♻️ refactor: migrate AI Rules to Claude Code Skills system
Migrate all AI Rules from .cursor/rules/ to .agents/skills/ directory:
- Move 23 skills to .agents/skills/ (main directory)
- Update symlinks: .claude/skills, .cursor/skills, .codex/skills
- Create project-overview skill from project documentation
- Add references/ subdirectories for complex skills
- Remove LobeChat references from skill descriptions
- Delete obsolete .cursor/rules/ and .claude/commands/prompts/ directories
Skills structure enables better portability and maintainability across AI tools.
* ♻️ refactor: migrate SkillStore and IntegrationDetailModal to imperative API
- Refactor SkillStore to use createModal imperative API instead of declarative Modal
- Refactor IntegrationDetailModal to use createModal with IntegrationDetailContent
- Remove open/setOpen state management from all calling components
- Add modal-imperative.mdc rule for modal best practices
- Reduce code complexity and improve maintainability
* 🐛 fix: keep modal open during OAuth flow until connection completes
Close modal only after isConnected becomes true, not immediately after
handleConnect returns. This ensures useSkillConnect listeners stay alive
to detect OAuth completion via postMessage/polling.
* 🔧 chore: update dependencies and refactor markdown handling
- Updated "@lobehub/ui" to version "^4.27.4" in package.json.
- Replaced "markdown-to-txt" with a local utility "markdownToTxt" for converting markdown to plain text across multiple components.
- Refactored imports in various files to utilize the new markdownToTxt utility, improving code consistency and maintainability.
Signed-off-by: Innei <tukon479@gmail.com>
---------
Signed-off-by: Innei <tukon479@gmail.com>
Consolidate skill files into a unified .agents directory structure.
- Move skills from .cursor/skills/ to .agents/
- Create symlinks in .codex/skills and .cursor/skills pointing to .agents
* refactor: reduce unused code
Signed-off-by: Innei <tukon479@gmail.com>
* 🐛 fix(desktop): prevent duplicate CORS headers in response
Only add CORS headers if they don't already exist in the server response.
This fixes issues with CDN resources (like cdn.jsdelivr.net) that already
return CORS headers, causing "multiple values" errors.
Fixes LOBE-2765
* 🔧 refactor(desktop): remove IpcServerMethod decorator and related metadata
This update simplifies the IPC method handling by removing the IpcServerMethod decorator and its associated metadata management. The changes include updates to documentation and code references, ensuring a cleaner and more maintainable IPC implementation.
No functional changes were introduced, but the codebase is now more streamlined for future development.
Signed-off-by: Innei <tukon479@gmail.com>
* ✨ feat(desktop): introduce HTTP headers utility functions
Added a new utility module for managing HTTP response headers in Electron, addressing case sensitivity issues. This includes functions to set, get, check existence, and delete headers. Updated the Browser class to utilize these utilities for setting CORS headers, ensuring no duplicates are present.
This enhancement improves code maintainability and simplifies header management in the application.
Signed-off-by: Innei <tukon479@gmail.com>
---------
Signed-off-by: Innei <tukon479@gmail.com>
chore: update package dependencies and import paths
* Updated @lobehub/ui from version 3.4.2 to 3.4.4 in package.json.
* Adjusted import paths for components from '@lobehub/ui/es/' to include '.mjs' extension where necessary.
* Refactored imports for motion library to use 'motion/react-m' instead of 'motion/react'.
* Cleaned up import statements in various components for consistency.
Signed-off-by: Innei <tukon479@gmail.com>
* refactor: client ipc
* refactor: server ipc
refactor: update IPC method names for consistency
Signed-off-by: Innei <tukon479@gmail.com>
fix: cast IPC return type to DesktopIpcServices for type safety
Signed-off-by: Innei <tukon479@gmail.com>
chore: add new workspace for desktop application in package.json
Signed-off-by: Innei <tukon479@gmail.com>
fix: export FileMetadata interface for improved accessibility
Signed-off-by: Innei <tukon479@gmail.com>
refactor: unify IPC mocking across test files for consistency
Signed-off-by: Innei <tukon479@gmail.com>
feat: enhance type-safe IPC flow with context propagation and service registry
- Introduced `getIpcContext()` and `runWithIpcContext()` for improved context management in IPC handlers.
- Updated `BrowserWindowsCtr` methods to utilize the new context handling.
- Added `McpInstallCtr` to the IPC constructors registry.
- Enhanced README with details on the new type-safe IPC features.
Signed-off-by: Innei <tukon479@gmail.com>
refactor: enhance IPC method registration for improved type safety
- Updated `registerMethod` in `IpcHandler` and `IpcService` to accept variable argument types, enhancing flexibility in method signatures.
- Simplified the `ExtractMethodSignature` type to support multiple arguments.
Signed-off-by: Innei <tukon479@gmail.com>
chore: add global type definitions and refactor import statements
- Introduced a new global type definition file to support Vite client imports.
- Refactored import statements in `App.ts` and `App.test.ts` to remove unnecessary type casting for `import.meta.glob`, improving code clarity.
Signed-off-by: Innei <tukon479@gmail.com>
* refactor: make groupName in BrowserWindowsCtr readonly for better encapsulation
Signed-off-by: Innei <tukon479@gmail.com>
* refactor: update IPC method registration and usage for improved type safety and consistency
- Replaced `@ipcClientEvent` with `@IpcMethod()` in various controllers to standardize IPC method definitions.
- Enhanced the usage of `ensureElectronIpc()` for type-safe IPC calls in service layers.
- Updated `BrowserWindowsCtr` and `NotificationCtr` to utilize the new IPC method structure, improving encapsulation and clarity.
- Refactored service methods to eliminate manual string concatenation for IPC event names, ensuring better maintainability.
Signed-off-by: Innei <tukon479@gmail.com>
---------
Signed-off-by: Innei <tukon479@gmail.com>
* add
* baseline
* ✅ test(store): add tests for discover store plugin and mcp slices
- Add comprehensive tests for discover/slices/plugin/action.ts (15 tests)
- Add comprehensive tests for discover/slices/mcp/action.ts (11 tests)
- Update test-coverage.md with new metrics and completed work
- Coverage: 74.24% overall (+26 tests, 2 new test files)
- Action files coverage: 29/40 tested (72.5%, +2 files)
Features tested:
- Plugin/MCP categories, detail, identifiers, and list fetching
- SWR key generation with locale and parameters
- SWR configuration verification
- Service integration with discoverService
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* 📝 docs(testing): add SWR hooks testing guide and subagent workflow
Testing Guide Updates:
- Add comprehensive SWR hooks testing section with examples
- Document key differences from regular action tests
- Add examples for testing SWR key generation and configuration
- Add examples for testing conditional fetching
- Update references to include SWR hook test examples
Test Coverage Guide Updates:
- Add detailed subagent workflow for parallel testing
- Document when and how to use subagents for testing
- Add complete workflow example using subagents
- Add benefits and best practices for subagent usage
- Clarify that subagents should NOT commit or update docs
- Add step-by-step guide for launching parallel subagents
Key improvements:
- Better documentation for testing SWR-based store actions
- Clear workflow for efficient parallel testing using subagents
- Single atomic commit strategy after all subagents complete
- Improved testing efficiency and organization
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* ♻️ refactor(test): fix SWR mock strategy to properly test fetcher
Previously, tests were hardcoding return values instead of calling
the fetcher function. This bypassed the actual service call logic.
Changes:
- Fix useSWR mock to call fetcher and return its Promise
- Update assertions to await Promise results
- Update testing guide with correct mock pattern
- Add explanation of why this approach is correct
Before (incorrect):
```typescript
useSWRMock.mockImplementation(((key, fetcher) => {
fetcher?.(); // Call but ignore result
return { data: mockData }; // Hardcoded
}) as any);
expect(result.current.data).toEqual(mockData);
```
After (correct):
```typescript
useSWRMock.mockImplementation(((key, fetcher) => {
const data = fetcher?.(); // Get Promise from fetcher
return { data }; // Return Promise
}) as any);
const resolvedData = await result.current.data;
expect(resolvedData).toEqual(mockData);
```
Benefits:
- ✅ Actually tests the fetcher function
- ✅ Mirrors real SWR behavior (data is Promise)
- ✅ Service calls are properly verified
- ✅ Tests are more accurate and maintainable
Updated files:
- .cursor/rules/testing-guide/zustand-store-action-test.mdc
- src/store/discover/slices/plugin/action.test.ts
- src/store/discover/slices/mcp/action.test.ts
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* 🐛 fix(test): correct SWR mock strategy to match project standards
- Remove useSWR mocking, use real SWR implementation instead
- Only mock service methods (fetchers) with vi.spyOn
- Use waitFor for async assertions
- Update testing guide with correct SWR pattern
- Add reference to src/store/chat/slices/message/action.test.ts
This fixes the incorrect mocking approach from previous commits.
All 13 tests pass with the corrected strategy.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* ✅ test(store): add comprehensive tests for high priority action files
- Add mcpStore action tests (41 tests, 624 LOC covered)
- MCP plugin installation flow (normal, resume, dependencies, config)
- Connection testing (HTTP and STDIO)
- Plugin lifecycle management
- Error handling and cancellation flows
- Add fileManager action tests (35 tests, 205 LOC covered)
- File upload and processing workflows
- Chunk embedding and parsing
- File list management and refresh
- SWR data fetching
Testing approach:
- Used parallel subagents for efficient development
- Followed zustand testing patterns from guide
- Proper test layering and per-test mocking
- All tests pass type-check and lint
Coverage improvement: 74.24% → ~76% (+76 tests, 2 files)
Action files: 29/40 → 31/40 tested (77.5%)
🏆 Milestone: All high priority files (>200 LOC) now have tests!
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* ✅ test(store): complete 100% action file coverage with 160 new tests
Added comprehensive tests for all remaining 9 medium-priority action files:
Discovery Store (33 tests):
- assistant/action.ts: 10 tests (SWR hooks, categories, detail, identifiers, list)
- provider/action.ts: 11 tests (SWR hooks, detail with readme, identifiers, list with filters)
- model/action.ts: 12 tests (SWR hooks, categories, detail, identifiers, list with params)
Knowledge Base Store (29 tests):
- crud/action.ts: 19 tests (create, update, remove, refresh, loading states, SWR hooks)
- content/action.ts: 10 tests (add files, remove files, error handling)
File Store (36 tests):
- upload/action.ts: 18 tests (base64 upload, file upload with progress, type detection, KB integration)
- chunk/action.ts: 18 tests (drawer management, highlight, semantic search)
AI Infrastructure Store (23 tests):
- aiModel/action.ts: 23 tests (CRUD, batch operations, remote sync, toggle enabled, SWR hooks)
Chat Store (39 tests):
- thread/action.ts: 39 tests (CRUD, messaging, AI title generation, validation, loading states)
Testing approach:
- Used 9 parallel subagents for efficient development
- Followed zustand testing patterns from guide
- SWR hook testing for discovery slices
- Complex async flows with proper error handling
- File operations with progress callbacks
- Semantic search and RAG integration
Coverage improvement: ~76% → ~80% (+160 tests, 9 files)
Action files: 31/40 → 40/40 tested (100%)
🎉 MILESTONE: All 40 action files now have comprehensive test coverage!
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* fix test
* fix test
* fix context-engine
* add tests
* remove
* remove tools bar
* pin bun version
* fix
---------
Co-authored-by: Claude <noreply@anthropic.com>
* docs: update fal provider invalid image links
* docs: add FAL model provider environment variables documentation
* 🐛 fix: update model type assignment in parseModels.ts to use dynamic lookup
* 📝 docs: add FAQ for resolving AI image generation timeout issues on Vercel
* ✨ feat: implement getModelPropertyWithFallback utility for dynamic model property retrieval
* 📝 docs: expand testing guide with best practices for mock data strategies, error handling, and module pollution prevention
* 🐛 fix: update model type in LobeOpenAICompatibleFactory tests to 'chat'
