits this. */
-  onArgsChange?: (args: Arguments) => void | Promise<void>;
-
-  /** Called on approve / skip / cancel. */
-  onInteractionAction?: (
-    action:
-      | { type: 'submit'; payload: Record<string, unknown> }
-      | { type: 'skip'; payload?: Record<string, unknown>; reason?: string }
-      | { type: 'cancel'; payload?: Record<string, unknown> },
-  ) => Promise<void>;
-
-  /** Register a callback to flush pending saves before approval. Returns cleanup. */
-  registerBeforeApprove?: (id: string, callback: () => void | Promise<void>) => () => void;
-}
-```
-
-### Canonical example — RunCommand Intervention
-
-`packages/builtin-tool-local-system/src/client/Intervention/RunCommand/index.tsx`:
-
-```tsx
-import type { RunCommandParams } from '@lobechat/electron-client-ipc';
-import type { BuiltinInterventionProps } from '@lobechat/types';
-import { Flexbox, Highlighter, Text } from '@lobehub/ui';
-import { memo } from 'react';
-
-const RunCommand = memo<BuiltinInterventionProps<RunCommandParams>>(({ args }) => {
-  const { description, command, timeout } = args;
-  return (
-    <Flexbox gap={8}>
-      <Flexbox horizontal justify="space-between">
-        {description && <Text>{description}</Text>}
-        {timeout && (
-          <Text style={{ fontSize: 12 }} type="secondary">
-            timeout: {formatTimeout(timeout)}
-          </Text>
-        )}
-      </Flexbox>
-      {command && (
-        <Highlighter wrap language="sh" showLanguage={false} variant="outlined">
-          {command}
-        </Highlighter>
-      )}
-    </Flexbox>
-  );
-});
-export default RunCommand;
-```
-
-### Intervention rules
-
-- **Show a preview, not a form by default.** Editing UI is opt-in via `onArgsChange` and is usually inline (click to edit a code block, etc.).
-- For args with debounced edit state (text fields), use `registerBeforeApprove(id, flushFn)` so the approve action waits for the debounce to flush. Always return the cleanup function.
-- Call `onInteractionAction({ type: 'submit', payload })` when the user approves; `'skip'` if they skip with a reason; `'cancel'` if they cancel the whole turn.
-- Add a corresponding `interventionAudit.ts` in the package root if the tool needs scope/path validation before approval (see `local-system/src/interventionAudit.ts`).
-
-### Intervention registry — `client/Intervention/index.ts`
-
-```ts
-import { LocalSystemApiName } from '../..';
-import EditLocalFile from './EditLocalFile';
-import RunCommand from './RunCommand';
-import WriteFile from './WriteFile';
-/* … */
-
-export const LocalSystemInterventions = {
-  [LocalSystemApiName.editLocalFile]: EditLocalFile,
-  [LocalSystemApiName.runCommand]: RunCommand,
-  [LocalSystemApiName.writeLocalFile]: WriteFile,
-  /* one entry per API that needs approval */
-};
-```
-
----
-
-## 6. Portal — Full-Screen Detail View (optional)
-
-**Lifecycle:** rendered when the user opens the tool message in a side panel or full-screen modal. One Portal per **tool**, not per API — the Portal switches on `apiName` internally.
-
-**Add for** tools whose results deserve a deep-dive view: search results with editable filters, page content with reader mode, code interpreter sessions.
-
-### Props (`BuiltinPortalProps<Args, State>`)
-
-```ts
-interface BuiltinPortalProps<Arguments = Record<string, any>, State = any> {
-  apiName?: string;
-  arguments: Arguments;
-  identifier: string;
-  messageId: string;
-  state: State;
-}
-```
-
-### Canonical example — Web-Browsing Portal
-
-`packages/builtin-tool-web-browsing/src/client/Portal/index.tsx`:
-
-```tsx
-import type { BuiltinPortalProps, CrawlPluginState, SearchQuery } from '@lobechat/types';
-import { memo } from 'react';
-
-import { WebBrowsingApiName } from '../../types';
-import PageContent from './PageContent';
-import PageContents from './PageContents';
-import Search from './Search';
-
-const Portal = memo<BuiltinPortalProps>(({ arguments: args, messageId, state, apiName }) => {
-  switch (apiName) {
-    case WebBrowsingApiName.search:
-      return <Search messageId={messageId} query={args as SearchQuery} response={state} />;
-
-    case WebBrowsingApiName.crawlSinglePage: {
-      const result = (state as CrawlPluginState).results.find((r) => r.originalUrl === args.url);
-      return <PageContent messageId={messageId} result={result} />;
-    }
-
-    case WebBrowsingApiName.crawlMultiPages:
-      return (
-        <PageContents
-          messageId={messageId}
-          results={(state as CrawlPluginState).results}
-          urls={args.urls}
-        />
-      );
-  }
-  return null;
-});
-export default Portal;
-```
-
-### Portal rules
-
-- One Portal per tool — the file is the routing layer, subcomponents implement each API's view.
-- Portals can read the chat store directly to detect "still streaming" and render a Skeleton internally (see `Search/index.tsx:20-46`).
-- Layout assumes more space than the Render — use `Flexbox` with `height={'100%'}` and structure for a side panel viewport.
-
-### Portal registry — `packages/builtin-tools/src/portals.ts`
-
-```ts
-import { WebBrowsingManifest, WebBrowsingPortal } from '@lobechat/builtin-tool-web-browsing/client';
-import { type BuiltinPortal } from '@lobechat/types';
-
-export const BuiltinToolsPortals: Record<string, BuiltinPortal> = {
-  [WebBrowsingManifest.identifier]: WebBrowsingPortal as BuiltinPortal,
-};
-```
-
----
-
-## 7. `client/components/` — Shared Subcomponents
-
-Cross-cutting building blocks used by multiple surfaces live here, not duplicated in each surface folder.
-
-Examples from `web-browsing/src/client/components/`:
-
-- `CategoryAvatar.tsx` — search category icon
-- `EngineAvatar.tsx` — search engine logo (used in Inspector chip + Render list + Portal header)
-- `SearchBar.tsx` — editable query bar (used in Render and Portal)
-
-Examples from `local-system/src/client/components/`:
-
-- `FileItem.tsx` — single file row (used in ListFiles Render, SearchFiles Render, MoveLocalFiles Render)
-- `FilePathDisplay.tsx` — path with truncation (used everywhere)
-
-### Rules
-
-- Live under `client/components/`, exported via `client/components/index.ts`.
-- Re-export from `client/index.ts` only if other packages need them; otherwise keep internal.
-- Keep them dumb — props in, JSX out, no store reads. The store reads belong in the surface that composes them.
-
----
-
-## 8. `client/index.ts` — Package Public API
-
-Re-exports everything the registries need plus useful types/manifest:
-
-```ts
-// Inspector — required
-export { TaskInspectors } from './Inspector';
-
-// Render — only if any API has one
-export { TaskRenders, CreateTaskRender, RunTasksRender } from './Render';
-
-// Placeholder / Streaming / Intervention — only if used
-export { LocalSystemListFilesPlaceholder, LocalSystemSearchFilesPlaceholder } from './Placeholder';
-export { LocalSystemStreamings } from './Streaming';
-export { LocalSystemInterventions } from './Intervention';
-
-// Portal — single export per tool
-export { default as WebBrowsingPortal } from './Portal';
-
-// Reusable components if other packages need them
-export { CategoryAvatar, EngineAvatar, SearchBar } from './components';
-
-// Re-export manifest, identifier, types for convenience
-export { TaskManifest, TaskIdentifier } from '../manifest';
-export * from '../types';
-```
-
----
-
-## 9. Diagnostic Quick-Lookup
-
-| Symptom                                         | Surface to check                                                                                                  |     |                           |
-| ----------------------------------------------- | ----------------------------------------------------------------------------------------------------------------- | --- | ------------------------- |
-| No header at all on the tool call               | Inspector missing from `client/Inspector/index.ts` registry                                                       |     |                           |
-| Header shows the API name but no chips          | Inspector missing \`args?.X                                                                                       |     | partialArgs?.X\` fallback |
-| Header doesn't pulse during loading             | Missing `shinyTextStyles.shinyText` on `isArgumentsStreaming \|\| isLoading`                                      |     |                           |
-| Empty result card under header                  | Render returned `<div />` instead of `null` when no data                                                          |     |                           |
-| Layout jump when result arrives                 | Placeholder dimensions don't match Render dimensions                                                              |     |                           |
-| Approval dialog never appears                   | Manifest missing `humanIntervention`, or Intervention not in registry                                             |     |                           |
-| Approval click doesn't wait for inline edit     | Missing `registerBeforeApprove(id, flushFn)`                                                                      |     |                           |
-| Portal opens but blank                          | Switch in `Portal/index.tsx` doesn't cover the apiName                                                            |     |                           |
-| Strings show as `builtins.lobe-foo.apiName.bar` | Missing i18n key in `src/locales/default/plugin.ts` (or not seeded in dev locale files)                           |     |                           |
-| Wrong color shade on `<Text type="secondary">`  | `type='secondary'` is lighter than `colorTextSecondary` — pass via `style={{ color: cssVar.colorTextSecondary }}` |     |                           |

.agents/skills/chat-sdk/SKILL.md

@@ -1,13 +1,7 @@
 ---
 name: chat-sdk
-description: >
-  Build multi-platform chat bots with Chat SDK (`chat` npm package). Use when developers want to
-  (1) Build a Slack, Teams, Google Chat, Discord, GitHub, or Linear bot,
-  (2) Use the Chat SDK to handle mentions, messages, reactions, slash commands, cards, modals, or streaming,
-  (3) Set up webhook handlers for chat platforms,
-  (4) Send interactive cards or stream AI responses to chat platforms.
-  Triggers on "chat sdk", "chat bot", "slack bot", "teams bot", "discord bot", "@chat-adapter",
-  building bots that work across multiple chat platforms.
+description: 'Build multi-platform chat bots with the chat SDK. Use for Slack, Teams, Google Chat, Discord, GitHub, Linear bots, webhooks, mentions, slash commands, cards, modals, or streaming responses.'
+user-invocable: false
 ---
 
 # Chat SDK

.agents/skills/cli-backend-testing/SKILL.md

@@ -1,218 +0,0 @@
----
-name: cli-backend-testing
-description: >
-  CLI + Backend integration testing workflow. Use when verifying backend API changes
-  (TRPC routers, services, models) via the LobeHub CLI against a local dev server.
-  Triggers on 'cli test', 'test with cli', 'verify with cli', 'local cli test',
-  'backend test with cli', or when needing to validate server-side changes end-to-end.
----
-
-# CLI + Backend Integration Testing
-
-Standard workflow for verifying backend changes using the LobeHub CLI (`lh`) against a local dev server.
-
-## When to Use
-
-- Verifying TRPC router / service / model changes end-to-end
-- Testing new API fields or response structure changes
-- Validating CLI command output after backend modifications
-- Debugging data flow issues between server and CLI
-
-## Prerequisites
-
-| Requirement  | Details                                                       |
-| ------------ | ------------------------------------------------------------- |
-| Dev server   | `localhost:3011` (Next.js)                                    |
-| CLI source   | `lobehub/apps/cli/`                                           |
-| CLI dev mode | Uses `LOBEHUB_CLI_HOME=.lobehub-dev` for isolated credentials |
-| Auth         | Device Code Flow login to local server                        |
-
-## Quick Reference
-
-All CLI dev commands run from `lobehub/apps/cli/`:
-
-```bash
-# Shorthand for all commands below
-CLI="LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts"
-```
-
-## Workflow
-
-### Step 1: Ensure Dev Server is Running
-
-Check if the dev server is already running:
-
-```bash
-curl -s -o /dev/null -w '%{http_code}' http://localhost:3011/ 2> /dev/null
-```
-
-- **If reachable** (returns any HTTP status): server is running. Skip to Step 2.
-- **If unreachable**: start the server:
-
-```bash
-# From cloud repo root
-pnpm run dev:next
-```
-
-To **restart** (pick up server-side code changes):
-
-```bash
-lsof -ti:3011 | xargs kill
-pnpm run dev:next
-```
-
-**Important:** Server-side code changes in the submodule (`lobehub/src/server/`, `lobehub/packages/`) require a server restart. Next.js hot-reload may not pick up changes in submodule packages.
-
-### Step 2: Check CLI Authentication
-
-Check if dev credentials already exist:
-
-```bash
-cat lobehub/apps/cli/.lobehub-dev/settings.json 2> /dev/null
-```
-
-- **If file exists and contains `"serverUrl": "http://localhost:3011"`**: already authenticated. Skip to Step 3.
-- **If file missing or points to wrong server**: login is needed. Ask the user to run:
-
-```bash
-! cd lobehub/apps/cli && LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts login --server http://localhost:3011
-```
-
-> Login requires interactive browser authorization (OIDC Device Code Flow), so the user must run it themselves via `!` prefix. After login, credentials are saved to `lobehub/apps/cli/.lobehub-dev/` and persist across sessions.
-
-### Step 3: Test with CLI Commands
-
-CLI runs from source (`bun src/index.ts`), so CLI-side code changes take effect immediately without rebuilding.
-
-```bash
-cd lobehub/apps/cli
-LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts <command>
-```
-
-### Step 4: Clean Up Test Data
-
-Delete any test data created during verification:
-
-```bash
-LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts task delete < id > -y
-LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts agent delete < id > -y
-```
-
-## Common Testing Patterns
-
-### Task System
-
-```bash
-# List tasks
-$CLI task list
-
-# Create test data with nesting
-$CLI task create -n "Root Task" -i "Test instruction"
-$CLI task create -n "Child Task" -i "Sub instruction" --parent T-1
-
-# View task detail (tests getTaskDetail service)
-$CLI task view T-1
-
-# View task tree
-$CLI task tree T-1
-
-# Test lifecycle
-$CLI task edit T-1 --status running
-$CLI task comment T-1 -m "Test comment"
-
-# Clean up
-$CLI task delete T-1 -y
-```
-
-### Agent System
-
-```bash
-# List agents
-$CLI agent list
-
-# View agent detail
-$CLI agent view <agent-id>
-
-# Run agent (tests agent execution pipeline)
-$CLI agent run <agent-id> -m "Test prompt"
-```
-
-### Document & Knowledge Base
-
-```bash
-# List documents
-$CLI doc list
-
-# Create and view
-$CLI doc create -t "Test Doc" -c "Content here"
-$CLI doc view <doc-id>
-
-# Knowledge base
-$CLI kb list
-$CLI kb tree <kb-id>
-```
-
-### Model & Provider
-
-```bash
-# List models and providers
-$CLI model list
-$CLI provider list
-
-# Test provider connectivity
-$CLI provider test <provider-id>
-```
-
-## Dev-Test Cycle
-
-The standard cycle for backend development:
-
-```
-1. Make code changes (service/model/router/type)
-         |
-2. Run unit tests (fast feedback)
-   bunx vitest run --silent='passed-only' '<test-file>'
-         |
-3. Restart dev server (if server-side changes)
-   lsof -ti:3011 | xargs kill && pnpm run dev:next
-         |
-4. CLI verification (end-to-end)
-   LOBEHUB_CLI_HOME=.lobehub-dev bun src/index.ts <command>
-         |
-5. Clean up test data
-```
-
-### When Server Restart is Needed
-
-| Change Location                           | Restart? |
-| ----------------------------------------- | -------- |
-| `lobehub/src/server/` (routers, services) | Yes      |
-| `lobehub/packages/database/` (models)     | Yes      |
-| `lobehub/packages/types/`                 | Yes      |
-| `lobehub/packages/prompts/`               | Yes      |
-| `lobehub/apps/cli/` (CLI code)            | No       |
-| `src/` (cloud overrides)                  | Yes      |
-
-### When Server Restart is NOT Needed
-
-CLI runs from source via `bun src/index.ts`, so any changes to `lobehub/apps/cli/src/` take effect immediately on next command invocation.
-
-## Troubleshooting
-
-| Issue                       | Solution                                                              |
-| --------------------------- | --------------------------------------------------------------------- |
-| `No authentication found`   | Run `login --server http://localhost:3011`                            |
-| `UNAUTHORIZED` on API calls | Token expired; re-run login                                           |
-| `ECONNREFUSED`              | Dev server not running; start with `pnpm run dev:next`                |
-| CLI shows old data/behavior | Server needs restart to pick up code changes                          |
-| `EADDRINUSE` on port 3011   | Server already running; kill with `lsof -ti:3011 \| xargs kill`       |
-| Login opens wrong server    | Must use `--server http://localhost:3011` flag (env var doesn't work) |
-
-## Credential Isolation
-
-| Mode       | Credential Dir                   | Server            |
-| ---------- | -------------------------------- | ----------------- |
-| Dev        | `lobehub/apps/cli/.lobehub-dev/` | `localhost:3011`  |
-| Production | `~/.lobehub/`                    | `app.lobehub.com` |
-
-The two environments are completely isolated. Dev mode credentials are gitignored.

.agents/skills/cli/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: cli
-description: LobeHub CLI (@lobehub/cli) development guide. Use when working on CLI commands, adding new subcommands, fixing CLI bugs, or understanding CLI architecture. Triggers on CLI development, command implementation, or `lh` command questions.
+description: LobeHub CLI (@lobehub/cli) development guide — commands, subcommands, architecture.
 disable-model-invocation: true
 ---
 

.agents/skills/cli/references/generate.md

@@ -111,7 +111,7 @@ Generate video from text prompt. This is an async operation.
 **Source**: `apps/cli/src/commands/generate/video.ts`
 
 ```bash
-lh gen video "A cat playing piano" -m <model> -p <provider> [options]
+lh gen video "A cat playing piano" -m < model > -p < provider > [options]
 ```
 
 | Option                      | Description              | Required |
@@ -259,13 +259,13 @@ Image and video generation use an async task pattern:
      UUID from the `async_tasks` table, not `gen_xxx`
    - Returns `{ status, error, generation }` (generation includes asset URLs on success)
    - Before querying, calls `checkTimeoutTasks` which marks tasks as `error` if they have been
-     `pending` or `processing` for more than ~5 minutes (`ASYNC_TASK_TIMEOUT = 298s`)
+     `pending` or `processing` for more than \~5 minutes (`ASYNC_TASK_TIMEOUT = 298s`)
 
 **Server routes**:
 
-- `src/server/routers/lambda/image/index.ts` — image creation (uses `authedProcedure` + `serverDatabase`)
-- `src/server/routers/lambda/video/index.ts` — video creation (uses `authedProcedure` + `serverDatabase`)
-- `src/server/routers/lambda/generation.ts` — status checking
+- `apps/server/src/routers/lambda/image/index.ts` — image creation (uses `authedProcedure` + `serverDatabase`)
+- `apps/server/src/routers/lambda/video/index.ts` — video creation (uses `authedProcedure` + `serverDatabase`)
+- `apps/server/src/routers/lambda/generation.ts` — status checking
 - `packages/database/src/models/asyncTask.ts` — `AsyncTaskModel` including `checkTimeoutTasks`
 
 **Note**: Image/video routes do NOT use the `keyVaults` middleware — they read API keys from the database via `initModelRuntimeFromDB` or `createAsyncCaller`.

.agents/skills/data-fetching-architecture/SKILL.md

@@ -0,0 +1,614 @@
+---
+name: data-fetching-architecture
+description: 'LobeHub data-fetching pipeline guide. Use for service layer, Zustand store, SWR, lambdaClient, useClientDataSWR, useFetchXxx hooks, or migrating useEffect fetches.'
+user-invocable: false
+---
+
+# LobeHub Data Fetching Architecture
+
+> **Related:** `store-data-structures` covers List vs Detail data shape rationale (Map vs Array).
+
+## Architecture Overview
+
+```text
+┌─────────────┐
+│  Component  │
+└──────┬──────┘
+       │ 1. Call useFetchXxx hook from store
+       ↓
+┌──────────────────┐
+│  Zustand Store   │
+│  (State + Hook)  │
+└──────┬───────────┘
+       │ 2. useClientDataSWR calls service
+       ↓
+┌──────────────────┐
+│  Service Layer   │
+│  (xxxService)    │
+└──────┬───────────┘
+       │ 3. Call lambdaClient
+       ↓
+┌──────────────────┐
+│  lambdaClient    │
+│  (TRPC Client)   │
+└──────────────────┘
+```
+
+## Core Principles
+
+### ✅ DO
+
+1. **Use Service Layer** for all API calls
+2. **Use Store SWR Hooks** for data fetching (not useEffect)
+3. **Use proper data structures** — see `store-data-structures` skill for List vs Detail patterns
+4. **Use lambdaClient.mutate** for write operations (create/update/delete)
+5. **Use lambdaClient.query** only inside service methods
+6. **Naming convention** — read hooks are `useFetchXxx`, cache invalidation helpers are `refreshXxx` (e.g. `useFetchBenchmarks` / `refreshBenchmarks`). Mutations then chain `refreshXxx()` after the service call.
+
+### ❌ DON'T
+
+1. **Never use useEffect** for data fetching
+2. **Never call lambdaClient** directly in components or stores
+3. **Never use useState** for server data
+4. **Never mix data structure patterns** — follow `store-data-structures` skill
+
+---
+
+## Layer 1: Service Layer
+
+### Purpose
+
+- Encapsulate all API calls to lambdaClient
+- Provide clean, typed interfaces
+- Single source of truth for API operations
+
+### Service Structure
+
+```typescript
+// src/services/agentEval.ts
+class AgentEvalService {
+  // Query methods - READ operations
+  async listBenchmarks() {
+    return lambdaClient.agentEval.listBenchmarks.query();
+  }
+
+  async getBenchmark(id: string) {
+    return lambdaClient.agentEval.getBenchmark.query({ id });
+  }
+
+  // Mutation methods - WRITE operations
+  async createBenchmark(params: CreateBenchmarkParams) {
+    return lambdaClient.agentEval.createBenchmark.mutate(params);
+  }
+
+  async updateBenchmark(params: UpdateBenchmarkParams) {
+    return lambdaClient.agentEval.updateBenchmark.mutate(params);
+  }
+
+  async deleteBenchmark(id: string) {
+    return lambdaClient.agentEval.deleteBenchmark.mutate({ id });
+  }
+}
+
+export const agentEvalService = new AgentEvalService();
+```
+
+### Service Guidelines
+
+1. **One service per domain** (e.g., agentEval, ragEval, aiAgent)
+2. **Export singleton instance** (`export const xxxService = new XxxService()`)
+3. **Method names match operations** (list, get, create, update, delete)
+4. **Clear parameter types** (use interfaces for complex params)
+
+---
+
+## Layer 2: Store with SWR Hooks
+
+### Purpose
+
+- Manage client-side state
+- Provide SWR hooks for data fetching
+- Handle cache invalidation
+
+### State Structure
+
+```typescript
+// src/store/eval/slices/benchmark/initialState.ts
+export interface BenchmarkSliceState {
+  // List data - simple array
+  benchmarkList: AgentEvalBenchmarkListItem[];
+  benchmarkListInit: boolean;
+
+  // Detail data - map for caching
+  benchmarkDetailMap: Record<string, AgentEvalBenchmark>;
+  loadingBenchmarkDetailIds: string[];
+
+  // Mutation states
+  isCreatingBenchmark: boolean;
+  isUpdatingBenchmark: boolean;
+  isDeletingBenchmark: boolean;
+}
+```
+
+> For complete initialState, reducer, and internal dispatch patterns, see the `store-data-structures` skill.
+
+### Actions
+
+```typescript
+// src/store/eval/slices/benchmark/action.ts
+const FETCH_BENCHMARKS_KEY = 'FETCH_BENCHMARKS';
+const FETCH_BENCHMARK_DETAIL_KEY = 'FETCH_BENCHMARK_DETAIL';
+
+export interface BenchmarkAction {
+  // SWR Hooks - for data fetching
+  useFetchBenchmarks: () => SWRResponse;
+  useFetchBenchmarkDetail: (id?: string) => SWRResponse;
+
+  // Refresh methods - for cache invalidation
+  refreshBenchmarks: () => Promise<void>;
+  refreshBenchmarkDetail: (id: string) => Promise<void>;
+
+  // Mutation actions
+  createBenchmark: (params: CreateParams) => Promise<any>;
+  updateBenchmark: (params: UpdateParams) => Promise<void>;
+  deleteBenchmark: (id: string) => Promise<void>;
+
+  // Internal methods - not for direct UI use
+  internal_dispatchBenchmarkDetail: (payload: BenchmarkDetailDispatch) => void;
+  internal_updateBenchmarkDetailLoading: (id: string, loading: boolean) => void;
+}
+
+export const createBenchmarkSlice: StateCreator<EvalStore, any, [], BenchmarkAction> = (
+  set,
+  get,
+) => ({
+  // Fetch list — simple array stored in benchmarkList
+  useFetchBenchmarks: () =>
+    useClientDataSWR(FETCH_BENCHMARKS_KEY, () => agentEvalService.listBenchmarks(), {
+      onSuccess: (data) => {
+        set({ benchmarkList: data, benchmarkListInit: true }, false, 'useFetchBenchmarks/success');
+      },
+    }),
+
+  // Fetch detail — null key disables the request when id is missing
+  useFetchBenchmarkDetail: (id) =>
+    useClientDataSWR(
+      id ? [FETCH_BENCHMARK_DETAIL_KEY, id] : null,
+      () => agentEvalService.getBenchmark(id!),
+      {
+        onSuccess: (data) => {
+          get().internal_dispatchBenchmarkDetail({
+            type: 'setBenchmarkDetail',
+            id: id!,
+            value: data,
+          });
+          get().internal_updateBenchmarkDetailLoading(id!, false);
+        },
+      },
+    ),
+
+  // Refresh methods
+  refreshBenchmarks: () => mutate(FETCH_BENCHMARKS_KEY),
+  refreshBenchmarkDetail: (id) => mutate([FETCH_BENCHMARK_DETAIL_KEY, id]),
+
+  // CREATE — refresh list after creation
+  createBenchmark: async (params) => {
+    set({ isCreatingBenchmark: true }, false, 'createBenchmark/start');
+    try {
+      const result = await agentEvalService.createBenchmark(params);
+      await get().refreshBenchmarks();
+      return result;
+    } finally {
+      set({ isCreatingBenchmark: false }, false, 'createBenchmark/end');
+    }
+  },
+
+  // UPDATE — optimistic update + refresh
+  updateBenchmark: async (params) => {
+    const { id } = params;
+
+    // 1. Optimistic update
+    get().internal_dispatchBenchmarkDetail({
+      type: 'updateBenchmarkDetail',
+      id,
+      value: params,
+    });
+    // 2. Set loading
+    get().internal_updateBenchmarkDetailLoading(id, true);
+
+    try {
+      // 3. Call service
+      await agentEvalService.updateBenchmark(params);
+      // 4. Refresh from server
+      await get().refreshBenchmarks();
+      await get().refreshBenchmarkDetail(id);
+    } finally {
+      get().internal_updateBenchmarkDetailLoading(id, false);
+    }
+  },
+
+  // DELETE — optimistic update + refresh
+  deleteBenchmark: async (id) => {
+    get().internal_dispatchBenchmarkDetail({ type: 'deleteBenchmarkDetail', id });
+    get().internal_updateBenchmarkDetailLoading(id, true);
+
+    try {
+      await agentEvalService.deleteBenchmark(id);
+      await get().refreshBenchmarks();
+    } finally {
+      get().internal_updateBenchmarkDetailLoading(id, false);
+    }
+  },
+
+  // Internal — dispatch to reducer (for detail map)
+  internal_dispatchBenchmarkDetail: (payload) => {
+    const currentMap = get().benchmarkDetailMap;
+    const nextMap = benchmarkDetailReducer(currentMap, payload);
+
+    // Skip set when nothing changed — avoids unnecessary re-renders
+    if (isEqual(nextMap, currentMap)) return;
+    set({ benchmarkDetailMap: nextMap }, false, `dispatchBenchmarkDetail/${payload.type}`);
+  },
+
+  // Internal — update loading state for specific detail
+  internal_updateBenchmarkDetailLoading: (id, loading) => {
+    set(
+      (state) => ({
+        loadingBenchmarkDetailIds: loading
+          ? [...state.loadingBenchmarkDetailIds, id]
+          : state.loadingBenchmarkDetailIds.filter((i) => i !== id),
+      }),
+      false,
+      'updateBenchmarkDetailLoading',
+    );
+  },
+});
+```
+
+### Store Guidelines
+
+1. **SWR keys as constants** at top of file
+2. **useClientDataSWR** for all data fetching (never useEffect)
+3. **onSuccess callback** updates store state
+4. **Refresh methods** use `mutate()` to invalidate cache
+5. **Loading states** in initialState, updated in onSuccess
+6. **Mutations** call service, then refresh relevant cache
+
+---
+
+## Layer 3: Component Usage
+
+### Fetching List Data
+
+```tsx
+// ✅ CORRECT
+const BenchmarkList = () => {
+  // 1. Get the hook from store
+  const useFetchBenchmarks = useEvalStore((s) => s.useFetchBenchmarks);
+
+  // 2. Get list data
+  const benchmarks = useEvalStore((s) => s.benchmarkList);
+  const isInit = useEvalStore((s) => s.benchmarkListInit);
+
+  // 3. Call the hook (SWR handles the data fetching)
+  useFetchBenchmarks();
+
+  // 4. Use the data
+  if (!isInit) return <Loading />;
+  return (
+    <div>
+      <h2>Total: {benchmarks.length}</h2>
+      {benchmarks.map((b) => (
+        <BenchmarkCard key={b.id} {...b} />
+      ))}
+    </div>
+  );
+};
+```
+
+### Fetching Detail Data
+
+```tsx
+// ✅ CORRECT
+const BenchmarkDetail = () => {
+  const { benchmarkId } = useParams<{ benchmarkId: string }>();
+
+  const useFetchBenchmarkDetail = useEvalStore((s) => s.useFetchBenchmarkDetail);
+
+  // Detail from map
+  const benchmark = useEvalStore((s) =>
+    benchmarkId ? s.benchmarkDetailMap[benchmarkId] : undefined,
+  );
+
+  // Per-item loading
+  const isLoading = useEvalStore((s) =>
+    benchmarkId ? s.loadingBenchmarkDetailIds.includes(benchmarkId) : false,
+  );
+
+  useFetchBenchmarkDetail(benchmarkId);
+
+  if (!benchmark) return <Loading />;
+  return (
+    <div>
+      <h1>{benchmark.name}</h1>
+      <p>{benchmark.description}</p>
+      {isLoading && <Spinner />}
+    </div>
+  );
+};
+```
+
+### Using Selectors (Recommended)
+
+```typescript
+// src/store/eval/slices/benchmark/selectors.ts
+export const benchmarkSelectors = {
+  getBenchmarkDetail: (id: string) => (s: EvalStore) => s.benchmarkDetailMap[id],
+  isLoadingBenchmarkDetail: (id: string) => (s: EvalStore) =>
+    s.loadingBenchmarkDetailIds.includes(id),
+};
+
+// Component with selectors
+const BenchmarkDetail = () => {
+  const { benchmarkId } = useParams();
+  const useFetchBenchmarkDetail = useEvalStore((s) => s.useFetchBenchmarkDetail);
+  const benchmark = useEvalStore(benchmarkSelectors.getBenchmarkDetail(benchmarkId!));
+
+  useFetchBenchmarkDetail(benchmarkId);
+
+  return <div>{benchmark && <h1>{benchmark.name}</h1>}</div>;
+};
+```
+
+### Anti-pattern
+
+```tsx
+// ❌ WRONG — Don't use useEffect for data fetching
+const BenchmarkList = () => {
+  const [data, setData] = useState([]);
+  const [loading, setLoading] = useState(false);
+
+  useEffect(() => {
+    setLoading(true);
+    lambdaClient.agentEval.listBenchmarks
+      .query()
+      .then(setData)
+      .finally(() => setLoading(false));
+  }, []);
+
+  return <div>...</div>;
+};
+```
+
+### Mutations in Components
+
+```tsx
+// Create — global mutation flag drives form loading
+const CreateBenchmarkModal = () => {
+  const createBenchmark = useEvalStore((s) => s.createBenchmark);
+  const isCreating = useEvalStore((s) => s.isCreatingBenchmark);
+
+  const handleSubmit = async (values) => {
+    try {
+      // Optimistic update + refresh happen inside createBenchmark
+      await createBenchmark(values);
+      message.success('Created successfully');
+      onClose();
+    } catch (error) {
+      message.error('Failed to create');
+    }
+  };
+
+  return (
+    <Form onSubmit={handleSubmit} loading={isCreating}>
+      ...
+    </Form>
+  );
+};
+
+// Update / delete — per-item loading so only the row being mutated spins
+const BenchmarkItem = ({ id }: { id: string }) => {
+  const updateBenchmark = useEvalStore((s) => s.updateBenchmark);
+  const deleteBenchmark = useEvalStore((s) => s.deleteBenchmark);
+  const isLoading = useEvalStore(benchmarkSelectors.isLoadingBenchmarkDetail(id));
+
+  const handleUpdate = async (data) => {
+    await updateBenchmark({ id, ...data });
+  };
+
+  const handleDelete = async () => {
+    await deleteBenchmark(id);
+  };
+
+  return (
+    <div>
+      {isLoading && <Spinner />}
+      <button onClick={handleUpdate}>Update</button>
+      <button onClick={handleDelete}>Delete</button>
+    </div>
+  );
+};
+```
+
+**Why two patterns:** create has no id yet, so a single `isCreatingXxx` flag is enough. Update/delete target a specific row, so global flags would freeze unrelated rows — keep per-item state in `loadingXxxIds`.
+
+---
+
+## Need a fuller worked example?
+
+The canonical `Benchmark` example above is the one to copy for a flat list + detail map. If you need to maintain a list **keyed by a parent id** (e.g. `datasetMap[benchmarkId]` because the same shape appears under multiple parents), read [`references/walkthrough.md`](./references/walkthrough.md) — it walks through the full 6 steps (service → reducer → slice → store wiring → selectors → component) for that variant.
+
+---
+
+## Common Patterns
+
+### Pattern 1: Pagination
+
+Cache key array must include every parameter that should trigger a refetch.
+
+```typescript
+useFetchTestCases: (params: { datasetId: string; limit: number; offset: number }) =>
+  useClientDataSWR(
+    params.datasetId ? [FETCH_TEST_CASES_KEY, params.datasetId, params.limit, params.offset] : null,
+    () => agentEvalService.listTestCases(params),
+    {
+      onSuccess: (data) =>
+        set({
+          testCaseList: data.data,
+          testCaseTotal: data.total,
+          isLoadingTestCases: false,
+        }),
+    },
+  );
+```
+
+### Pattern 2: Dependent Fetching
+
+Both hooks run in parallel — SWR dedupes, no manual sequencing needed.
+
+```tsx
+const BenchmarkDetail = () => {
+  const { benchmarkId } = useParams();
+  const useFetchBenchmarkDetail = useEvalStore((s) => s.useFetchBenchmarkDetail);
+  const useFetchDatasets = useEvalStore((s) => s.useFetchDatasets);
+
+  useFetchBenchmarkDetail(benchmarkId);
+  useFetchDatasets(benchmarkId);
+
+  return <div>...</div>;
+};
+```
+
+### Pattern 3: Conditional Fetching
+
+Pass `undefined` to disable the hook entirely.
+
+```tsx
+// only fetch when modal is open AND id present
+useFetchDatasetDetail(open && datasetId ? datasetId : undefined);
+```
+
+### Pattern 4: Cross-domain Refresh
+
+```typescript
+deleteBenchmark: async (id) => {
+  await agentEvalService.deleteBenchmark(id);
+  await get().refreshBenchmarks();
+  await get().refreshDatasets(id); // related cache invalidated too
+};
+```
+
+---
+
+## Migration Guide: useEffect → Store SWR
+
+### Before (❌ Wrong)
+
+```tsx
+const TestCaseList = ({ datasetId }: Props) => {
+  const [data, setData] = useState<any[]>([]);
+  const [loading, setLoading] = useState(false);
+
+  useEffect(() => {
+    setLoading(true);
+    lambdaClient.agentEval.listTestCases
+      .query({ datasetId })
+      .then((r) => setData(r.data))
+      .finally(() => setLoading(false));
+  }, [datasetId]);
+
+  return <Table data={data} loading={loading} />;
+};
+```
+
+### After (✅ Correct)
+
+```typescript
+// 1. Add service method
+class AgentEvalService {
+  async listTestCases(params: { datasetId: string }) {
+    return lambdaClient.agentEval.listTestCases.query(params);
+  }
+}
+
+// 2. Add store slice hook
+export const createTestCaseSlice: StateCreator<...> = (set) => ({
+  useFetchTestCases: (params) =>
+    useClientDataSWR(
+      params.datasetId ? [FETCH_TEST_CASES_KEY, params.datasetId] : null,
+      () => agentEvalService.listTestCases(params),
+      {
+        onSuccess: (data) =>
+          set({ testCaseList: data.data, isLoadingTestCases: false }),
+      },
+    ),
+});
+
+// 3. Component reads from store
+const TestCaseList = ({ datasetId }: Props) => {
+  const useFetchTestCases = useEvalStore((s) => s.useFetchTestCases);
+  const data = useEvalStore((s) => s.testCaseList);
+  const loading = useEvalStore((s) => s.isLoadingTestCases);
+
+  useFetchTestCases({ datasetId });
+
+  return <Table data={data} loading={loading} />;
+};
+```
+
+---
+
+## Troubleshooting
+
+| Symptom                     | Check                                                               |
+| --------------------------- | ------------------------------------------------------------------- |
+| Data never loads            | Hook called? Key not `null`/`undefined`? Network tab shows request? |
+| Stale data after mutation   | Did `refreshXxx` run? Cache key matches what the hook uses?         |
+| Loading state stuck `true`  | `onSuccess` writes loading=false? Promise rejected silently?        |
+| Detail map missing an entry | Reducer dispatch ran? `isEqual` short-circuited on stale data?      |
+
+---
+
+## Summary Checklist
+
+When adding new data fetching:
+
+### Step 1: Types & State
+
+See `store-data-structures` for details.
+
+- [ ] Define types in `@lobechat/types`: Detail type + List item type
+- [ ] State structure: `xxxList: XxxListItem[]`, `xxxDetailMap: Record<string, Xxx>`, `loadingXxxDetailIds: string[]`
+- [ ] Reducer if optimistic updates are needed
+
+### Step 2: Service Layer
+
+- [ ] Create service in `src/services/xxxService.ts`
+- [ ] Methods: `listXxx()`, `getXxx(id)`, `createXxx()`, `updateXxx()`, `deleteXxx()`
+
+### Step 3: Store Actions
+
+- [ ] `initialState.ts` with state structure
+- [ ] `action.ts` with:
+  - [ ] `useFetchXxxList()`, `useFetchXxxDetail(id)` — SWR hooks
+  - [ ] `refreshXxxList()`, `refreshXxxDetail(id)` — cache invalidation
+  - [ ] CRUD methods calling service
+  - [ ] `internal_dispatch`, `internal_updateLoading` if using reducer
+- [ ] `selectors.ts` (optional but recommended)
+- [ ] Integrate slice into main store + initialState
+
+### Step 4: Component Usage
+
+- [ ] Use store hooks (NOT useEffect)
+- [ ] List pages: access `xxxList` array
+- [ ] Detail pages: access `xxxDetailMap[id]`
+- [ ] Use loading states for UI feedback
+
+**Mental model:** Types → Service → Reducer → Slice → Component 🎯
+
+---
+
+## Related Skills
+
+- **`store-data-structures`** — How to structure List and Detail data in stores
+- **`zustand`** — General Zustand patterns and best practices

.agents/skills/data-fetching-architecture/references/walkthrough.md

@@ -0,0 +1,244 @@
+# Walkthrough: Adding a New Feature End-to-End
+
+This is a worked example of the canonical 6-step recipe applied to a new entity (`Dataset`), showing a variant of the main skill's pattern: **a list keyed by a parent id** (`datasetMap[benchmarkId]`), useful when the same shape appears under different parents.
+
+If you only need the canonical (single-array) pattern, the main `SKILL.md` already shows it for `Benchmark`. Read this file when you need the parent-keyed Map variant, or when you want a checklist-style walkthrough.
+
+## Step 1: Add Service methods
+
+```typescript
+class AgentEvalService {
+  async listDatasets(benchmarkId: string) {
+    return lambdaClient.agentEval.listDatasets.query({ benchmarkId });
+  }
+  async getDataset(id: string) {
+    return lambdaClient.agentEval.getDataset.query({ id });
+  }
+  async createDataset(params: CreateDatasetParams) {
+    return lambdaClient.agentEval.createDataset.mutate(params);
+  }
+  // updateDataset / deleteDataset follow the same shape
+}
+```
+
+## Step 2: Reducer (optimistic updates)
+
+```typescript
+// src/store/eval/slices/dataset/reducer.ts
+export type DatasetDispatch =
+  | { type: 'addDataset'; value: Dataset }
+  | { type: 'updateDataset'; id: string; value: Partial<Dataset> }
+  | { type: 'deleteDataset'; id: string };
+
+export const datasetReducer = (state: Dataset[] = [], payload: DatasetDispatch): Dataset[] =>
+  produce(state, (draft) => {
+    switch (payload.type) {
+      case 'addDataset':
+        draft.unshift(payload.value);
+        break;
+      case 'updateDataset': {
+        const i = draft.findIndex((item) => item.id === payload.id);
+        if (i !== -1) draft[i] = { ...draft[i], ...payload.value };
+        break;
+      }
+      case 'deleteDataset': {
+        const i = draft.findIndex((item) => item.id === payload.id);
+        if (i !== -1) draft.splice(i, 1);
+        break;
+      }
+    }
+  });
+```
+
+## Step 3: Store slice
+
+```typescript
+// src/store/eval/slices/dataset/initialState.ts
+export interface DatasetData {
+  currentPage: number;
+  hasMore: boolean;
+  isLoading: boolean;
+  items: Dataset[];
+  pageSize: number;
+  total: number;
+}
+
+export interface DatasetSliceState {
+  // Map keyed by benchmarkId — multiple parent contexts share the slice
+  datasetMap: Record<string, DatasetData>;
+  // Single item for modal display
+  datasetDetail: Dataset | null;
+  isLoadingDatasetDetail: boolean;
+  loadingDatasetIds: string[];
+}
+
+export const datasetInitialState: DatasetSliceState = {
+  datasetMap: {},
+  datasetDetail: null,
+  isLoadingDatasetDetail: false,
+  loadingDatasetIds: [],
+};
+```
+
+```typescript
+// src/store/eval/slices/dataset/action.ts
+const FETCH_DATASETS_KEY = 'FETCH_DATASETS';
+const FETCH_DATASET_DETAIL_KEY = 'FETCH_DATASET_DETAIL';
+
+export const createDatasetSlice: StateCreator<EvalStore, any, [], DatasetAction> = (set, get) => ({
+  // Cache key includes benchmarkId so each parent has its own SWR entry
+  useFetchDatasets: (benchmarkId) =>
+    useClientDataSWR(
+      benchmarkId ? [FETCH_DATASETS_KEY, benchmarkId] : null,
+      () => agentEvalService.listDatasets(benchmarkId!),
+      {
+        onSuccess: (data) => {
+          set({
+            datasetMap: {
+              ...get().datasetMap,
+              [benchmarkId!]: {
+                currentPage: 1,
+                hasMore: false,
+                isLoading: false,
+                items: data,
+                pageSize: data.length,
+                total: data.length,
+              },
+            },
+          });
+        },
+      },
+    ),
+
+  useFetchDatasetDetail: (id) =>
+    useClientDataSWR(
+      id ? [FETCH_DATASET_DETAIL_KEY, id] : null,
+      () => agentEvalService.getDataset(id!),
+      {
+        onSuccess: (data) => set({ datasetDetail: data, isLoadingDatasetDetail: false }),
+      },
+    ),
+
+  refreshDatasets: (benchmarkId) => mutate([FETCH_DATASETS_KEY, benchmarkId]),
+  refreshDatasetDetail: (id) => mutate([FETCH_DATASET_DETAIL_KEY, id]),
+
+  // CREATE with optimistic update — note the temp id pattern
+  createDataset: async (params) => {
+    const tmpId = Date.now().toString();
+    const { benchmarkId } = params;
+
+    get().internal_dispatchDataset(
+      { type: 'addDataset', value: { ...params, id: tmpId, createdAt: Date.now() } as any },
+      benchmarkId,
+    );
+    get().internal_updateDatasetLoading(tmpId, true);
+
+    try {
+      const result = await agentEvalService.createDataset(params);
+      await get().refreshDatasets(benchmarkId);
+      return result;
+    } finally {
+      get().internal_updateDatasetLoading(tmpId, false);
+    }
+  },
+
+  // UPDATE / DELETE follow the same optimistic + refresh pattern as BenchmarkSlice
+  // (see the main SKILL.md)
+
+  // Internal — dispatch reducer scoped to a parent
+  internal_dispatchDataset: (payload, benchmarkId) => {
+    const currentData = get().datasetMap[benchmarkId];
+    const nextItems = datasetReducer(currentData?.items, payload);
+
+    // Skip set when nothing changed — avoids unnecessary re-renders
+    if (isEqual(nextItems, currentData?.items)) return;
+
+    set({
+      datasetMap: {
+        ...get().datasetMap,
+        [benchmarkId]: {
+          ...currentData,
+          currentPage: currentData?.currentPage ?? 1,
+          hasMore: currentData?.hasMore ?? false,
+          isLoading: false,
+          items: nextItems,
+          pageSize: currentData?.pageSize ?? nextItems.length,
+          total: currentData?.total ?? nextItems.length,
+        },
+      },
+    });
+  },
+
+  internal_updateDatasetLoading: (id, loading) => {
+    set((state) => ({
+      loadingDatasetIds: loading
+        ? [...state.loadingDatasetIds, id]
+        : state.loadingDatasetIds.filter((i) => i !== id),
+    }));
+  },
+});
+```
+
+## Step 4: Wire into the store
+
+```typescript
+// src/store/eval/store.ts
+export type EvalStore = EvalStoreState & BenchmarkAction & DatasetAction & RunAction;
+
+const createStore: StateCreator<EvalStore, [['zustand/devtools', never]]> = (set, get, store) => ({
+  ...initialState,
+  ...createBenchmarkSlice(set, get, store),
+  ...createDatasetSlice(set, get, store),
+  ...createRunSlice(set, get, store),
+});
+
+// src/store/eval/initialState.ts
+export const initialState: EvalStoreState = {
+  ...benchmarkInitialState,
+  ...datasetInitialState,
+  ...runInitialState,
+};
+```
+
+## Step 5: Selectors (optional but recommended)
+
+```typescript
+export const datasetSelectors = {
+  getDatasetData: (benchmarkId: string) => (s: EvalStore) => s.datasetMap[benchmarkId],
+  getDatasets: (benchmarkId: string) => (s: EvalStore) => s.datasetMap[benchmarkId]?.items ?? [],
+  isLoadingDataset: (id: string) => (s: EvalStore) => s.loadingDatasetIds.includes(id),
+};
+```
+
+## Step 6: Use in component
+
+```tsx
+// List scoped to a parent
+const DatasetList = ({ benchmarkId }: { benchmarkId: string }) => {
+  const useFetchDatasets = useEvalStore((s) => s.useFetchDatasets);
+  const datasets = useEvalStore(datasetSelectors.getDatasets(benchmarkId));
+  const datasetData = useEvalStore(datasetSelectors.getDatasetData(benchmarkId));
+
+  useFetchDatasets(benchmarkId);
+
+  if (datasetData?.isLoading) return <Loading />;
+  return (
+    <div>
+      <h2>Total: {datasetData?.total ?? 0}</h2>
+      <List data={datasets} />
+    </div>
+  );
+};
+
+// Single item for modal — conditional fetching pattern
+const DatasetImportModal = ({ open, datasetId }: Props) => {
+  const useFetchDatasetDetail = useEvalStore((s) => s.useFetchDatasetDetail);
+  const dataset = useEvalStore((s) => s.datasetDetail);
+  const isLoading = useEvalStore((s) => s.isLoadingDatasetDetail);
+
+  // Only fetch when modal is open AND id present
+  useFetchDatasetDetail(open && datasetId ? datasetId : undefined);
+
+  return <Modal open={open}>{isLoading ? <Loading /> : <div>{dataset?.name}</div>}</Modal>;
+};
+```

.agents/skills/db-migrations/SKILL.md

@@ -1,10 +1,71 @@
 ---
 name: db-migrations
-description: 'Use when generating or regenerating Drizzle migration files, changing database schema tables or columns, resolving migration sequence conflicts after rebase, reviewing migration SQL for idempotent patterns, or renaming migration files.'
+description: 'Use for Drizzle migrations: schema/table/column changes, migration generation or regeneration, sequence conflicts after rebase, idempotent SQL review, or migration renames.'
+user-invocable: false
 ---
 
 # Database Migrations Guide
 
+## Development-stage schema changes
+
+Schema changes churn during feature development. When the schema changes before the migration has shipped, do not hand-edit the existing migration SQL to chase the new schema shape. Delete the draft migration artifacts added by this branch (SQL file, matching snapshot, and matching journal entry), then run the generator again and re-apply the normal migration review steps below.
+
+For example, if this branch's draft migration is `0110_add_verify_tables_and_ai_infra_id`:
+
+```bash
+# 1. Delete the draft SQL and its snapshot
+rm packages/database/migrations/0110_add_verify_tables_and_ai_infra_id.sql
+rm packages/database/migrations/meta/0110_snapshot.json
+
+# 2. Remove the matching 0110 entry from the journal's "entries" array
+#    packages/database/migrations/meta/_journal.json
+
+# 3. Regenerate from the current schema
+bun run db:generate
+```
+
+This keeps the generated SQL, snapshot, and journal aligned with the actual schema. Manual SQL edits are reserved for review-time hardening such as idempotent clauses, custom extension SQL, and meaningful filename/tag updates.
+
+Before release, if a feature branch accumulated multiple development-only migrations, consolidate them into one migration when possible. Production does not need to replay every intermediate draft shape, and fewer migrations reduce deploy-time risk.
+
+For example, if this branch added `0110`, `0111`, and `0112`, delete all three drafts and regenerate a single migration:
+
+```bash
+# 1. Delete every draft SQL and snapshot this branch added
+rm packages/database/migrations/011{0,1,2}_*.sql
+rm packages/database/migrations/meta/011{0,1,2}_snapshot.json
+
+# 2. Remove the 0110/0111/0112 entries from the journal's "entries" array
+#    packages/database/migrations/meta/_journal.json
+
+# 3. Regenerate one migration covering the full schema delta
+bun run db:generate
+```
+
+Do not make a migration compatible with earlier development-only versions of the same branch. While the migration has not shipped, there is no production history to preserve. Fix local/dev databases directly with whatever SQL is simplest (drop the draft table, rename a column, delete draft rows), then regenerate the branch migration from the current schema.
+
+For example, if an earlier draft on this branch created `signup_attempt_id` and you have since renamed it to `user_signup_log_id`, do not add a compatibility `ALTER ... RENAME` to the migration. Just fix the dev DB directly (see the `access-pg` skill for the `bun -e` + `pg` pattern), then regenerate:
+
+```bash
+# Fix the dev DB to match the new schema (simplest SQL wins)
+set -a && source .env && set +a && bun -e '
+import pg from "pg";
+const client = new pg.Client({ connectionString: process.env.DATABASE_URL });
+await client.connect();
+await client.query("ALTER TABLE user_signup_logs DROP COLUMN signup_attempt_id");
+await client.end();
+'
+
+# Regenerate so the migration reflects only the final shape
+bun run db:generate
+```
+
+After a migration has reached production or the target default branch, treat it as immutable: add a follow-up migration instead of rewriting it.
+
+## Rebase conflicts
+
+When a rebase conflicts in migration files, keep the upstream/default-branch migrations and remove all migrations introduced by the current feature branch. Complete the rebase, then regenerate this branch's migration from the rebased schema. This avoids merging two independent snapshots or hand-splicing journal entries.
+
 ## Step 1: Generate Migrations
 
 ```bash

.agents/skills/debug-package/SKILL.md

@@ -0,0 +1,66 @@
+---
+name: debug-package
+description: 'LobeHub debug package and log namespace guide. Use when adding debug() logging, choosing lobe-* namespaces, troubleshooting DEBUG output, localStorage.debug, or log format specifiers.'
+user-invocable: false
+---
+
+# Debug Package Usage Guide
+
+## Basic Usage
+
+```typescript
+import debug from 'debug';
+
+// Format: lobe-[module]:[submodule]
+const log = debug('lobe-server:market');
+
+log('Simple message');
+log('With variable: %O', object);
+log('Formatted number: %d', number);
+```
+
+## Namespace Conventions
+
+- Desktop: `lobe-desktop:[module]`
+- Server: `lobe-server:[module]`
+- Client: `lobe-client:[module]`
+- Router: `lobe-[type]-router:[module]`
+
+## Format Specifiers
+
+- `%O` - Object expanded (recommended for complex objects)
+- `%o` - Object
+- `%s` - String
+- `%d` - Number
+
+## Enable Debug Output
+
+### Browser
+
+```javascript
+localStorage.debug = 'lobe-*';
+```
+
+### Node.js
+
+```bash
+DEBUG=lobe-* npm run dev
+DEBUG=lobe-* pnpm dev
+```
+
+### Electron
+
+```typescript
+process.env.DEBUG = 'lobe-*';
+```
+
+## Example
+
+```typescript
+// apps/server/src/routers/edge/market/index.ts
+import debug from 'debug';
+
+const log = debug('lobe-edge-router:market');
+
+log('getAgent input: %O', input);
+```

.agents/skills/debug/SKILL.md

@@ -1,66 +0,0 @@
----
-name: debug
-description: Debug package usage guide. Use when adding debug logging, understanding log namespaces, or implementing debugging features. Triggers on debug logging requests or logging implementation.
-user-invocable: false
----
-
-# Debug Package Usage Guide
-
-## Basic Usage
-
-```typescript
-import debug from 'debug';
-
-// Format: lobe-[module]:[submodule]
-const log = debug('lobe-server:market');
-
-log('Simple message');
-log('With variable: %O', object);
-log('Formatted number: %d', number);
-```
-
-## Namespace Conventions
-
-- Desktop: `lobe-desktop:[module]`
-- Server: `lobe-server:[module]`
-- Client: `lobe-client:[module]`
-- Router: `lobe-[type]-router:[module]`
-
-## Format Specifiers
-
-- `%O` - Object expanded (recommended for complex objects)
-- `%o` - Object
-- `%s` - String
-- `%d` - Number
-
-## Enable Debug Output
-
-### Browser
-
-```javascript
-localStorage.debug = 'lobe-*';
-```
-
-### Node.js
-
-```bash
-DEBUG=lobe-* npm run dev
-DEBUG=lobe-* pnpm dev
-```
-
-### Electron
-
-```typescript
-process.env.DEBUG = 'lobe-*';
-```
-
-## Example
-
-```typescript
-// src/server/routers/edge/market/index.ts
-import debug from 'debug';
-
-const log = debug('lobe-edge-router:market');
-
-log('getAgent input: %O', input);
-```

.agents/skills/desktop/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: desktop
-description: Electron desktop development guide. Use when implementing desktop features, IPC handlers, controllers, preload scripts, window management, menu configuration, or Electron-specific functionality. Triggers on desktop app development, Electron IPC, or desktop local tools implementation.
+description: Electron desktop development guide — IPC handlers, controllers, preload scripts, window/menu management.
 disable-model-invocation: true
 ---
 

.agents/skills/docs-changelog/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: docs-changelog
-description: 'Writing guide for website changelog pages under docs/changelog/*.mdx. Use when creating or editing product update posts in EN/ZH. Not for GitHub Release notes.'
+description: 'Write website changelog pages under docs/changelog/*.mdx. Use for EN/ZH product update posts, changelog posts, update-log copy, or docs changelog edits; not GitHub Release notes.'
 ---
 
 # Docs Changelog Writing Guide

.agents/skills/drizzle/SKILL.md

@@ -1,20 +1,29 @@
 ---
 name: drizzle
-description: Drizzle ORM schema and database guide. Use when working with database schemas (src/database/schemas/*), defining tables, creating migrations, or database model code. Triggers on Drizzle schema definition, database migrations, or ORM usage questions.
+description: 'LobeHub Drizzle ORM schema and query style. Use for pgTable schemas, indexes, joins, inferred types, db.select/db.query, schema fields, foreign keys, junction tables, or postgres query patterns.'
+user-invocable: false
 ---
 
 # Drizzle ORM Schema Style Guide
 
+> **Adding a Model or Repository?** Ship a sibling test in the same PR — every new
+> file under `packages/database/src/models/**` or `src/repositories/**` needs a
+> matching `__tests__/<name>.test.ts`. See the **testing** skill
+> (`.agents/skills/testing/references/db-model-test.md`) for the `getTestDB()`
+> integration pattern, user-isolation tests, the BM25 `describe.skipIf(!isServerDB)`
+> guard, and schema gotchas. CI's coverage patch gate won't reliably catch a brand-new
+> untested file, so this is on you.
+
 ## Configuration
 
 - Config: `drizzle.config.ts`
-- Schemas: `src/database/schemas/`
-- Migrations: `src/database/migrations/`
+- Schemas: `packages/database/src/schemas/`
+- Migrations: `packages/database/migrations/`
 - Dialect: `postgresql` with `strict: true`
 
 ## Helper Functions
 
-Location: `src/database/schemas/_helpers.ts`
+Location: `packages/database/src/schemas/_helpers.ts`
 
 - `timestamptz(name)`: Timestamp with timezone
 - `createdAt()`, `updatedAt()`, `accessedAt()`: Standard timestamp columns
@@ -24,16 +33,42 @@ Location: `src/database/schemas/_helpers.ts`
 
 - **Tables**: Plural snake_case (`users`, `session_groups`)
 - **Columns**: snake_case (`user_id`, `created_at`)
+- **New tables**: Check nearby existing tables before naming a new one. Preserve
+  the established noun family and suffix. For example, if the user-scoped table
+  is `user_xxx_logs`, the workspace-scoped counterpart should be
+  `workspace_xxx_logs`, not `workspace_xxx_records` or another new synonym.
+
+```typescript
+// ✅ Good: follows the existing user/workspace table family.
+export const userSignupLogs = pgTable('user_signup_logs', { ... });
+export const workspaceSignupLogs = pgTable('workspace_signup_logs', { ... });
+
+// ❌ Bad: introduces a new suffix for the same concept.
+export const workspaceSignupRecords = pgTable('workspace_signup_records', { ... });
+```
 
 ## Column Definitions
 
 ### Primary Keys
 
+Do not use auto-incrementing primary keys (`serial`, `bigserial`, generated
+identity columns). They create sequence-state problems during cross-database
+migrations, restores, and data copy jobs. Prefer text IDs from application
+generators (`idGenerator`, `createNanoId`) or `uuid` for internal tables.
+
+Keep `$defaultFn(...)` when a table normally owns ID generation. Callers can
+still pass an explicit `id`; the default only runs when the insert omits it. Do
+not remove the default just because one flow needs to supply a request-scoped ID.
+
 ```typescript
+// ✅ Good: app-generated text ID; explicit inserts can still override it.
 id: text('id')
   .primaryKey()
   .$defaultFn(() => idGenerator('agents'))
   .notNull(),
+
+// ❌ Bad: sequence state is fragile across DB migrations and restores.
+id: serial('id').primaryKey(),
 ```
 
 ID prefixes make entity types distinguishable. For internal tables, use `uuid`.
@@ -52,6 +87,80 @@ userId: text('user_id')
 ...timestamps,  // Spread from _helpers.ts
 ```
 
+### Optional and Undefined Values
+
+Do not introduce artificial sentinel strings for missing values, such as
+`unknown`, unless the domain already has that explicit state and existing code
+uses it consistently. Prefer nullable columns, optional TypeScript fields, or a
+separate concrete status enum when the value is genuinely absent.
+
+```typescript
+// ✅ Good: absent until the final stage writes a real decision.
+export type UserSignupLogFinalDecision = 'allow' | 'block' | 'error';
+
+finalDecision: varchar('final_decision', { length: 32 }).$type<UserSignupLogFinalDecision>(),
+
+// ❌ Bad: invents a new state that callers now need to handle everywhere.
+export type UserSignupLogFinalDecision = 'allow' | 'block' | 'error' | 'unknown';
+
+finalDecision: varchar('final_decision', { length: 32 })
+  .$type<UserSignupLogFinalDecision>()
+  .notNull()
+  .default('unknown');
+```
+
+### Field Descriptions
+
+For columns whose meaning is not obvious from the name alone, add JSDoc on the
+schema field. Include a concrete example when it clarifies the stored value or
+the lifecycle moment that writes it. This is especially important for external
+IDs, lifecycle statuses, denormalized snapshots, JSONB signals, and fields whose
+name could mean either a request ID or a persisted row ID.
+
+```typescript
+// ✅ Good: explain the table's business object first, then only document
+// non-obvious lifecycle or risk-control fields.
+/**
+ * User signup logs - one row per signup flow, collecting stage-level
+ * risk-control decisions before and after the auth provider creates a user.
+ */
+export const userSignupLogs = pgTable('user_signup_logs', {
+  /** Final signup outcome reason, for example user_created, llm_block, or guard_error */
+  finalReason: text('final_reason'),
+
+  /** Aggregated risk level derived from stage decisions, for example block -> high */
+  riskLevel: varchar('risk_level', { length: 16 }).$type<UserSignupLogRiskLevel>(),
+
+  /** Ordered stage-level decisions and metadata grouped by signup review stage */
+  stageResults: jsonb('stage_results').$type<UserSignupLogStageResults>(),
+});
+
+// ❌ Bad: comments restate obvious column names without adding domain meaning.
+/** User email */
+email: text('email'),
+```
+
+### JSONB Types
+
+Avoid `Record<string, unknown>` or similarly loose JSONB types for schema
+columns. Define a concrete interface that describes the expected JSON shape, even
+when most properties are optional. This keeps callers, migrations, and review
+queries aligned on the same data contract.
+
+```typescript
+interface UserSignupLogMetadata {
+  payloadPath?: string;
+  requestPath?: string;
+}
+
+metadata: jsonb('metadata').$type<UserSignupLogMetadata>(),
+```
+
+```typescript
+// ❌ Bad: hides the contract and makes downstream access untyped.
+metadata: jsonb('metadata').$type<Record<string, unknown>>(),
+```
+
 ### Indexes
 
 ```typescript
@@ -125,11 +234,7 @@ The relational API generates complex lateral joins with `json_build_array` that
 
 ```typescript
 // ✅ Good
-const [result] = await this.db
-  .select()
-  .from(agents)
-  .where(eq(agents.id, id))
-  .limit(1);
+const [result] = await this.db.select().from(agents).where(eq(agents.id, id)).limit(1);
 return result;
 
 // ❌ Bad: relational API
@@ -177,6 +282,78 @@ const rows = await this.db
   .groupBy(agentEvalDatasets.id);
 ```
 
+### Raw SQL and Advanced Queries
+
+Prefer Drizzle builders whenever the query reads clearly with `select`,
+`insert().select()`, `update().from()`, joins, CTEs, and `groupBy` — this keeps
+table/column references tied to schema, so changes surface as TypeScript errors.
+Within a builder, expression-level `sql<T>` is fine for features lacking a helper
+(JSON path, casts, aggregates, `CASE`, `NOW()`). Row locks are clauses, not
+expressions — use `.for('update')`, never raw `FOR UPDATE`.
+
+Use `COALESCE` only when null-handling is part of required DB semantics (nullable
+JSONB append/merge, "keep first non-null"). Don't scatter
+`COALESCE(excluded.col, current.col)` across ordinary upsert scalars just to avoid
+an update object — build `set` from defined values only, and hide any remaining
+SQL behind named helpers (`appendJsonbArray`, `mergeJsonbObject`, `keepFirstValue`)
+so the method reads as business intent, not SQL plumbing.
+
+```typescript
+// ✅ Scalars included only when present; SQL hidden behind a named helper.
+const updateValues = compactUndefined({
+  email: record.email ?? undefined,
+  ip: record.ip ?? undefined,
+});
+await db.insert(userSignupLogs).values(values).onConflictDoUpdate({
+  set: { ...updateValues, stageResults: appendStageResult(stage, result), updatedAt: now },
+  target: userSignupLogs.id,
+});
+
+// ❌ Every scalar becomes SQL plumbing.
+set: {
+  email: sql`COALESCE(excluded.email, ${userSignupLogs.email})`,
+  ip: sql`COALESCE(excluded.ip, ${userSignupLogs.ip})`,
+}
+```
+
+When refactoring raw SQL:
+
+- Preserve query shape on latency-sensitive paths. If raw SQL is one roundtrip,
+  don't split it into multiple depth-based queries just to drop `execute`.
+- Use `$with(...)` + `insert().select()` / `update().from()` for multi-step
+  single-roundtrip writes Drizzle can express.
+- Don't rely on `execute<MyRow>(sql...)` for safety — it types rows but doesn't keep
+  selected columns in sync with schema changes.
+- If only a PostgreSQL feature Drizzle can't express works, keep the raw SQL and
+  tighten it: schema refs in interpolations, explicit user scope, a narrow row
+  interface, and regression tests.
+
+Recursive CTEs are the canonical "keep raw" case — there's no clean `WITH RECURSIVE`
+builder, and a rewrite would add depth-based roundtrips:
+
+```typescript
+interface TaskTreeRow {
+  id: string;
+  parent_task_id: string | null;
+}
+
+// execute<T> acceptable: no clean WITH RECURSIVE builder. Keep schema refs in the
+// interpolations and scope every leg to the user.
+const { rows } = await db.execute<TaskTreeRow>(sql`
+  WITH RECURSIVE task_tree AS (
+    SELECT ${tasks.id}, ${tasks.parentTaskId}
+    FROM ${tasks}
+    WHERE ${tasks.id} = ${rootTaskId} AND ${tasks.createdByUserId} = ${userId}
+    UNION ALL
+    SELECT ${tasks.id}, ${tasks.parentTaskId}
+    FROM ${tasks}
+    JOIN task_tree ON ${tasks.parentTaskId} = task_tree.id
+    WHERE ${tasks.createdByUserId} = ${userId}
+  )
+  SELECT * FROM task_tree
+`);
+```
+
 ### One-to-Many (Separate Queries)
 
 When you need a parent record with its children, use two queries instead of relational `with:`:

.agents/skills/heterogeneous-agent/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: heterogeneous-agent
-description: Guide for implementing and debugging LobeHub heterogeneous agent integrations such as Claude Code, Codex, and future external CLI agents. Use when working on adapter event mapping, Electron IPC transport, renderer persistence, tool-call chaining, subagent threads, resume/session handling, or regressions like mixed multi-tool messages, broken step boundaries, stuck tool loading, and orphan tool messages. Triggers on 'heterogeneous agent', 'hetero agent', '异构 agent', 'claude code adapter', 'codex adapter', 'external agent CLI', '孤立 tool 消息', 'raw Codex trace', or adapter/executor bugs.
+description: 'Implement or debug LobeHub heterogeneous agents. Use for Claude Code/Codex adapters, external CLI agents, event mapping, IPC, persistence, tool-call chains, sessions, traces, or adapter bugs.'
 ---
 
 # Heterogeneous Agent Development

.agents/skills/heterogeneous-agent/references/debug-workflow.md

@@ -241,6 +241,6 @@ When the bug comes from a real trace, distill it into the closest existing test
 3. Add or update the narrowest failing test near the broken layer.
 4. Fix the smallest layer that can explain the symptom.
 5. Re-run focused tests.
-6. Only then do an Electron smoke test with the `local-testing` skill if UI confirmation is still needed.
+6. Only then do an Electron smoke test with the `agent-testing` skill if UI confirmation is still needed.
 
 Do not start with a broad Electron repro if a raw trace or adapter test can prove the fault zone faster.

.agents/skills/hotkey/SKILL.md

@@ -1,6 +1,7 @@
 ---
 name: hotkey
-description: Guide for adding keyboard shortcuts. Use when implementing new hotkeys, registering shortcuts, or working with keyboard interactions. Triggers on hotkey implementation or keyboard shortcut tasks.
+description: 'Add or edit LobeHub keyboard shortcuts. Use for HotkeyEnum, HOTKEYS_REGISTRATION, combineKeys, useHotkeyById, tooltip hotkeys, shortcut scope, conflicts, or Cmd/Ctrl key combos.'
+user-invocable: false
 ---
 
 # Adding Keyboard Shortcuts Guide

.agents/skills/i18n/SKILL.md

@@ -1,6 +1,7 @@
 ---
 name: i18n
-description: Internationalization guide using react-i18next. Use when adding translations, creating i18n keys, or working with localized text in React components (.tsx files). Triggers on translation tasks, locale management, or i18n implementation.
+description: 'LobeHub i18n with react-i18next. Use for user-facing strings, locale keys, namespaces, useTranslation, t(), interpolation, zh-CN/en-US previews, hardcoded UI copy, or pnpm i18n.'
+user-invocable: false
 ---
 
 # LobeHub Internationalization Guide

.agents/skills/linear/SKILL.md

@@ -1,55 +1,55 @@
 ---
 name: linear
-description: "Linear issue management. MUST USE when: (1) user mentions LOBE-xxx issue IDs (e.g. LOBE-4540), (2) user says 'linear', 'linear issue', 'link linear', (3) creating PRs that reference Linear issues. Provides workflows for retrieving issues, updating status, and adding comments."
+description: 'Linear issue management. Use for LOBE-xxx issues, Linear links, PRs referencing Linear, retrieving issues, updating status, completion comments, or sub-issue trees.'
+user-invocable: false
 ---
 
 # Linear Issue Management
 
 Before using Linear workflows, search for `linear` MCP tools. If not found, treat as not installed.
 
-## ⚠️ CRITICAL: PR Creation with Linear Issues
+## PR Creation with Linear Issues
 
-**When creating a PR that references Linear issues (LOBE-xxx), you MUST:**
+A PR that fixes a Linear issue has **two separate jobs to do**, and both matter:
 
-1. Create the PR with magic keywords (`Fixes LOBE-xxx`)
-2. **IMMEDIATELY after PR creation**, add completion comments to ALL referenced Linear issues
-3. Do NOT consider the task complete until Linear comments are added
+1. **`Fixes LOBE-xxx` in the PR body** — Linear watches GitHub for these magic keywords and auto-links the PR and auto-closes the issue on merge. This is the machine-readable side.
+2. **A completion comment on the Linear issue** — gives the reviewer/PM/teammate landing in Linear a human-readable summary of what changed and why, without forcing them to click through to GitHub and read a diff.
 
-This is NON-NEGOTIABLE. Skipping Linear comments is a workflow violation.
+If you only do step 1, Linear watchers (often non-engineers) hit the issue and see no context. So pair PR creation with the Linear comment as part of the same task — finish both before considering the work done.
 
 ## Workflow
 
 1. **Retrieve issue details** before starting: `mcp__linear-server__get_issue`
-2. **Read images**: If the issue description contains images, MUST use `mcp__linear-server__extract_images` to read image content for full context
-3. **Check for sub-issues**: Use `mcp__linear-server__list_issues` with `parentId` filter
-4. **Mark as In Progress**: When starting to plan or implement an issue, immediately update status to **"In Progress"** via `mcp__linear-server__update_issue`
+2. **Read images** — issue descriptions often contain screenshots with critical context (mockups, error states, before/after). Use `mcp__linear-server__extract_images` so you actually see them; reading raw markdown alone misses what the reporter was looking at.
+3. **Check for sub-issues**: `mcp__linear-server__list_issues` with `parentId` filter
+4. **Mark as In Progress** at the moment you start planning or implementing — this signals to teammates the issue is owned, so they don't double-pick it up.
 5. **Update issue status** when completing: `mcp__linear-server__update_issue`
-6. **Add completion comment** (REQUIRED): `mcp__linear-server__create_comment`
+6. **Add completion comment** (see [format below](#completion-comment-format))
 
 ## Creating Issues
 
-When creating issues with `mcp__linear-server__create_issue`, **MUST add the `claude code` label**.
+When creating issues with `mcp__linear-server__create_issue`, add the `claude code` label. Reason: the label is how the team filters/audits AI-generated issues; without it those issues vanish into the general backlog and the team loses visibility into AI contribution patterns.
 
 ## Language
 
-Issue titles, descriptions, and comments **MUST follow the language of the current conversation**, not default to English.
+Match the issue language to the conversation that produced it — if you're discussing in 中文，write the issue in 中文；if discussing in English, write it in English. Reason: the issue is a continuation of the conversation, and forcing a language switch creates translation friction for the collaborator who started the thread.
 
-- Conversation in 中文 → issue body in 中文；technical terms (file paths, identifiers, library names, commands, error messages) stay in English.
-- Conversation in English → issue body in English.
+Specifics:
+
+- 中文 conversation → 中文 body; technical terms (file paths, identifiers, library names, commands, error messages) stay in English.
+- English conversation → English body.
 - Code blocks, file paths, and quoted strings always stay in their original form regardless of surrounding language.
-- This applies equally to **updates** — when editing an existing issue (description **and titles**), preserve the language of the conversation that triggered the edit; do not switch the issue language during a refactor (Chinese → English or vice versa).
-
-Rationale: the issue is a continuation of the conversation. Forcing English when the discussion is in Chinese creates translation friction for the collaborator who came from that thread.
+- This applies equally to **updates** — when editing an existing issue (description **and titles**), preserve the language of the conversation that triggered the edit; don't switch the issue language mid-refactor.
 
 ## Creating Sub-issue Trees
 
 When breaking a parent issue into a tree of sub-issues (e.g., task decomposition for LOBE-xxx), follow these rules — they work around real limitations of the Linear MCP tools.
 
-### 1. ALWAYS prefix titles with an ordering index
+### 1. Prefix titles with an ordering index
 
-The Linear Sub-issues panel displays children by `sortOrder`, which **defaults to newest-first** (most recently created appears on top). Neither parallel nor serial creation will produce the intended top-to-bottom reading order, and the MCP `save_issue` tool does **not expose a `sortOrder` parameter** — you cannot set order at create time.
+The Linear Sub-issues panel orders children by `sortOrder`, which **defaults to newest-first** (most recently created appears on top). Neither parallel nor serial creation produces the intended top-to-bottom reading order, and the MCP `save_issue` tool does **not expose a `sortOrder` parameter** — you can't set order at create time.
 
-**Workaround**: encode execution order in the title itself:
+Workaround: encode execution order in the title itself:
 
 ```plaintext
 [1]     [db]       add schema fields
@@ -100,7 +100,7 @@ The implementer may open only the sub-issue, not the parent — don't rely on co
 
 ## Completion Comment Format
 
-Every completed issue MUST have a comment summarizing work done:
+Each completed issue gets a comment summarizing the work, so reviewers and future readers don't have to reconstruct it from the PR diff:
 
 ```markdown
 ## Changes Summary
@@ -116,34 +116,28 @@ Every completed issue MUST have a comment summarizing work done:
 - ...
 ```
 
-This is critical for:
+This gives team visibility, code-review context, and a paper trail for future reference.
 
-- Team visibility
-- Code review context
-- Future reference
+## PR Association
 
-## PR Association (REQUIRED)
-
-When creating PRs for Linear issues, include magic keywords in PR body:
+When creating PRs for Linear issues, include magic keywords in the PR body:
 
 - `Fixes LOBE-123`
 - `Closes LOBE-123`
 - `Resolves LOBE-123`
 
+These trigger Linear's auto-link + auto-close on merge.
+
 ## Per-Issue Completion Rule
 
-When working on multiple issues, update EACH issue IMMEDIATELY after completing it:
+When working on multiple issues, close out **each one before starting the next** — don't batch all the Linear updates to the end. Batching is where comments get forgotten and issues stay stuck in "In Progress" days after the PR shipped.
+
+For each issue:
 
 1. Complete implementation
 2. Run `bun run type-check`
 3. Run related tests
 4. Create PR if needed
-5. Update status to **"In Review"** (NOT "Done")
-6. **Add completion comment immediately**
-7. Move to next issue
-
-**Note:** Status → "In Review" when PR created. "Done" only after PR merged.
-
-**❌ Wrong:** Complete all → Create PR → Forget Linear comments
-
-**✅ Correct:** Complete → Create PR → Add Linear comments → Task done
+5. Update status to **"In Review"** (not "Done" — "Done" is for after the PR merges)
+6. Add the completion comment
+7. Move to the next issue

.agents/skills/local-testing/SKILL.md

@@ -1,520 +0,0 @@
----
-name: local-testing
-description: >
-  Local app and bot testing. Uses agent-browser CLI for Electron/web app UI testing,
-  and osascript (AppleScript) for controlling native macOS apps (WeChat, Discord, Telegram, Slack, Lark/飞书, QQ)
-  to test bots. Triggers on 'local test', 'test in electron', 'test desktop', 'test bot',
-  'bot test', 'test in discord', 'test in telegram', 'test in slack', 'test in weixin',
-  'test in wechat', 'test in lark', 'test in feishu', 'test in qq',
-  'manual test', 'osascript', or UI/bot verification tasks.
----
-
-# Local App & Bot Testing
-
-Two approaches for local testing on macOS:
-
-| Approach                    | Tool                | Best For                                             |
-| --------------------------- | ------------------- | ---------------------------------------------------- |
-| **agent-browser + CDP**     | `agent-browser` CLI | Electron apps, web apps (DOM access, JS eval)        |
-| **osascript (AppleScript)** | `osascript -e`      | Native macOS apps (WeChat, Discord, Telegram, Slack) |
-
----
-
-# Part 1: agent-browser (Electron / Web Apps)
-
-Use `agent-browser` to automate Chromium-based apps via Chrome DevTools Protocol.
-
-Install via `npm i -g agent-browser`, `brew install agent-browser`, or `cargo install agent-browser`. Run `agent-browser install` to download Chrome. Run `agent-browser upgrade` to update.
-
-## Core Workflow
-
-Every browser automation follows this pattern:
-
-1. **Navigate**: `agent-browser open <url>`
-2. **Snapshot**: `agent-browser snapshot -i` (get element refs like `@e1`, `@e2`)
-3. **Interact**: Use refs to click, fill, select
-4. **Re-snapshot**: After navigation or DOM changes, get fresh refs
-
-```bash
-agent-browser open https://example.com/form
-agent-browser snapshot -i
-# Output: @e1 [input type="email"], @e2 [input type="password"], @e3 [button] "Submit"
-
-agent-browser fill @e1 "user@example.com"
-agent-browser fill @e2 "password123"
-agent-browser click @e3
-agent-browser wait --load networkidle
-agent-browser snapshot -i # Check result
-```
-
-## Command Chaining
-
-```bash
-# Chain open + wait + snapshot in one call
-agent-browser open https://example.com && agent-browser wait --load networkidle && agent-browser snapshot -i
-```
-
-Use `&&` when you don't need to read intermediate output. Run commands separately when you need to parse output first (e.g., snapshot to discover refs, then interact).
-
-## Essential Commands
-
-```bash
-# Navigation
-agent-browser open <url>              # Navigate (aliases: goto, navigate)
-agent-browser close                   # Close browser
-agent-browser close --all             # Close all active sessions
-
-# Snapshot
-agent-browser snapshot -i             # Interactive elements with refs (recommended)
-agent-browser snapshot -s "#selector" # Scope to CSS selector
-
-# Interaction (use @refs from snapshot)
-agent-browser click @e1               # Click element
-agent-browser click @e1 --new-tab     # Click and open in new tab
-agent-browser fill @e2 "text"         # Clear and type text
-agent-browser type @e2 "text"         # Type without clearing
-agent-browser select @e1 "option"     # Select dropdown option
-agent-browser check @e1               # Check checkbox
-agent-browser press Enter             # Press key
-agent-browser keyboard type "text"    # Type at current focus (no selector)
-agent-browser keyboard inserttext "text"  # Insert without key events
-agent-browser scroll down 500         # Scroll page
-agent-browser scroll down 500 --selector "div.content"  # Scroll within container
-
-# Get information
-agent-browser get text @e1            # Get element text
-agent-browser get url                 # Get current URL
-agent-browser get title               # Get page title
-agent-browser get cdp-url             # Get CDP WebSocket URL
-
-# Wait
-agent-browser wait @e1                # Wait for element
-agent-browser wait --load networkidle # Wait for network idle
-agent-browser wait --url "**/page"    # Wait for URL pattern
-agent-browser wait 2000               # Wait milliseconds
-agent-browser wait --text "Welcome"   # Wait for text to appear
-agent-browser wait --fn "!document.body.innerText.includes('Loading...')"  # Wait for text to disappear
-agent-browser wait "#spinner" --state hidden  # Wait for element to disappear
-
-# Downloads
-agent-browser download @e1 ./file.pdf          # Click element to trigger download
-agent-browser wait --download ./output.zip     # Wait for any download to complete
-
-# Network
-agent-browser network requests                 # Inspect tracked requests
-agent-browser network requests --type xhr,fetch  # Filter by resource type
-agent-browser network requests --method POST   # Filter by HTTP method
-agent-browser network route "**/api/*" --abort # Block matching requests
-agent-browser network har start                # Start HAR recording
-agent-browser network har stop ./capture.har   # Stop and save HAR file
-
-# Viewport & Device Emulation
-agent-browser set viewport 1920 1080          # Set viewport size (default: 1280x720)
-agent-browser set viewport 1920 1080 2        # 2x retina
-agent-browser set device "iPhone 14"          # Emulate device (viewport + user agent)
-
-# Capture
-agent-browser screenshot              # Screenshot to temp dir
-agent-browser screenshot --full       # Full page screenshot
-agent-browser screenshot --annotate   # Annotated screenshot with numbered element labels
-agent-browser pdf output.pdf          # Save as PDF
-
-# Clipboard
-agent-browser clipboard read          # Read text from clipboard
-agent-browser clipboard write "text"  # Write text to clipboard
-agent-browser clipboard copy          # Copy current selection
-agent-browser clipboard paste         # Paste from clipboard
-
-# Dialogs (alert, confirm, prompt, beforeunload)
-agent-browser dialog accept           # Accept dialog
-agent-browser dialog accept "input"   # Accept prompt dialog with text
-agent-browser dialog dismiss          # Dismiss/cancel dialog
-agent-browser dialog status           # Check if dialog is open
-
-# Diff (compare page states)
-agent-browser diff snapshot                        # Compare current vs last snapshot
-agent-browser diff screenshot --baseline before.png  # Visual pixel diff
-agent-browser diff url <url1> <url2>               # Compare two pages
-
-# Streaming
-agent-browser stream enable           # Start WebSocket streaming
-agent-browser stream status           # Inspect streaming state
-agent-browser stream disable          # Stop streaming
-```
-
-## Batch Execution
-
-```bash
-echo '[
-  ["open", "https://example.com"],
-  ["snapshot", "-i"],
-  ["click", "@e1"],
-  ["screenshot", "result.png"]
-]' | agent-browser batch --json
-```
-
-## Authentication
-
-```bash
-# Option 1: Auth vault (credentials stored encrypted)
-echo "$PASSWORD" | agent-browser auth save myapp --url https://app.example.com/login --username user --password-stdin
-agent-browser auth login myapp
-
-# Option 2: Session name (auto-save/restore cookies + localStorage)
-agent-browser --session-name myapp open https://app.example.com/login
-agent-browser close                                                       # State auto-saved
-agent-browser --session-name myapp open https://app.example.com/dashboard # Auto-restored
-
-# Option 3: Persistent profile
-agent-browser --profile ~/.myapp open https://app.example.com/login
-
-# Option 4: State file
-agent-browser state save auth.json
-agent-browser state load auth.json
-```
-
-### LobeHub dev server — inject better-auth cookie
-
-`agent-browser --headed` on macOS can create an off-screen Chromium window, blocking manual login. For a local LobeHub dev server (e.g. `localhost:3011`), copy the `better-auth.session_token` cookie out of a **Network request** in the user's own Chrome DevTools and load it via `state load`. See [references/agent-browser-login.md](./references/agent-browser-login.md) for the full recipe.
-
-## Semantic Locators (Alternative to Refs)
-
-```bash
-agent-browser find text "Sign In" click
-agent-browser find label "Email" fill "user@test.com"
-agent-browser find role button click --name "Submit"
-agent-browser find placeholder "Search" type "query"
-agent-browser find testid "submit-btn" click
-```
-
-## JavaScript Evaluation (eval)
-
-```bash
-# Simple expressions
-agent-browser eval 'document.title'
-
-# Complex JS: use --stdin with heredoc (RECOMMENDED)
-agent-browser eval --stdin << 'EVALEOF'
-JSON.stringify(
-  Array.from(document.querySelectorAll("img"))
-    .filter(i => !i.alt)
-    .map(i => ({ src: i.src.split("/").pop(), width: i.width }))
-)
-EVALEOF
-
-# Base64 encoding (avoids all shell escaping issues)
-agent-browser eval -b "$(echo -n 'document.title' | base64)"
-```
-
-## Ref Lifecycle
-
-Refs (`@e1`, `@e2`, etc.) are invalidated when the page changes. Always re-snapshot after clicking links/buttons that navigate, form submissions, or dynamic content loading.
-
-## Annotated Screenshots (Vision Mode)
-
-```bash
-agent-browser screenshot --annotate
-# Output includes the image path and a legend:
-#   [1] @e1 button "Submit"
-#   [2] @e2 link "Home"
-agent-browser click @e2 # Click using ref from annotated screenshot
-```
-
-## Parallel Sessions
-
-```bash
-agent-browser --session site1 open https://site-a.com
-agent-browser --session site2 open https://site-b.com
-agent-browser session list
-```
-
-## Connect to Existing Chrome
-
-```bash
-agent-browser --auto-connect snapshot # Auto-discover running Chrome
-agent-browser --cdp 9222 snapshot     # Explicit CDP port
-```
-
-## iOS Simulator (Mobile Safari)
-
-```bash
-agent-browser device list
-agent-browser -p ios --device "iPhone 16 Pro" open https://example.com
-agent-browser -p ios snapshot -i
-agent-browser -p ios tap @e1
-agent-browser -p ios swipe up
-agent-browser -p ios screenshot mobile.png
-agent-browser -p ios close
-```
-
-## Observability Dashboard
-
-```bash
-agent-browser dashboard install
-agent-browser dashboard start # Background server on port 4848
-agent-browser dashboard stop
-```
-
-## Cloud Providers
-
-Use `-p <provider>` to run against cloud browsers: `agentcore`, `browserbase`, `browserless`, `browseruse`, `kernel`.
-
-## Browser Engine Selection
-
-```bash
-agent-browser --engine lightpanda open example.com # 10x faster, 10x less memory
-```
-
-## Electron (LobeHub Desktop)
-
-### Setup / Teardown
-
-Use the `electron-dev.sh` script to manage the Electron dev environment. It handles process lifecycle, waits for SPA readiness, and reliably kills all child processes (main + helpers + vite).
-
-```bash
-SCRIPT=".agents/skills/local-testing/scripts/electron-dev.sh"
-
-# Start Electron dev with CDP (idempotent — skips if already running)
-$SCRIPT start
-
-# Check if Electron is running and CDP is reachable
-$SCRIPT status
-
-# Kill all Electron-related processes (main + helper + vite)
-$SCRIPT stop
-
-# Force fresh restart
-$SCRIPT restart
-```
-
-After `start` succeeds, connect with: `agent-browser --cdp 9222 snapshot -i`
-
-**Always run `$SCRIPT stop` when done testing** — `pkill -f "Electron"` alone won't catch all helper processes.
-
-#### Environment Variables
-
-| Variable          | Default                 | Description                              |
-| ----------------- | ----------------------- | ---------------------------------------- |
-| `CDP_PORT`        | `9222`                  | Chrome DevTools Protocol port            |
-| `ELECTRON_LOG`    | `/tmp/electron-dev.log` | Electron process log                     |
-| `ELECTRON_WAIT_S` | `60`                    | Max seconds to wait for Electron process |
-| `RENDERER_WAIT_S` | `60`                    | Max seconds to wait for SPA to load      |
-
-### LobeHub-Specific Patterns
-
-#### Access Zustand Store State
-
-```bash
-agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
-(function() {
-  var chat = window.__LOBE_STORES.chat();
-  var ops = Object.values(chat.operations);
-  return JSON.stringify({
-    ops: ops.map(function(o) { return { type: o.type, status: o.status }; }),
-    activeAgent: chat.activeAgentId,
-    activeTopic: chat.activeTopicId,
-  });
-})()
-EVALEOF
-```
-
-#### Find and Use the Chat Input
-
-```bash
-# The chat input is contenteditable — must use -C flag
-agent-browser --cdp 9222 snapshot -i -C 2>&1 | grep "editable"
-
-agent-browser --cdp 9222 click @e48
-agent-browser --cdp 9222 type @e48 "Hello world"
-agent-browser --cdp 9222 press Enter
-```
-
-#### Wait for Agent to Complete
-
-```bash
-agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
-(function() {
-  var chat = window.__LOBE_STORES.chat();
-  var ops = Object.values(chat.operations);
-  var running = ops.filter(function(o) { return o.status === 'running'; });
-  return running.length === 0 ? 'done' : 'running: ' + running.length;
-})()
-EVALEOF
-```
-
-#### Install Error Interceptor
-
-```bash
-agent-browser --cdp 9222 eval --stdin << 'EVALEOF'
-(function() {
-  window.__CAPTURED_ERRORS = [];
-  var orig = console.error;
-  console.error = function() {
-    var msg = Array.from(arguments).map(function(a) {
-      if (a instanceof Error) return a.message;
-      return typeof a === 'object' ? JSON.stringify(a) : String(a);
-    }).join(' ');
-    window.__CAPTURED_ERRORS.push(msg);
-    orig.apply(console, arguments);
-  };
-  return 'installed';
-})()
-EVALEOF
-
-# Later, check captured errors:
-agent-browser --cdp 9222 eval "JSON.stringify(window.__CAPTURED_ERRORS)"
-```
-
-## Chrome / Web Apps
-
-```bash
-/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome \
-  --remote-debugging-port=9222 \
-  --user-data-dir=/tmp/chrome-test-profile \
-  "<URL>" &
-sleep 5
-agent-browser --cdp 9222 snapshot -i
-
-# Or auto-discover running Chrome with remote debugging
-agent-browser --auto-connect snapshot -i
-```
-
----
-
-# Part 2: osascript (Native macOS App Bot Testing)
-
-Use AppleScript via `osascript` to control native macOS desktop apps for bot testing. Works with any app that supports macOS Accessibility, no CDP or Chromium needed.
-
-The pattern is the same for every platform:
-
-1. **Activate** the app (`tell application "X" to activate`)
-2. **Navigate** to a channel/chat (Quick Switcher `Cmd+K` or Search `Cmd+F`)
-3. **Send** a message (clipboard paste `Cmd+V` + Enter)
-4. **Wait** for the bot response
-5. **Screenshot** for verification (`screencapture` + `Read` tool)
-
-## Per-Platform References
-
-Pick the file for your target platform — each contains activation, navigation, send-message, and verification snippets specific to that app:
-
-| Platform      | Reference                                          | Quick switcher |
-| ------------- | -------------------------------------------------- | -------------- |
-| Discord       | [references/discord.md](./references/discord.md)   | `Cmd+K`        |
-| Slack         | [references/slack.md](./references/slack.md)       | `Cmd+K`        |
-| Telegram      | [references/telegram.md](./references/telegram.md) | `Cmd+F`        |
-| WeChat / 微信 | [references/wechat.md](./references/wechat.md)     | `Cmd+F`        |
-| Lark / 飞书   | [references/lark.md](./references/lark.md)         | `Cmd+K`        |
-| QQ            | [references/qq.md](./references/qq.md)             | `Cmd+F`        |
-
-For **shared osascript patterns** (activate, type, paste, screenshot, read accessibility, common workflow template, gotchas), see [references/osascript-common.md](./references/osascript-common.md). Read this first if you're new to osascript automation.
-
----
-
-# Scripts
-
-Ready-to-use scripts in `.agents/skills/local-testing/scripts/`:
-
-| Script                    | Usage                                               |
-| ------------------------- | --------------------------------------------------- |
-| `electron-dev.sh`         | Manage Electron dev env (start/stop/status/restart) |
-| `capture-app-window.sh`   | Capture screenshot of a specific app window         |
-| `record-electron-demo.sh` | Record Electron app demo with ffmpeg                |
-| `record-app-screen.sh`    | Record app screen (video + screenshots, start/stop) |
-| `test-discord-bot.sh`     | Send message to Discord bot via osascript           |
-| `test-slack-bot.sh`       | Send message to Slack bot via osascript             |
-| `test-telegram-bot.sh`    | Send message to Telegram bot via osascript          |
-| `test-wechat-bot.sh`      | Send message to WeChat bot via osascript            |
-| `test-lark-bot.sh`        | Send message to Lark / 飞书 bot via osascript       |
-| `test-qq-bot.sh`          | Send message to QQ bot via osascript                |
-
-### Window Screenshot Utility
-
-`capture-app-window.sh` captures a screenshot of a specific app window using `screencapture -l <windowID>`. It uses Swift + CGWindowList to find the window by process name, so screenshots work correctly even when the window is on an external monitor or behind other windows.
-
-```bash
-# Standalone usage
-./.agents/skills/local-testing/scripts/capture-app-window.sh "Discord" /tmp/discord.png
-./.agents/skills/local-testing/scripts/capture-app-window.sh "Slack" /tmp/slack.png
-./.agents/skills/local-testing/scripts/capture-app-window.sh "WeChat" /tmp/wechat.png
-```
-
-All bot test scripts use this utility automatically for their screenshots.
-
-### Bot Test Scripts
-
-All bot test scripts share the same interface:
-
-```bash
-./scripts/test-<platform>-bot.sh <channel_or_contact> <message> [wait_seconds] [screenshot_path]
-```
-
-Examples:
-
-```bash
-# Discord — test a bot in #bot-testing channel
-./.agents/skills/local-testing/scripts/test-discord-bot.sh "bot-testing" "!ping"
-./.agents/skills/local-testing/scripts/test-discord-bot.sh "bot-testing" "/ask Tell me a joke" 30
-
-# Slack — test a bot in #bot-testing channel
-./.agents/skills/local-testing/scripts/test-slack-bot.sh "bot-testing" "@mybot hello"
-./.agents/skills/local-testing/scripts/test-slack-bot.sh "bot-testing" "/ask What is 2+2?" 20
-
-# Telegram — test a bot by username
-./.agents/skills/local-testing/scripts/test-telegram-bot.sh "MyTestBot" "/start"
-./.agents/skills/local-testing/scripts/test-telegram-bot.sh "GPTBot" "Hello" 60
-
-# WeChat — test a bot or send to a contact
-./.agents/skills/local-testing/scripts/test-wechat-bot.sh "文件传输助手" "test message" 5
-./.agents/skills/local-testing/scripts/test-wechat-bot.sh "MyBot" "Tell me a joke" 30
-
-# Lark/飞书 — test a bot in a group chat
-./.agents/skills/local-testing/scripts/test-lark-bot.sh "bot-testing" "@MyBot hello"
-./.agents/skills/local-testing/scripts/test-lark-bot.sh "bot-testing" "Help me with this" 30
-
-# QQ — test a bot in a group or direct chat
-./.agents/skills/local-testing/scripts/test-qq-bot.sh "bot-testing" "Hello bot" 15
-./.agents/skills/local-testing/scripts/test-qq-bot.sh "MyBot" "/help" 10
-```
-
-Each script: activates the app, navigates to the channel/contact, pastes the message via clipboard, sends, waits, and takes a screenshot. Use the `Read` tool on the screenshot for visual verification.
-
----
-
-# Screen Recording
-
-Record automated demos using `record-app-screen.sh` (start/stop lifecycle, CDP screenshots + ffmpeg assembly). See [references/record-app-screen.md](references/record-app-screen.md) for full documentation.
-
-```bash
-./.agents/skills/local-testing/scripts/electron-dev.sh start
-./.agents/skills/local-testing/scripts/record-app-screen.sh start my-demo
-# ... run automation ...
-./.agents/skills/local-testing/scripts/record-app-screen.sh stop
-```
-
-Outputs to `.records/` directory (gitignored): `<name>.mp4` (video) + `<name>/` (screenshots every 3s).
-
----
-
-# Gotchas
-
-### agent-browser
-
-- **Daemon can get stuck** — if commands hang, `agent-browser close --all` or `pkill -f agent-browser` to reset
-- **HMR invalidates everything** — after code changes, refs break. Re-snapshot or restart
-- **`snapshot -i` doesn't find contenteditable** — use `snapshot -i -C` for rich text editors
-- **`fill` doesn't work on contenteditable** — use `type` for chat inputs
-- **Screenshots go to `~/.agent-browser/tmp/screenshots/`** — read them with the `Read` tool
-- **Dialogs block all commands** — if commands time out, check `agent-browser dialog status`
-- **Default timeout is 25s** — override with `AGENT_BROWSER_DEFAULT_TIMEOUT` (ms) or use explicit waits
-- **Shell quoting corrupts eval** — use `eval --stdin <<'EVALEOF'` for complex JS
-
-### Electron-specific
-
-- **Always use `electron-dev.sh stop` to clean up** — `pkill -f "Electron"` only kills the main process; helper processes (GPU, renderer, network) survive. The script finds and kills all of them via PID matching against the project's electron binary path.
-- **`npx electron-vite dev` must run from `apps/desktop/`** — running from project root fails silently. The `electron-dev.sh` script handles this automatically.
-- **Don't resize the Electron window after load** — resizing triggers full SPA reload
-- **Store is at `window.__LOBE_STORES`** not `window.__ZUSTAND_STORES__`
-
-### osascript
-
-See [references/osascript-common.md](./references/osascript-common.md#gotchas) for the full osascript gotchas list (accessibility permissions, `keystroke` non-ASCII issues, locale-specific app names, rate limiting, etc.).

.agents/skills/local-testing/references/agent-browser-login.md

@@ -1,110 +0,0 @@
-# Log `agent-browser` into a local LobeHub dev server
-
-`agent-browser --headed` on macOS often creates the Chromium window off-screen — the user can't see or interact with it, so manual login inside the agent-browser session fails. Instead of sharing the user's real Chrome profile, copy the **better-auth session cookie** out of a request in DevTools and inject it into the agent-browser session as a Playwright-style state file.
-
-## When to use
-
-- You need `agent-browser` to reach an authenticated page on `http://localhost:<port>` (e.g. `localhost:3011`).
-- The user already has a logged-in tab of the same dev server in their own Chrome.
-- Spawning a headed Chromium to let the user log in manually is unreliable (window off-screen, no interaction).
-
-Do **not** use this on production URLs — only local dev. Treat the cookie as a secret: don't paste it into shared logs, PRs, or commit it anywhere.
-
-## Step 1 — Ask the user to copy the cookie from a Network request, NOT `document.cookie`
-
-`document.cookie` will not return HttpOnly cookies, which is exactly where better-auth puts its session. Instruct the user:
-
-1. Open the logged-in tab (`http://localhost:<port>/…`) in their own Chrome.
-2. `Cmd+Option+I` → **Network** tab.
-3. Refresh, click any same-origin request (e.g. the top-level document request).
-4. In the right pane under **Request Headers**, right-click the `Cookie:` line → **Copy value** (or copy the entire header).
-5. Paste the string into chat.
-
-You only need the better-auth pieces. Everything else (Clerk, `LOBE_LOCALE`, HMR hash, theme vars) is noise and can stay. The minimum viable set is:
-
-```
-better-auth.session_token=<value>; better-auth.state=<value>
-```
-
-## Step 2 — Build a Playwright-style state file
-
-`agent-browser state load` expects Playwright's `storageState` format: a JSON with a `cookies` array and an `origins` array.
-
-```bash
-cat > /tmp/mkstate.py << 'PY'
-import json, sys, time
-
-# Read the Cookie header from stdin (allows optional "Cookie: " prefix).
-raw = sys.stdin.read().strip()
-if raw.lower().startswith("cookie:"):
-    raw = raw.split(":", 1)[1].strip()
-
-# Keep only better-auth cookies. Extend this set if the app genuinely needs more.
-WANTED = {"better-auth.session_token", "better-auth.state"}
-
-cookies = []
-exp = int(time.time()) + 30 * 24 * 3600  # 30 days
-for pair in raw.split("; "):
-    if "=" not in pair:
-        continue
-    name, _, value = pair.partition("=")
-    if name not in WANTED:
-        continue
-    cookies.append({
-        "name": name,
-        "value": value,
-        "domain": "localhost",
-        "path": "/",
-        "expires": exp,
-        "httpOnly": False,
-        "secure": False,
-        "sameSite": "Lax",
-    })
-
-if not cookies:
-    sys.stderr.write("no better-auth cookies found in input\n")
-    sys.exit(1)
-
-print(json.dumps({"cookies": cookies, "origins": []}, indent=2))
-PY
-
-# Feed the copied Cookie header in via env var or heredoc.
-printf '%s' "$COOKIE_HEADER" | python3 /tmp/mkstate.py > /tmp/state.json
-```
-
-**Note on `httpOnly`**: the real cookie in the user's browser is HttpOnly, but `storageState` doesn't enforce the flag on load — it just attaches the value. Storing with `httpOnly: false` is fine for local dev and sidesteps a CDP-context quirk where HttpOnly cookies sometimes fail to attach.
-
-## Step 3 — Load state and navigate
-
-```bash
-SESSION="my-test" # any stable session name
-
-agent-browser --session "$SESSION" state load /tmp/state.json
-agent-browser --session "$SESSION" open "http://localhost:3011/"
-agent-browser --session "$SESSION" get url
-# Expect NOT /signin?callbackUrl=… — if you still see signin, cookie didn't apply.
-```
-
-## Step 4 — Verify
-
-```bash
-agent-browser --session "$SESSION" snapshot -i | head -20
-# Look for the user's avatar/name in the sidebar, or absence of the signin form.
-```
-
-## Common failure modes
-
-| Symptom                                         | Cause                                                                   | Fix                                                  |
-| ----------------------------------------------- | ----------------------------------------------------------------------- | ---------------------------------------------------- |
-| Still redirects to `/signin` after `state load` | User pasted from `document.cookie` → missed HttpOnly session            | Re-pull from Network request Headers, not console    |
-| `state load` reports 0 cookies                  | Separator wrong, or user pasted URL-decoded value                       | Keep the raw `Cookie:` header as-is; split on `"; "` |
-| Login works briefly then expires                | `better-auth.session_token` rotated (user logged out / signed in again) | Re-copy and re-load                                  |
-| Domain mismatch                                 | Use `domain: "localhost"` literally, no leading dot for local dev       | —                                                    |
-
-## Scope
-
-Only covers authenticating an **agent-browser** session into a **local** LobeHub dev server. It does not:
-
-- Work for production — production cookies are `Secure; HttpOnly; Domain=.lobehub.com` and must be delivered over HTTPS.
-- Replace real OAuth flows — tests that must exercise the login UI need a real Chromium with `--remote-debugging-port` or a bot account.
-- Flow cookies back to the user's Chrome — injection is one-way (into agent-browser only).

.agents/skills/local-testing/references/discord.md

@@ -1,97 +0,0 @@
-# Discord Bot Testing
-
-**App name:** `Discord` | **Process name:** `Discord`
-
-See [osascript-common.md](./osascript-common.md) for shared patterns.
-
-## Activate & Navigate
-
-```bash
-# Activate Discord
-osascript -e 'tell application "Discord" to activate'
-sleep 1
-
-# Open Quick Switcher (Cmd+K) to navigate to a channel
-osascript -e 'tell application "System Events" to keystroke "k" using command down'
-sleep 0.5
-osascript -e 'tell application "System Events" to keystroke "bot-testing"'
-sleep 1
-osascript -e 'tell application "System Events" to key code 36' # Enter
-sleep 2
-```
-
-## Send Message to Bot
-
-```bash
-# The message input is focused after navigating to a channel
-# Type a message
-osascript -e 'tell application "System Events" to keystroke "/hello"'
-sleep 0.5
-osascript -e 'tell application "System Events" to key code 36' # Enter
-```
-
-## Send Long Message (via clipboard)
-
-```bash
-osascript -e '
-tell application "Discord" to activate
-delay 0.5
-set the clipboard to "Write a 3000 word essay about space exploration"
-tell application "System Events"
-    keystroke "v" using command down
-    delay 0.3
-    key code 36  -- Enter
-end tell
-'
-```
-
-## Verify Bot Response
-
-```bash
-# Wait for bot to respond, then screenshot
-sleep 10
-screencapture /tmp/discord-bot-response.png
-# Read with the Read tool for visual verification
-```
-
-## Full Bot Test Example
-
-```bash
-#!/usr/bin/env bash
-# test-discord-bot.sh — Send message and verify bot response
-
-# 1. Activate Discord and navigate to channel
-osascript -e '
-tell application "Discord" to activate
-delay 1
--- Quick Switcher
-tell application "System Events" to keystroke "k" using command down
-delay 0.5
-tell application "System Events" to keystroke "bot-testing"
-delay 1
-tell application "System Events" to key code 36
-delay 2
-'
-
-# 2. Send test message
-osascript -e '
-set the clipboard to "!ping"
-tell application "System Events"
-    keystroke "v" using command down
-    delay 0.3
-    key code 36
-end tell
-'
-
-# 3. Wait for response and capture
-sleep 5
-screencapture /tmp/discord-test-result.png
-echo "Screenshot saved to /tmp/discord-test-result.png"
-```
-
-## Script
-
-```bash
-./.agents/skills/local-testing/scripts/test-discord-bot.sh "bot-testing" "!ping"
-./.agents/skills/local-testing/scripts/test-discord-bot.sh "bot-testing" "/ask Tell me a joke" 30
-```

.agents/skills/local-testing/references/lark.md

@@ -1,61 +0,0 @@
-# Lark / 飞书 Bot Testing
-
-**App name:** `Lark` or `飞书` | **Process name:** `Lark` or `飞书`
-
-See [osascript-common.md](./osascript-common.md) for shared patterns.
-
-## Activate & Navigate
-
-```bash
-# Activate Lark (auto-detects Lark or 飞书)
-osascript -e 'tell application "Lark" to activate' 2> /dev/null \
-  || osascript -e 'tell application "飞书" to activate'
-sleep 1
-
-# Quick Switcher / Search (Cmd+K)
-osascript -e 'tell application "System Events" to keystroke "k" using command down'
-sleep 0.5
-osascript -e '
-set the clipboard to "bot-testing"
-tell application "System Events"
-    keystroke "v" using command down
-    delay 1.5
-    key code 36  -- Enter
-end tell
-'
-sleep 2
-```
-
-## Send Message to Bot
-
-```bash
-osascript -e '
-set the clipboard to "@MyBot help me with this task"
-tell application "System Events"
-    keystroke "v" using command down
-    delay 0.3
-    key code 36  -- Enter
-end tell
-'
-```
-
-## Verify Response
-
-```bash
-sleep 10
-screencapture /tmp/lark-bot-response.png
-```
-
-## Lark-Specific Notes
-
-- App name varies: `Lark` (international) vs `飞书` (China mainland) — the script auto-detects
-- Uses `Cmd+K` for quick search (same as Discord/Slack)
-- Enter sends message by default
-- Always use clipboard paste for CJK characters
-
-## Script
-
-```bash
-./.agents/skills/local-testing/scripts/test-lark-bot.sh "bot-testing" "@MyBot hello"
-./.agents/skills/local-testing/scripts/test-lark-bot.sh "bot-testing" "Help me with this" 30
-```

.agents/skills/local-testing/references/qq.md

@@ -1,62 +0,0 @@
-# QQ Bot Testing
-
-**App name:** `QQ` | **Process name:** `QQ`
-
-See [osascript-common.md](./osascript-common.md) for shared patterns.
-
-## Activate & Navigate
-
-```bash
-osascript -e 'tell application "QQ" to activate'
-sleep 1
-
-# Search for contact/group (Cmd+F)
-osascript -e '
-tell application "System Events"
-    keystroke "f" using command down
-    delay 0.8
-end tell
-'
-osascript -e '
-set the clipboard to "bot-testing"
-tell application "System Events"
-    keystroke "v" using command down
-    delay 1.5
-    key code 36  -- Enter
-end tell
-'
-sleep 2
-```
-
-## Send Message to Bot
-
-```bash
-osascript -e '
-set the clipboard to "Hello bot!"
-tell application "System Events"
-    keystroke "v" using command down
-    delay 0.3
-    key code 36  -- Enter
-end tell
-'
-```
-
-## Verify Response
-
-```bash
-sleep 10
-screencapture /tmp/qq-bot-response.png
-```
-
-## QQ-Specific Notes
-
-- Enter sends message by default; Shift+Enter for newlines
-- Uses `Cmd+F` for search (not `Cmd+K` like Discord/Slack/Lark)
-- Always use clipboard paste for CJK characters
-
-## Script
-
-```bash
-./.agents/skills/local-testing/scripts/test-qq-bot.sh "bot-testing" "Hello bot" 15
-./.agents/skills/local-testing/scripts/test-qq-bot.sh "MyBot" "/help" 10
-```

.agents/skills/local-testing/references/slack.md

@@ -1,73 +0,0 @@
-# Slack Bot Testing
-
-**App name:** `Slack` | **Process name:** `Slack`
-
-See [osascript-common.md](./osascript-common.md) for shared patterns.
-
-## Activate & Navigate
-
-```bash
-# Activate Slack
-osascript -e 'tell application "Slack" to activate'
-sleep 1
-
-# Quick Switcher (Cmd+K)
-osascript -e 'tell application "System Events" to keystroke "k" using command down'
-sleep 0.5
-osascript -e 'tell application "System Events" to keystroke "bot-testing"'
-sleep 1
-osascript -e 'tell application "System Events" to key code 36' # Enter
-sleep 2
-```
-
-## Send Message to Bot
-
-```bash
-# Direct message input (focused after channel nav)
-osascript -e 'tell application "System Events" to keystroke "@mybot hello"'
-sleep 0.3
-osascript -e 'tell application "System Events" to key code 36'
-```
-
-## Send Long Message
-
-```bash
-osascript -e '
-tell application "Slack" to activate
-delay 0.5
-set the clipboard to "A long test message for the bot..."
-tell application "System Events"
-    keystroke "v" using command down
-    delay 0.3
-    key code 36
-end tell
-'
-```
-
-## Slash Command Test
-
-```bash
-osascript -e '
-tell application "Slack" to activate
-delay 0.5
-tell application "System Events"
-    keystroke "/ask What is the meaning of life?"
-    delay 0.5
-    key code 36
-end tell
-'
-```
-
-## Verify Response
-
-```bash
-sleep 10
-screencapture /tmp/slack-bot-response.png
-```
-
-## Script
-
-```bash
-./.agents/skills/local-testing/scripts/test-slack-bot.sh "bot-testing" "@mybot hello"
-./.agents/skills/local-testing/scripts/test-slack-bot.sh "bot-testing" "/ask What is 2+2?" 20
-```

.agents/skills/local-testing/references/telegram.md

@@ -1,80 +0,0 @@
-# Telegram Bot Testing
-
-**App name:** `Telegram` | **Process name:** `Telegram`
-
-See [osascript-common.md](./osascript-common.md) for shared patterns.
-
-## Activate & Navigate
-
-```bash
-# Activate Telegram
-osascript -e 'tell application "Telegram" to activate'
-sleep 1
-
-# Search for a bot (Cmd+F or click search)
-osascript -e '
-tell application "System Events"
-    keystroke "f" using command down
-    delay 0.5
-    keystroke "MyTestBot"
-    delay 1
-    key code 36  -- Enter to select
-end tell
-'
-sleep 2
-```
-
-## Send Message to Bot
-
-```bash
-# After navigating to bot chat, input is focused
-osascript -e '
-tell application "System Events"
-    keystroke "/start"
-    delay 0.3
-    key code 36
-end tell
-'
-```
-
-## Send Long Message
-
-```bash
-osascript -e '
-tell application "Telegram" to activate
-delay 0.5
-set the clipboard to "Tell me about quantum computing in detail"
-tell application "System Events"
-    keystroke "v" using command down
-    delay 0.3
-    key code 36
-end tell
-'
-```
-
-## Verify Response
-
-```bash
-sleep 10
-screencapture /tmp/telegram-bot-response.png
-```
-
-## Telegram Bot API (programmatic alternative)
-
-For sending messages directly to the bot's chat without UI:
-
-```bash
-# Send message as the bot (for testing webhooks/responses)
-curl -s "https://api.telegram.org/bot$TELEGRAM_BOT_TOKEN/sendMessage" \
-  -d "chat_id=$CHAT_ID&text=test message"
-
-# Get recent updates
-curl -s "https://api.telegram.org/bot$TELEGRAM_BOT_TOKEN/getUpdates?limit=5" | jq .
-```
-
-## Script
-
-```bash
-./.agents/skills/local-testing/scripts/test-telegram-bot.sh "MyTestBot" "/start"
-./.agents/skills/local-testing/scripts/test-telegram-bot.sh "GPTBot" "Hello" 60
-```

.agents/skills/local-testing/references/wechat.md

@@ -1,81 +0,0 @@
-# WeChat / 微信 Bot Testing
-
-**App name:** `微信` or `WeChat` | **Process name:** `WeChat`
-
-See [osascript-common.md](./osascript-common.md) for shared patterns.
-
-## Activate & Navigate
-
-```bash
-# Activate WeChat
-osascript -e 'tell application "微信" to activate'
-sleep 1
-
-# Search for a contact/bot (Cmd+F)
-osascript -e '
-tell application "System Events"
-    keystroke "f" using command down
-    delay 0.5
-    keystroke "TestBot"
-    delay 1
-    key code 36  -- Enter to select
-end tell
-'
-sleep 2
-```
-
-## Send Message
-
-```bash
-# After navigating to a chat, the input is focused
-osascript -e '
-tell application "System Events"
-    keystroke "Hello bot!"
-    delay 0.3
-    key code 36
-end tell
-'
-```
-
-## Send Long Message (clipboard)
-
-```bash
-osascript -e '
-tell application "微信" to activate
-delay 0.5
-set the clipboard to "Please help me with this task..."
-tell application "System Events"
-    keystroke "v" using command down
-    delay 0.3
-    key code 36
-end tell
-'
-```
-
-## Verify Response
-
-```bash
-sleep 10
-screencapture /tmp/wechat-bot-response.png
-```
-
-## WeChat-Specific Notes
-
-- WeChat macOS app name can be `微信` or `WeChat` depending on system language. Try both:
-  ```bash
-  osascript -e 'tell application "微信" to activate' 2> /dev/null \
-    || osascript -e 'tell application "WeChat" to activate'
-  ```
-- WeChat uses **Enter** to send (not Cmd+Enter by default, but configurable)
-- For multi-line messages without sending, use **Shift+Enter**:
-  ```bash
-  osascript -e 'tell application "System Events" to key code 36 using shift down'
-  ```
-- Always use clipboard paste for CJK characters — `keystroke` mangles non-ASCII
-
-## Script
-
-```bash
-./.agents/skills/local-testing/scripts/test-wechat-bot.sh "文件传输助手" "test message" 5
-./.agents/skills/local-testing/scripts/test-wechat-bot.sh "MyBot" "Tell me a joke" 30
-```

.agents/skills/microcopy/SKILL.md

@@ -1,10 +1,16 @@
 ---
 name: microcopy
-description: UI copy and microcopy guidelines. Use when writing UI text, buttons, error messages, empty states, onboarding, or any user-facing copy. Triggers on i18n translation, UI text writing, or copy improvement tasks. Supports both Chinese and English.
+description: 'UI copy and microcopy guidelines. Use for user-facing copy, buttons, errors, empty states, onboarding, i18n wording, translation, or copy improvements in Chinese or English.'
+user-invocable: false
 ---
 
 # LobeHub UI Microcopy Guidelines
 
+This file is the quick-reference summary. For full prompt-style guidelines with extensive examples (anti-patterns, tone matrices, scenario walk-throughs), load the language-specific reference:
+
+- **中文文案** — [`references/zh.md`](./references/zh.md)
+- **English copy** — [`references/en.md`](./references/en.md)
+
 Brand: **Where Agents Collaborate** - Focus on collaborative agent system, not just "generation".
 
 ## Fixed Terminology

.agents/skills/modal/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: modal
-description: MUST use when creating, editing, or writing modal dialogs or imperative modals. Prefer createModal / useModalContext / confirmModal from @lobehub/ui/base-ui; root @lobehub/ui is legacy (antd Modal). Covers patterns, ModalHost, and migration notes.
+description: 'LobeHub imperative modal conventions. Use when creating or migrating modals, dialogs, popups, confirm flows, ModalHost wiring, createModal, confirmModal, useModalContext, or base-ui modal APIs.'
 user-invocable: false
 ---