On Linux, the ESC key doesn't exit the Claude CLI, causing a 30s timeout.
This fix:
1. Adds SIGTERM fallback 2s after ESC fails
2. Returns captured data on timeout instead of failing
Tested: ~19s on Linux instead of 30s timeout.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Replace console.error calls with createLogger for consistent logging across
the AgentService. This improves debuggability and makes logger calls testable.
Changes:
- Add createLogger import from @automaker/utils
- Add private logger instance initialized with 'AgentService' prefix
- Replace all 7 console.error calls with this.logger.error
- Update test mocks to use vi.hoisted() for proper mock access
- Update settings-helpers test to create mockLogger inside vi.mock()
Test Impact:
- All 774 tests passing
- Logger error calls are now verifiable in tests
- Mock logger properly accessible via vi.hoisted() pattern
Resolves Gemini Code Assist suggestions:
- "Make logger mockable for test assertions"
- "Use logger instead of console.error in AgentService"
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Move Cursor-specific duplicate text handling from auto-mode-service.ts
into CursorProvider.deduplicateTextBlocks() for cleaner separation.
This handles:
- Duplicate consecutive text blocks (same text twice in a row)
- Final accumulated text block (contains ALL previous text)
Also update REFACTORING-ANALYSIS.md with SpawnStrategy types for
future CLI providers (wsl, npx, direct, cmd).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Replace 'claude' | 'cursor' literal unions with ModelProvider type
from @automaker/types for better extensibility when adding new providers.
- Update ProviderFactory.getProviderNameForModel() return type
- Update RunningFeature.provider type in auto-mode-service
- Update getRunningAgents() return type
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add Record<string, string> type to planningPrompts object to fix TypeScript
error when using string as index.
Error fixed:
Element implicitly has an 'any' type because expression of type 'string'
can't be used to index type '{ lite: string; ... }'.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Remove caching from Auto Mode and Agent services to allow custom prompts
to take effect immediately without requiring app restart.
Changes:
- Auto Mode: Load prompts on every feature execution instead of caching
- Agent Service: Load prompts on every chat message instead of caching
- Remove unused class fields: planningPrompts, agentSystemPrompt
This makes custom prompts work consistently across all features:
✓ Auto Mode - hot reload enabled
✓ Agent Runner - hot reload enabled
✓ Backlog Plan - already had hot reload
✓ Enhancement - already had hot reload
Users can now modify prompts in Settings and see changes immediately
without restarting the app.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Add comprehensive prompt customization system allowing users to customize
all AI prompts (Auto Mode, Agent Runner, Backlog Plan, Enhancement) through
the Settings UI.
## Features
### Core Customization System
- New TypeScript types for prompt customization with enabled flag
- CustomPrompt interface with value and enabled state
- Prompts preserved even when disabled (no data loss)
- Merged prompt system (custom overrides defaults when enabled)
- Persistent storage in ~/.automaker/settings.json
### Settings UI
- New "Prompt Customization" section in Settings
- 4 tabs: Auto Mode, Agent, Backlog Plan, Enhancement
- Toggle-based editing (read-only default → editable custom)
- Dynamic textarea height based on prompt length (120px-600px)
- Visual state indicators (Custom/Default labels)
### Warning System
- Critical prompt warnings for Backlog Plan (JSON format requirement)
- Field-level warnings when editing critical prompts
- Info banners for Auto Mode planning markers
- Color-coded warnings (blue=info, amber=critical)
### Backend Integration
- Auto Mode service loads prompts from settings
- Agent service loads prompts from settings
- Backlog Plan service loads prompts from settings
- Enhancement endpoint loads prompts from settings
- Settings sync includes promptCustomization field
### Files Changed
- libs/types/src/prompts.ts - Type definitions
- libs/prompts/src/defaults.ts - Default prompt values
- libs/prompts/src/merge.ts - Merge utilities
- apps/ui/src/components/views/settings-view/prompts/ - UI components
- apps/server/src/lib/settings-helpers.ts - getPromptCustomization()
- All service files updated to use customizable prompts
## Technical Details
Prompt storage format:
```json
{
"promptCustomization": {
"autoMode": {
"planningLite": {
"value": "Custom prompt text...",
"enabled": true
}
}
}
}
```
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
- Add MCPTestService for testing MCP server connections
- Support stdio, SSE, and HTTP transport types
- Implement workaround for SSE headers bug (SDK Issue #436)
- Create API routes for /api/mcp/test and /api/mcp/tools
- Add API client methods for MCP operations
- Create MCPToolsList component with collapsible schema display
- Add Test button to MCP servers section with status indicators
- Add Headers field for HTTP/SSE servers
- Add Environment Variables field for stdio servers
- Fix text overflow in tools list display
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Introduced a new pipeline service to manage custom workflow steps that execute after a feature is marked "In Progress".
- Added API endpoints for configuring, saving, adding, updating, deleting, and reordering pipeline steps.
- Enhanced the UI to support pipeline settings, including a dialog for managing steps and integration with the Kanban board.
- Updated the application state management to handle pipeline configurations per project.
- Implemented dynamic column generation in the Kanban board to display pipeline steps between "In Progress" and "Waiting Approval".
- Added documentation for the new pipeline feature, including usage instructions and configuration details.
This feature allows for a more structured workflow, enabling automated processes such as code reviews and testing after feature implementation.
- Introduced a new environment variable `AUTOMAKER_DEBUG_RAW_OUTPUT` to enable raw output logging for agent streams.
- Added a new endpoint `/raw-output` to retrieve raw JSONL output for debugging purposes.
- Implemented functionality in `AutoModeService` to log raw output events and save them to `raw-output.jsonl`.
- Enhanced `FeatureLoader` to provide access to raw output files.
- Updated UI components to clean fragmented streaming text for better log parsing.
- Updated AutoModeService to track model and provider for running features, improving logging and state management.
- Modified AddFeatureDialog to handle model selection for both Claude and Cursor, adjusting thinking level logic accordingly.
- Expanded ModelSelector to allow provider selection and dynamically display models based on the selected provider.
- Introduced new model constants for Cursor models, integrating them into the existing model management structure.
- Updated README and project plan to reflect the completion of task execution integration for Cursor models.
Add Model Context Protocol (MCP) server integration to extend AI agent
capabilities with external tools. This allows users to configure MCP
servers (stdio, SSE, HTTP) in global settings and have agents use them.
Note: MCP servers are currently configured globally. Per-project MCP
server configuration is planned for a future update.
Features:
- New MCP Servers settings section with full CRUD operations
- Import/Export JSON configs (Claude Code format compatible)
- Configurable permission settings:
- Auto-approve MCP tools (bypass permission prompts)
- Unrestricted tools (allow all tools when MCP enabled)
- Refresh button to reload from settings file
Implementation:
- Added MCPServerConfig and MCPToolInfo types
- Added store actions for MCP server management
- Updated claude-provider to use configurable MCP permissions
- Updated sdk-options factory functions for MCP support
- Added settings helpers for loading MCP configs
Removed all debug console.log statements from agent-service.ts to avoid
polluting production logs. This addresses code review feedback from
gemini-code-assist.
Removed debug logs for:
- sendMessage() entry and session state
- Event emissions (started, message, stream, complete)
- Provider execution
- SDK session ID capture
- Tool use detection
- Queue processing
- emitAgentEvent() calls
Kept console.error logs for actual errors (session not found, execution
errors, etc.) as they are useful for troubleshooting.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
The sandbox configuration was set in createChatOptions() and
createAutoModeOptions(), but was never passed to the ClaudeProvider.
This caused the sandbox to never actually be enabled.
Changes:
- Add sandbox field to ExecuteOptions interface
- Pass sandbox config from AgentService to provider
- Pass sandbox config from AutoModeService to provider
- Forward sandbox config in ClaudeProvider to SDK options
Now the sandbox configuration from settings is properly used.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Add a global setting to enable/disable sandbox mode for Claude Agent SDK.
This allows users to control sandbox behavior based on their authentication
setup and system compatibility.
Changes:
- Add enableSandboxMode to GlobalSettings (default: true)
- Add sandbox mode checkbox in Claude settings UI
- Wire up setting through app store and settings service
- Update createChatOptions and createAutoModeOptions to use setting
- Add getEnableSandboxModeSetting helper function
- Remove hardcoded sandbox configuration from ClaudeProvider
- Add detailed logging throughout agent execution flow
The sandbox mode requires API key or OAuth token authentication. Users
experiencing issues with CLI-only auth can disable it in settings.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
- Introduced a new utility function to filter out CLAUDE.md from context files when autoLoadClaudeMd is enabled, enhancing clarity and preventing duplication.
- Updated AgentService and AutoModeService to utilize the new filtering function, streamlining context file management.
- Improved documentation for the new utility, detailing its purpose and usage in context file handling.
- Updated both services to conditionally load context files while excluding CLAUDE.md when autoLoadClaudeMd is enabled, preventing duplication.
- Improved the structure and clarity of the context files prompt, emphasizing the importance of following project-specific rules and conventions.
- Ensured consistent handling of context file loading across different methods in both services.
- Updated both services to conditionally load CLAUDE.md based on the autoLoadClaudeMd setting, preventing duplication.
- Improved clarity in comments regarding the loading process of context files.
- Ensured consistent retrieval of the autoLoadClaudeMd setting across different methods.
- Removed the private method for getting the autoLoadClaudeMd setting from AgentService and AutoModeService.
- Updated both services to utilize the new settings helper for retrieving the autoLoadClaudeMd setting, improving code reusability and clarity.
- Adjusted error handling in the settings helper to throw errors instead of returning false when the settings service is unavailable.
- Updated API routes to accept an optional settings service for loading the autoLoadClaudeMd setting.
- Introduced a new settings helper utility for retrieving project-specific settings.
- Enhanced feature generation and spec generation processes to utilize the autoLoadClaudeMd setting.
- Refactored relevant route handlers to support the new settings integration across various endpoints.
- Introduced a new setting to enable automatic loading of CLAUDE.md files from project-specific directories.
- Updated relevant services and components to support the new setting, including the AgentService and AutoModeService.
- Added UI controls for managing the auto-load setting in the settings view.
- Enhanced SDK options to incorporate settingSources for CLAUDE.md loading.
- Updated global and project settings interfaces to include autoLoadClaudeMd property.
- Introduced a new utility function `loadContextFiles` to load project-specific context files from the `.automaker/context/` directory, enhancing agent prompts with project rules and guidelines.
- Updated `AgentService` and `AutoModeService` to utilize the new context loading functionality, combining context prompts with existing system prompts for improved agent performance.
- Added comprehensive documentation on the context files system, including usage examples and metadata structure, to facilitate better understanding and implementation.
- Removed redundant context loading logic from `AutoModeService`, streamlining the codebase.
These changes aim to improve the agent's contextual awareness and adherence to project-specific conventions.
This commit updates various modules to utilize the secure file system operations from the secureFs module instead of the native fs module. Key changes include:
- Replaced fs imports with secureFs in multiple route handlers and services to enhance security and consistency in file operations.
- Added centralized validation for working directories in the sdk-options module to ensure all AI model invocations are secure.
These changes aim to improve the security and maintainability of file handling across the application.
- Wrapped the terminal settings retrieval in a try-catch block to handle potential errors and respond with a 500 status and error details.
- Updated path normalization logic to skip resolution for WSL UNC paths, preventing potential issues with path handling in Windows Subsystem for Linux.
- Enhanced unit tests for session termination to include timer-based assertions for graceful session killing.
Platform-specific implementations:
- macOS: Uses 'expect' command (unchanged, working)
- Windows: Uses node-pty for PTY support
Also fixes 'which' vs 'where' for checking Claude CLI availability.
- Changed the token generation method to use slice instead of substr for better readability.
- Enhanced maxSessions validation in the settings update handler to check for undefined values and ensure the input is a number before processing.
- Updated the maximum terminal session limit from 500 to 1000 to accommodate more concurrent sessions.
- Enhanced path handling in the editor and HTTP API client to normalize file paths for both Unix and Windows systems, ensuring consistent URL encoding.
Replace Unix-only 'expect' command with node-pty library which works
on Windows, macOS, and Linux. Also fixes 'which' command to use 'where'
on Windows for checking if Claude CLI is available.
- Added path validation in resolveWorkingDirectory to reject paths with null bytes and normalize paths.
- Improved killSession method to attempt graceful termination with SIGTERM before falling back to SIGKILL after a delay.
- Enhanced logging for session termination to provide clearer feedback on the process.
Adds a Claude usage tracking feature that displays session, weekly, and Sonnet usage stats. Uses the Claude CLI's /usage command to fetch data (no API key required).
Features:
- Usage popover in board header showing session, weekly, and Sonnet limits
- Progress bars with color-coded status (green/orange/red)
- Auto-refresh with configurable interval
- Caching of usage data with stale indicator
- Settings section for refresh interval configuration
Server:
- ClaudeUsageService: Executes Claude CLI via PTY (expect) to fetch usage
- New /api/claude/usage endpoint
UI:
- ClaudeUsagePopover component with usage cards
- ClaudeUsageSection in settings for configuration
- Integration with app store for persistence
- Refactored session limit checks in terminal settings to use constants for minimum and maximum session values.
- Enhanced terminal layout saving mechanism with debouncing to prevent excessive writes during rapid changes.
- Updated error messages to reflect new session limit constants.
- Added new endpoints for terminal settings: GET and PUT /settings to retrieve and update terminal configurations.
- Implemented session limit checks during session creation, returning a 429 status when the limit is reached.
- Introduced a new TerminalSection in settings view for customizing terminal appearance and behavior, including font family, default font size, line height, and screen reader mode.
- Added support for new terminal features such as search functionality and improved error handling with a TerminalErrorBoundary component.
- Updated terminal layout persistence to include session IDs for reconnection and enhanced terminal state management.
- Introduced new keyboard shortcuts for terminal actions, including creating new terminal tabs.
- Enhanced UI with scrollbar theming for terminal components.
This commit introduces several improvements to the security and file handling mechanisms across the application. Key changes include:
- Updated the Dockerfile to pin the GitHub CLI version for reproducible builds.
- Refactored the secure file system operations to ensure consistent path validation and type handling.
- Removed legacy path management functions and streamlined the allowed paths logic in the security module.
- Enhanced route handlers to validate path parameters against the ALLOWED_ROOT_DIRECTORY, improving security against unauthorized access.
- Updated the settings service to focus solely on the Anthropic API key, removing references to Google and OpenAI keys.
These changes aim to enhance security, maintainability, and clarity in the codebase.
Tests: All unit tests passing.
Changes:
- Removed specific compiled file patterns from .gitignore to simplify ignore rules.
- Modified error handling in feature-loader.ts to rethrow errors instead of keeping original paths, preventing potential broken references.
- Added ".js" extensions to import statements in types package for ESM compliance.
Benefits:
✅ Cleaner .gitignore for better maintainability
✅ Improved error handling logic in feature-loader
✅ Consistent import paths for ESM compatibility
All tests passing.
