- Introduced a helper function to escape strings for AppleScript to ensure proper command execution.
- Updated terminal command definitions to utilize the new escape function, improving compatibility with special characters.
- Adjusted command parameters to use double quotes for directory paths, enhancing consistency and reliability.
- Introduced a new IPC channel for quitting the application.
- Updated ipc.ts to handle the App_Quit channel, allowing the app to quit when invoked.
- Added corresponding quit method in the preload API for client-side access.
- Fixed a minor URL check in WindowService to ensure proper navigation handling.
- Add MCP server configuration UI for agent settings
- Update agent and session forms to include MCP server selection
- Fix MCP API service logging and tools handling
- Add Chinese localization for MCP settings
- Update type definitions to support MCP server arrays
This enables agents to use MCP (Model Control Protocol) servers
as additional tools and capabilities in their execution context.
* fix(assistant): update translate assistant content handling for QwenMT model
- Adjusted content assignment in getDefaultTranslateAssistant to use store settings only when the model is not a QwenMT model, ensuring correct translation behavior.
* lint err
* refactor(assistant): encapsulate content handling logic for translation
- Introduced a new function, getTranslateContent, to streamline content assignment in getDefaultTranslateAssistant.
- This change improves readability and maintains correct translation behavior for QwenMT models.
* format code
- Log each content block when processing user or assistant messages
- Maintain existing switch case logic for text block handling
- Improve debugging visibility for multi-block message processing
- Correct toolName lookup to use tool_use_id instead of blockKey in tool-result chunks
- Add debug logging for stream event handling
- Update contentBlockState key to use event.content_block.id for tool_use events
- Correctly reference toolName from contentBlockState using blockKey instead of block.tool_use_id
- Ensure proper tool result chunk generation when handling assistant messages
- Maintain consistent data structure for tool call processing
- Populate toolName in tool-result chunks from contentBlockState
- Add onSessionCreated callback to SessionModal for post-creation actions
- Return created session from useSessions hook and update SWR cache optimistically
- Introduced a new method in ConfigManager to generate and retrieve a unique client ID.
- Updated AppUpdater to include the client ID in the request headers alongside the user agent.
- Rename built_in_tools field to tools for consistency
- Add type field to Tool schema (builtin/mcp/custom)
- Consolidate tool handling in BaseService with listMcpTools method
- Remove unused CreateSessionResponse and related schemas
- Clean up unused imports and dead code in session handlers
- Unify agent and session tool resolution logic
- Add shared Anthropic utilities package with OAuth and API key client creation
- Implement provider-specific message routing alongside existing v1 API
- Enhance authentication middleware with priority handling (API key > Bearer token)
- Add comprehensive auth middleware test suite with timing attack protection
- Update session handling and message transformation for Claude Code integration
- Improve error handling and validation across message processing pipeline
- Standardize import formatting and code structure across affected modules
This establishes the foundation for Claude Code OAuth authentication while maintaining
backward compatibility with existing API key authentication methods.
* Refactor agent streaming from EventEmitter to ReadableStream
Replaced EventEmitter-based agent streaming with ReadableStream for
better compatibility with AI SDK patterns. Modified
SessionMessageService to return stream/completion pair instead of event
emitter, updated HTTP handlers to use stream pumping, and added IPC
contract for renderer-side message persistence.
* Add accessible paths management to agent configuration
Move accessible paths functionality from session modal to agent modal,
add validation requiring at least one path, and update form handling to
inherit agent paths in sessions.
* Add provider_name field to model objects and improve display
- Add provider_name field to ApiModel schema and transformation logic
- Update model options to include providerName for better display
- Improve provider label fallback chain in model transformation
- Fix agent hook to use proper SWR key and conditional fetching
- Enhance option rendering with better truncation and provider display
* fix(i18n): Auto update translations for PR #10276
* Optimize chat components with memoization and shared layout
- Wrap `SessionMessages` and `SessionInputBar` in `useMemo` to prevent unnecessary re-renders
- Refactor `AgentSessionMessages` to use shared layout components and message grouping
- Extract common styled components to `shared.tsx` for reuse across message components
* Add smooth animations to SessionsTab and Sessions components
- Replace static conditional rendering with Framer Motion animations for no-agent and session states
- Animate session list items with staggered entrance and exit transitions
- Add loading spinner animation with fade effect
- Apply motion to session creation button with delayed entrance
* Add loading state with spinner and i18n support to SessionsTab
- Replace static "No active agent" message with a spinner and loading text
- Integrate react-i18next for translation of loading message
- Adjust animation timing and styling for smoother loading state transition
* Support API models with provider_name field in getModelName
- Add ApiModel type import and update function signature to accept ApiModel
- Return formatted name using provider_name field for API models
- Maintain backward compatibility for legacy models by looking up provider in store
* Simplify provider display name logic and add debug logging
- Replace complex fallback chain for provider display name with direct provider name access
- Add console.log for model debugging in getModelName function
* Extract model name from session model string
- Use split and pop to isolate the model name after the colon
- Fall back to the full model string if no colon is present
- Maintain provider and group identifiers for model object consistency
* Improve model name resolution for agent sessions
- Extract actual model ID from session model string and resolve model details
- Use resolved model name, provider, and group when available instead of defaults
- Remove redundant API model handling in getModelName function
* Set default active agent and session on load
- Automatically select first agent if none active after loading
- Automatically select first session per agent if none active after loading
- Prevent empty selection states in UI components
---------
Co-authored-by: GitHub Action <action@github.com>
- Remove unused persistence tracking variables in message handler
- Simplify finalizeResponse logic by removing unnecessary checks
- Change 'finish' event type to 'complete' for consistency
- Add debug logging for streaming events
- Clean up dead code and improve readability
- Extract getServersFromRedux to shared utility getMCPServersFromRedux
- Implement 5-minute TTL cache for MCP servers and providers
- Reduce redundant Redux store queries in API server
- Improve response times for frequently accessed data
Introduce UpdateSessionResponse type and schema to support session updates. Implement update session methods across client, service, and handler layers to enable session modifications.
Align frontend and backend types for agents list response. The API now returns paginated data with limit/offset and renamed 'agents' field to 'data' for consistency. Update related type definitions and usage across the codebase.
Introduce CreateSessionResponse type and schema to clearly define the return type of session creation operations. This improves type safety and consistency across the codebase when handling session responses.
* feat(CodeTools): add support for terminal selection on macOS
- Introduced terminal selection functionality in CodeTools, allowing users to choose from available terminal applications.
- Implemented caching for terminal availability checks to enhance performance.
- Updated CodeToolsService to preload available terminals and check their availability.
- Enhanced UI in CodeToolsPage to display terminal options and handle user selection.
- Added new IPC channel for retrieving available terminals from the main process.
* lint errs
* format
* support wezterm
* support terminal
* support ghostty
* support warp kitty
* fix github scanner issues
* fix all github issues
* support windows
* support windows
* suppport hyper
* Refactor terminal command execution for macOS applications to use shell scripts instead of AppleScript, improving compatibility and performance.
* Remove Hyper terminal configuration from shared constants
* update lint
* fix(i18n): Auto update translations for PR #10192
* fix platform checking
* format
* feat: add Tabby terminal configuration for macOS
* fix wrap terminal
* delete warp
---------
Co-authored-by: GitHub Action <action@github.com>
- Replace process spawning with @anthropic-ai/claude-code SDK query function
- Remove complex process management, stdout/stderr parsing, and JSON buffering
- Directly iterate over typed SDKMessages from AsyncGenerator
- Simplify error handling and completion logic
- Maintain full compatibility with existing SessionMessageService interface
- Eliminate ~130 lines of process management code
- Improve reliability by removing JSON parsing edge cases
* feat: Add automatic database migration system for agents service
- Add migrations tracking schema with version, tag, and timestamp
- Implement MigrationService to automatically run pending migrations
- Integrate migration check into BaseService initialization
- Read migration files from drizzle/ directory and journal.json
- Track applied migrations to prevent re-execution
- Ensure database is always at latest version on service startup
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
* refactor: Improve migration logging and enhance database path configuration
* chore: harden migration bootstrap flow
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Happy <yesreply@happy.engineering>
- Added `aisdk-stream-protocel.md` to document text and data stream protocols.
- Created `ClaudeCodeService` for invoking and streaming responses from the Claude Code CLI.
- Introduced built-in tools for Claude Code, including Bash, Edit, and WebFetch.
- Developed transformation functions to convert Claude Code messages to AI SDK format.
- Enhanced OCR utility with delayed loading of the Sharp module.
- Updated agent types and session message structures to accommodate new features.
- Modified API tests to reflect changes in session creation and message streaming.
- Upgraded `uuid` package to version 13.0.0 for improved UUID generation.
- Install electron-reload package for automatic app reloading during development
- Configure electron-reload in main process with development-only activation
- Enable automatic restart when source files change during yarn dev
- Use hardResetMethod: 'exit' for clean app restarts
