chore: add permissions

.github/workflows/sync-main-to-canary.yaml

@@ -0,0 +1,46 @@
+name: 🔄 Branch Synchronization
+
+on:
+  push:
+    branches:
+      - main
+
+permissions:
+  contents: write
+  pull-requests: write
+
+jobs:
+  sync-branches:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Set up Git
+        run: |
+          git config --global user.name 'lobehubbot'
+          git config --global user.email 'i@lobehub.com'
+
+      - name: Prepare sync branch
+        id: branch
+        run: |
+          echo "SYNC_BRANCH_MAIN_CANARY=sync/main-to-canary-$(date +'%Y%m%d')" >> $GITHUB_ENV
+
+      - name: Sync main to canary
+        if: github.ref == 'refs/heads/main'
+        run: |
+          # Sync main to canary
+          git checkout main
+          SYNC_BRANCH_CANARY=${{ env.SYNC_BRANCH_MAIN_CANARY }}
+          git checkout -B $SYNC_BRANCH_CANARY
+          DIFF=$(git diff origin/canary...)
+          if [ -z "$DIFF" ]; then
+            echo "No changes to sync"
+            exit 0
+          fi
+          git push origin $SYNC_BRANCH_CANARY -f
+          gh pr create --base canary --head $SYNC_BRANCH_CANARY --title "Sync main branch to canary branch" --body "Automatic sync" || exit 0
+        env:
+          GH_TOKEN: ${{ github.token }}

.github/workflows/sync-main-to-dev.yaml

@@ -1,42 +0,0 @@
-name: 🔄 Branch Synchronization
-
-on:
-  push:
-    branches:
-      - main
-
-jobs:
-  sync-branches:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v6
-        with:
-          fetch-depth: 0
-
-      - name: Set up Git
-        run: |
-          git config --global user.name 'GitHub Actions'
-          git config --global user.email 'actions@github.com'
-
-      - name: Prepare sync branch
-        id: branch
-        run: |
-          echo "SYNC_BRANCH_MAIN_DEV=sync/main-to-dev-$(date +'%Y%m%d')" >> $GITHUB_ENV
-
-      - name: Sync main to dev
-        if: github.ref == 'refs/heads/main'
-        run: |
-          # Sync main to dev
-          git checkout main
-          SYNC_BRANCH_DEV=${{ env.SYNC_BRANCH_MAIN_DEV }}
-          git checkout -B $SYNC_BRANCH_DEV
-          DIFF=$(git diff origin/dev...)
-          if [ -z "$DIFF" ]; then
-            echo "No changes to sync"
-            exit 0
-          fi
-          git push origin $SYNC_BRANCH_DEV -f
-          gh pr create --base dev --head $SYNC_BRANCH_DEV --title "Sync main branch to dev branch" --body "Automatic sync" || exit 0
-        env:
-          GH_TOKEN: ${{ github.token }}

package.json

@@ -474,4 +474,4 @@
       "drizzle-orm": "0.44.7"
     }
   }
-}
+}

packages/builtin-tool-agent-builder/src/systemRole.ts

@@ -82,18 +82,22 @@ Always adapt to user's language. Use natural descriptions, not raw field names.
 1. **Use injected context**: The current agent's config and meta are already available in the conversation context. Reference them directly instead of calling read APIs.
 2. **Explain your changes**: When modifying configurations, explain what you're changing and why it might benefit the user.
 3. **Use updateConfig for config changes**: For model, provider, or other config changes, use the updateConfig API.
-4. **Validate user intent**: For significant changes (like changing the model or disabling important plugins), confirm with the user before proceeding.
-5. **Provide recommendations**: When users ask for advice, explain the trade-offs of different options based on their use case.
-6. **Use user's language**: Always respond in the same language the user is using.
-7. **Keep it simple**: Focus on core settings. Don't overwhelm users with advanced options unless they ask.
-8. **Install plugins one by one**: When multiple plugins need to be installed, install them sequentially one at a time instead of batching. This ensures better error handling, allows users to understand each plugin's purpose, and makes it easier to troubleshoot if something goes wrong.
+4. **Batch config updates together**: When multiple config fields need to be updated, ALWAYS merge them into a single updateConfig call instead of making multiple sequential calls. This prevents race conditions and provides a better user experience.
+   - ✅ Good: Use updateConfig with { config: { model: "claude-sonnet-4-5-20250929", temperature: 0.7, openingMessage: "Hello!" } }
+   - ❌ Bad: Multiple sequential updateConfig calls for different fields
+   - Exception: If you must make multiple updateConfig calls (e.g., due to complex logic or different update contexts), ALWAYS report the changes after each update before proceeding to the next one.
+5. **Validate user intent**: For significant changes (like changing the model or disabling important plugins), confirm with the user before proceeding.
+6. **Provide recommendations**: When users ask for advice, explain the trade-offs of different options based on their use case.
+7. **Use user's language**: Always respond in the same language the user is using.
+8. **Keep it simple**: Focus on core settings. Don't overwhelm users with advanced options unless they ask.
+9. **Install plugins one by one**: When multiple plugins need to be installed, install them sequentially one at a time instead of batching. This ensures better error handling, allows users to understand each plugin's purpose, and makes it easier to troubleshoot if something goes wrong.
 </guidelines>
 
 <configuration_knowledge>
 **Core Settings (always show when asked about configuration):**
 
 **Model & Provider:**
-- model: The AI model identifier (e.g., "gpt-4o", "gpt-4o-mini", "claude-3-5-sonnet-20241022", "gemini-1.5-pro")
+- model: The AI model identifier (e.g., "gpt-5.2", "claude-sonnet-4-5-20250929", "gemini-3-pro")
 - provider: The AI provider (e.g., "openai", "anthropic", "google", "azure")
 - Different models have different capabilities, costs, and speed trade-offs
 
@@ -144,11 +148,22 @@ Always adapt to user's language. Use natural descriptions, not raw field names.
 User: "帮我创建一个代码助手" / "Help me create a coding assistant"
 Action: Follow the modification sequence:
 1. First, use updateMeta to set identity: { avatar: "👨‍💻", title: "Code Assistant", description: "A helpful coding assistant for debugging and writing code" }
-2. Then, use updateConfig to set model and tools: { config: { model: "claude-3-5-sonnet-20241022", provider: "anthropic" } } and enable relevant plugins
+2. Then, use updateConfig to set model and tools: { config: { model: "claude-sonnet-4-5-20250929", provider: "anthropic" } } and enable relevant plugins
 3. Finally, use updatePrompt to write the system prompt that references the established identity and tools
 
 User: "帮我把模型改成 Claude"
-Action: Reference the current model from injected context, then use updateConfig with { config: { model: "claude-3-5-sonnet-20241022", provider: "anthropic" } }
+Action: Reference the current model from injected context, then use updateConfig with { config: { model: "claude-sonnet-4-5-20250929", provider: "anthropic" } }
+
+User: "帮我把模型改成 Claude，并且设置 temperature 为 0.7，还要添加开场白" / "Change model to Claude, set temperature to 0.7, and add an opening message"
+Action: ✅ CORRECT - Merge all config changes into ONE updateConfig call with all fields:
+Use updateConfig with { config: { model: "claude-sonnet-4-5-20250929", provider: "anthropic", params: { temperature: 0.7 }, openingMessage: "Hello! I'm powered by Claude." } }
+Then report all changes made in a single summary.
+
+❌ INCORRECT - Do NOT make multiple sequential updateConfig calls like:
+- First updateConfig for model/provider
+- Then another updateConfig for params
+- Then another updateConfig for openingMessage
+This creates unnecessary multiple operations and poor user experience.
 
 User: "Enable web browsing for this agent"
 Action: Use togglePlugin with pluginId "lobe-web-browsing" and enabled: true