Address PR review comment - use process.cwd() for Windows too instead of
USERPROFILE/homedir fallback for consistency.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Use node-pty on all platforms instead of expect on macOS for more reliable PTY handling
- Use process.cwd() as working directory (project dir is likely already trusted)
- Add detection for new trust prompt text variants ("Ready to code here", "permission to work")
- Add specific error handling for trust prompt pending state
- Show helpful UI message when trust prompt needs manual approval
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
CLI-based providers (OpenCode, etc.) only accept a single prompt via
stdin/args and don't support separate system/user message channels like
Claude SDK. When systemPrompt is passed to these providers, it was
silently dropped, causing:
- BacklogPlan JSON parsing failures with OpenCode/GPT-5.2 (missing
"output ONLY JSON" formatting instruction)
- Loss of critical formatting/schema instructions for structured outputs
This fix adds embedSystemPromptIntoPrompt() method to CliProvider base
class that:
- Prepends systemPrompt to the user prompt before CLI execution
- Handles both string and array prompts (vision support)
- Handles both string systemPrompt and SystemPromptPreset objects
- Uses standard \n\n---\n\n separator (consistent with codebase)
- Sets systemPrompt to undefined to prevent double-injection
Benefits OpencodeProvider immediately (uses base executeQuery).
CursorProvider still uses manual workarounds (overrides executeQuery).
Fixes the immediate BacklogPlan + OpenCode bug while maintaining
backward compatibility with existing Cursor workarounds.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Updated create-pr.ts to improve commit error handling and logging.
- Enhanced project-switcher.tsx with new folder opening functionality and state management for project setup.
- Expanded icon-picker.tsx to include a comprehensive list of icons organized by category.
- Replaced dialog components with popover components for auto mode and plan settings, improving UI responsiveness.
- Refactored board-view components to streamline feature management and enhance user experience.
- Removed outdated dialog components and replaced them with popover alternatives for better accessibility.
These changes aim to improve the overall usability and functionality of the project management interface.
- Fix handleResolveConflicts to use origin/${worktree.branch} instead of
hardcoded origin/main for pull and resolve conflicts
- Add defaultBaseBranch prop to CreatePRDialog to use selected branch
- Fix branchCardCounts to use primary worktree branch as default
- Enable PR status and Address PR Comments for main branch tab
- Add automatic PR detection from GitHub for branches without stored metadata
This allows users working on release branches (like v0.11.0rc) to properly
pull from their branch's remote and see PR status for any branch.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Resolved conflict in worktree-panel.tsx by combining imports:
- DevServerLogsPanel from this branch
- WorktreeMobileDropdown, WorktreeActionsDropdown, BranchSwitchDropdown from v0.11.0rc
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add the ability to view dev server logs in a dedicated panel with:
- Real-time log streaming via WebSocket events
- ANSI color support using xterm.js
- Scrollback buffer (50KB) for log history on reconnect
- Output throttling to prevent UI flooding
- "View Logs" option in worktree dropdown menu
Server changes:
- Add scrollback buffer and event emission to DevServerService
- Add GET /api/worktree/dev-server-logs endpoint
- Add dev-server:started, dev-server:output, dev-server:stopped events
UI changes:
- Add reusable XtermLogViewer component
- Add DevServerLogsPanel dialog component
- Add useDevServerLogs hook for WebSocket subscription
Closes#462
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Modified command arguments in tests to include '--add-dir' for better context.
- Updated error messages for authentication and timeout scenarios to provide clearer guidance.
- Adjusted timer values in tests to align with implementation delays, ensuring accurate simulation of usage data retrieval.
- Skip learning extraction when ANTHROPIC_API_KEY is not available
- Add reasoningEffort parameter to simpleQuery for Codex model configuration
- Add stdinData support to spawnProcess for CLI stdin input
- Update UI API types for model override with reasoning support
- Support Codex and OpenCode models in issue validation
- Add reasoningEffort parameter for Codex model configuration
- Update validation logic to use structured output for Claude/Codex
- Update UI hooks and types for multi-provider model selection
- Add OpenAI API key storage to store-api-key handler
- Include Google/OpenAI key status in credentials API responses
- Add unified ModelId type for Claude, Codex, Cursor, OpenCode, and dynamic providers
- Update PhaseModelEntry to support all provider model types
When an OpenAI API key is stored in settings or environment, use SDK mode
instead of CLI mode. This bypasses the MCP transport layer which was
failing with 'TokenRefreshFailed' errors due to OAuth token issues.
The SDK uses the API key directly via @openai/codex-sdk, avoiding the
OAuth token refresh mechanism that was causing mid-execution failures.
Fix Prettier formatting in two files:
- apps/server/src/lib/sdk-options.ts: Split long arrays to one item per line
- docs/docker-isolation.md: Align markdown table columns
Resolves CI format check failures.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
- Replaced dynamic import of the query function with a call to the new Simple Query Service for improved clarity and maintainability.
- Streamlined the response handling by directly utilizing the result from the simple query, enhancing code readability.
- Updated the prompt and options structure to align with the new service's requirements, ensuring consistent behavior in learning extraction.
- Introduced a new Simple Query Service to streamline basic AI queries, allowing for structured JSON outputs.
- Updated existing routes to utilize the new service, replacing direct SDK calls with a unified interface for querying.
- Enhanced provider handling in various routes, including generate-spec, generate-features-from-spec, and validate-issue, to support both Claude and Cursor models seamlessly.
- Added structured output support for improved response handling and error management across the application.
- Added validation to check if the specified worktree path exists before generating commit messages.
- Implemented a check to ensure the worktree path is a valid git repository by verifying the presence of the .git directory.
- Improved error handling by returning appropriate responses for invalid paths and non-git repositories.
- Implemented a new endpoint to generate commit messages based on git diffs.
- Updated worktree routes to include the AI commit message generation functionality.
- Enhanced the UI to support automatic generation of commit messages when the commit dialog opens, based on user settings.
- Added settings for enabling/disabling AI-generated commit messages and configuring the model used for generation.
Integrate Claude Haiku to automatically generate commit messages when
committing worktree changes. Shows a sparkle animation while generating
and auto-populates the commit message field.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Introduced a flag to track if a branch was successfully deleted, improving response clarity.
- Updated the response structure to include the new branchDeleted flag.
- Enhanced projectPath validation in init-script to ensure it is a non-empty string before processing.
- Replace busy-wait loop in refreshModels with Promise-based approach
- Remove duplicate error logging in opencode-models.ts handlers
- Fix multi-slash parsing in provider-icon.tsx (only handle exactly one slash)
- Use dynamic icon resolution for selected OpenCode model in trigger
- Fix misleading comment about merge precedence (static takes precedence)
- Add enabledOpencodeModels and opencodeDefaultModel to settings sync
- Add clarifying comments about session-only dynamic model settings
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Update error event interface to handle nested error objects with
name/data/message structure from OpenCode CLI
- Extract meaningful error messages from provider errors in normalizeEvent
- Add error type handling in executeWithProvider to throw errors with
actual provider messages instead of returning empty response
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Update isOpencodeModel() to detect dynamic models with provider/model format
(e.g., github-copilot/gpt-4o, google/gemini-2.5-pro, zai-coding-plan/glm-4.7)
- Update resolveModelString() to recognize and pass through OpenCode models
- Update enhance route to route OpenCode models to OpenCode provider
- Fix OpenCode CLI command format: use --format json (not stream-json)
- Remove unsupported -q and - flags from CLI arguments
- Update normalizeEvent() to handle actual OpenCode JSON event format
- Add dynamic model configuration UI with provider grouping
- Cache providers and models in app store for snappier navigation
- Show authenticated providers in OpenCode CLI status card
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Updated apply handler to accept an optional branchName from the request body.
- Modified BoardView and BacklogPlanDialog components to pass currentBranch to the apply API.
- Enhanced ElectronAPI and HttpApiClient to include branchName in the apply method.
This change allows users to specify a branch when applying backlog plans, improving flexibility in feature management.
The TodoWrite tool was missing from the fullAccess and chat tool
presets, causing the Claude Agent SDK to crash with exit code 1
when the agent attempted to use it for task tracking.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add libs/platform/src/editor.ts with cross-platform editor detection and launching
- Handles Windows .cmd batch scripts (cursor.cmd, code.cmd, etc.)
- Supports macOS app bundles in /Applications and ~/Applications
- Includes caching with 5-minute TTL for performance
- Refactor open-in-editor.ts to use @automaker/platform utilities
- Add POST /api/worktree/refresh-editors endpoint to clear cache
- Add refresh button to Settings > Account for IDE selection
- Update useAvailableEditors hook with refresh() and isRefreshing
Fixes Windows issue where "Open in Editor" was falling back to Explorer
because execFile cannot run .cmd scripts without shell:true.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Introduced a new parameter `preEnhancementDescription` to capture the original description before enhancements.
- Updated the `update` method in `FeatureLoader` to handle the new parameter and maintain a history of original descriptions.
- Enhanced UI components to support tracking and restoring pre-enhancement descriptions across various dialogs.
- Improved history management in `AddFeatureDialog`, `EditFeatureDialog`, and `FollowUpDialog` to include original text for better user experience.
This change enhances the ability to revert to previous descriptions, improving the overall functionality of the feature enhancement process.
Cache management improvements:
- Remove separate cachedEditor variable; derive default from cachedEditors
- Update isCacheValid() to check cachedEditors existence
- detectDefaultEditor() now always goes through detectAllEditors()
to ensure cache TTL is respected consistently
Editor fallback improvements:
- Log warning when requested editorCommand is not found in available editors
- Include list of available editor commands in warning message
- Make fallback to default editor explicit rather than silent
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Security improvements in open-in-editor.ts:
- Use execFile with argument arrays instead of shell interpolation
in commandExists() to prevent command injection
- Replace shell `test -d` commands with Node.js fs/promises access()
in findMacApp() for safer file system checks
- Add cache TTL (5 minutes) for editor detection to prevent stale data
UX improvements in worktree-actions-dropdown.tsx:
- Add error handling for clipboard copy operation
- Show success toast when path is copied
- Show error toast if clipboard write fails
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
## Security Fix (Command Injection)
- Use `execFile` with argument arrays instead of string interpolation
- Add `safeOpenInEditor` helper that properly handles `open -a` commands
- Validate that worktreePath is an absolute path before execution
- Prevents shell metacharacter injection attacks
## Shared Type Definition
- Move `EditorInfo` interface to `@automaker/types` package
- Server and UI now import from shared package to prevent drift
- Re-export from use-available-editors.ts for convenience
## Remove Unused Code
- Remove unused `defaultEditorName` prop from WorktreeActionsDropdown
- Remove prop from WorktreeTab component interface
- Remove useDefaultEditor hook usage from WorktreePanel
- Export new hooks from hooks/index.ts
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Addresses feedback from gemini-code-assist and coderabbitai reviewers:
## Duplicate Code (High Priority)
- Extract `getEffectiveDefaultEditor` logic into shared `useEffectiveDefaultEditor` hook
- Both account-section.tsx and worktree-actions-dropdown.tsx now use the shared hook
## Performance (Medium Priority)
- Refactor `detectAllEditors` to use `Promise.all` for parallel editor detection
- Replace sequential `await tryAddEditor()` calls with parallel `findEditor()` checks
## Code Quality (Medium Priority)
- Remove verbose IIFE pattern for editor icon rendering
- Pre-compute icon components before JSX return statement
## Bug Fixes
- Use `os.homedir()` instead of `~` fallback which doesn't expand in shell
- Normalize Select value to 'auto' when saved editor command not found in editors
- Add defensive check for empty editors array in useEffectiveDefaultEditor
- Improve mock openInEditor to correctly map all editor commands to display names
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
