* feat: add optional mode field to slash command front matter
- Add mode field to Command interface
- Update command parsing to extract mode from frontmatter
- Modify RunSlashCommandTool to automatically switch mode when specified
- Add comprehensive tests for mode field parsing and switching
- Update existing tests to include mode field
* Make it work for manual slash commands too
---------
Co-authored-by: Roo Code <roomote@roocode.com>
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
* Add support for skills
* fix: use type-only import for ClineProvider and relative paths in skills section
---------
Co-authored-by: Roo Code <roomote@roocode.com>
Adds userAgentAppId configuration to the BedrockRuntimeClient in the
code indexing embedder, matching the implementation pattern already
used in the main Bedrock API provider.
This enables proper user agent identification in CloudTrail AWS requests
when using Bedrock for code indexing embeddings.
Fixes#10165
Co-authored-by: Roo Code <roomote@roocode.com>
* Add a way to save screenshots from the browser tool
* fix: use cross-platform paths in BrowserSession screenshot tests
* fix: validate screenshot paths to prevent filesystem escape
---------
Co-authored-by: Roo Code <roomote@roocode.com>
* feat: add AWS Bedrock support for codebase indexing
- Add bedrock as a new EmbedderProvider type
- Add AWS Bedrock embedding model profiles (titan-embed-text models)
- Create BedrockEmbedder class with support for Titan and Cohere models
- Add Bedrock configuration support to config manager and interfaces
- Update service factory to create BedrockEmbedder instances
- Add comprehensive tests for BedrockEmbedder
- Add localization strings for Bedrock support
Closes#8658
* fix: add missing bedrockOptions to loadConfiguration return type
* Fix various issues that the original PR missed.
* Remove debug logs
* Rename AWS Bedrock -> Amazon Bedrock
* Remove some 'as any's
* Revert README changes
* Add translations
* More translations
* Remove leftover code from a debugging session.
* fix: add bedrock to codebaseIndexModelsSchema and update brace-expansion override
- Add bedrock provider to codebaseIndexModelsSchema type definition to fix empty model dropdown in UI
- Update pnpm override for brace-expansion from '>=2.0.2' to '^2.0.2' to resolve ESM/CommonJS compatibility issues
* Improvements to AWS Bedrock embeddings support
- Enhanced bedrock.ts embedder implementation
- Added comprehensive test coverage in bedrock.spec.ts
- Updated config-manager.ts for better Bedrock configuration handling
- Improved service-factory.ts integration
- Updated embeddingModels.ts with Bedrock models
- Enhanced CodeIndexPopover.tsx UI for Bedrock options
- Added auto-populate test for CodeIndexPopover
- Updated pnpm-lock.yaml dependencies
* Restore openrouter config
* Remove debug log
* Fix config-manager.spec.ts unit test.
* Add translations for "optional"
* Revert unnecessary change related to open ia embedder
---------
Co-authored-by: Roo Code <roomote@roocode.com>
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
Co-authored-by: Smartsheet-JB-Brown <jb.brown@smartsheet.com>
* feat: add OpenRouter embedding provider support
Implement comprehensive OpenRouter embedding provider support for codebase indexing with the following features:
- New OpenRouterEmbedder class with full API compatibility
- Support for OpenRouter's OpenAI-compatible embedding endpoint
- Rate limiting and retry logic with exponential backoff
- Base64 embedding handling to bypass OpenAI package limitations
- Global rate limit state management across embedder instances
- Configuration updates for API key storage and provider selection
- UI integration for OpenRouter provider settings
- Comprehensive test suite with mocking
- Model dimension support for OpenRouter's embedding models
This adds OpenRouter as the 7th supported embedding provider alongside OpenAI, Ollama, OpenAI-compatible, Gemini, Mistral, and Vercel AI Gateway.
* Add translation key
* Fix mutex double release bug
* Add translations
* Add more translations
* Fix failing tests
* code-index(openrouter): fix HTTP-Referer header to RooCodeInc/Roo-Code; i18n: add and wire OpenRouter Code Index strings; test: assert default headers in embedder
---------
Co-authored-by: daniel-lxs <ricciodaniel98@gmail.com>
* feat: improve @ file search for large projects
- Increase default file limit from 5,000 to 10,000 (configurable up to 500,000)
- Respect VSCode search settings (useIgnoreFiles, useGlobalIgnoreFiles, useParentIgnoreFiles)
- Add 'maximumIndexedFilesForFileSearch' configuration setting
- Add tests for new functionality
Conservative default of 10k keeps memory usage low while still providing 2x
improvement. Users with large projects can opt-in to higher limits (up to 500k).
This is a simplified alternative to PR #5723 that solves the same problem
without the complexity of caching. Ripgrep is already fast enough for
10k+ files, and the benefit of caching doesn't justify 2,200+ lines of
additional code and maintenance burden.
Fixes#5721
* fix: add missing translations for maximumIndexedFilesForFileSearch setting
* test: improve file-search tests to verify configuration behavior
* fix: prevent MCP server restart when toggling tool permissions
Add isProgrammaticUpdate flag to distinguish between programmatic config
updates and user-initiated file changes. Skip file watcher processing
during programmatic updates to prevent unnecessary server restarts.
* fix(mcp): prevent server reconnection when toggling disabled state
Fixed bug where MCP servers would reconnect instead of staying disabled when toggled off. The issue was that toggleServerDisabled() used stale in-memory config instead of reading the fresh config from disk after writing the disabled flag.
Changes:
Added readServerConfigFromFile() helper to read and validate server config from disk
Updated disable path to read fresh config before calling connectToServer()
Updated enable path to read fresh config before calling connectToServer()
This ensures the disabled: true flag is properly read, causing connectToServer() to create a disabled placeholder connection instead of actually connecting the server.
+ refactor(mcp): use safeWriteJson for atomic config writes
Replace JSON.stringify + fs.writeFile with safeWriteJson in McpHub.ts
to prevent data corruption through atomic writes with file locking.
* fix(mcp): prevent race condition in isProgrammaticUpdate flag
Replace multiple independent reset timers with a single timer that gets
cleared and rescheduled on each programmatic config update. This prevents
the flag from being reset prematurely when multiple rapid updates occur,
which could cause unwanted server restarts during the file watcher's
debounce period.
+ fix(mcp): ensure isProgrammaticUpdate flag cleanup with try-finally
Wrap safeWriteJson() calls in try-finally blocks to guarantee the
isProgrammaticUpdate flag is always reset, even if the write operation
fails. This prevents the flag from being stuck at true indefinitely,
which would cause subsequent user-initiated config changes to be
silently ignored.
* feat(checkpoints): create checkpoint on user message send
* fix(checkpoints): suppress implicit user-message checkpoint row; keep current checkpoint updated without a chat row
* Fix checkpoint suppression for user messages
- Propagate suppressMessage flag through event chain properly
- Update ChatView to check checkpoint metadata for suppressMessage flag
- Ensure checkpoint messages are created but not rendered when suppressed
- Fix bug where checkpointSave(false) should have been checkpointSave(true)
* fix: only create checkpoint on user message when files have changed
- Changed allowEmpty from true to false in checkpointSave call
- Checkpoints will now only be created when there are actual file changes
- This avoids creating empty commits in the shadow git repository
* test: update checkpoint test to include suppressMessage parameter
- Fixed test expectation to match the new function signature
- saveCheckpoint now expects both allowEmpty and suppressMessage parameters
---------
Co-authored-by: Roo Code <roomote@roocode.com>
Co-authored-by: Hannes Rudolph <hrudolph@gmail.com>
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
* feat: update /init command to focus only on non-obvious discoveries
- Emphasize that only non-obvious information discovered by reading files should be included
- Update all mode-specific examples to show truly non-obvious rules
- Add CRITICAL section explaining what qualifies as non-obvious
- Exclude standard patterns and framework defaults
- Focus on gotchas, hidden requirements, and counterintuitive patterns
* fix: ensure /init command checks for and updates existing AGENTS.md files
- Check for existing AGENTS.md in root directory
- Check for existing AGENTS.md in all .roo/rules-*/ directories
- Read and improve existing files rather than replacing them
- Update discovery phase to explicitly look for these files
* refactor: clean up /init command to properly handle existing AGENTS.md files
- Consolidate all AGENTS.md path checking into step 1 with CRITICAL emphasis
- Add explicit instructions for deep dive improvement of existing files
- Remove redundant mentions throughout the workflow
- Emphasize iterating and enhancing existing files, not just appending
- Add clear guidance on cleaning up outdated info and reorganizing
- Make discovery phase the single source of truth for file checking
* fix: strengthen /init command to aggressively remove obvious information
- Add explicit instructions to DELETE obvious information first
- Emphasize that files should get SHORTER, not longer
- Use stronger language: CRITICALLY EVALUATE, AGGRESSIVELY DELETE
- Make clear that standard practices must be removed even if previously included
- Add success metric: files should be more concise and valuable
- Goal is to fix the tendency to just append instead of clean up
* fix: ensure .roo directory is created in project root, not system root
CRITICAL FIX: Paths were being interpreted as absolute, causing .roo to be created at system root
- Clarify all paths are relative to PROJECT/WORKSPACE root
- Add explicit warnings that .roo must be in project root, not system root
- Update example to show proper project structure
- Add IMPORTANT notes throughout emphasizing relative paths
- Fix prevents accidental creation of /.roo at system level
* test: update built-in-commands test to match new init command content
- Replace 'mode-specific rule directories' with 'non-obvious'
- Replace 'analysis_workflow' with 'discovered by reading files'
- Tests now align with the new focus on non-obvious discoveries
Fixes#6756
When ripgrep encounters file access errors (e.g., permission denied), it would cause the search_files tool to return empty results. Adding the --no-messages flag suppresses these error messages while still showing pattern syntax errors, allowing the search to continue and return valid results from accessible files.
Co-authored-by: Roo Code <roomote@roocode.com>
* Revert "fix: prevent unnecessary MCP server refresh on settings save (#6772) (#6779)"
This reverts commit 8d05bc179b.
* fix(mcp): Revert changes causing startup issues and temporarily disable notifications
- Reverted PR #6779 which prevented unnecessary MCP server refreshes but caused startup failures
- Temporarily disabled MCP notification popups as a stopgap solution
- Added TODO comments explaining the temporary nature of disabled notifications
- This allows MCP servers to function properly while a more robust solution is developed
* test(mcp): restore mcpEnabled toggle coverage to verify delegation to McpHub
* refactor(mcp): remove info notifications during refresh; rely on UI indicator
* fix: add missing MCP error translation keys
- Added refresh_after_disable translation key
- Added refresh_after_enable translation key
- Added disconnect_servers_partial translation key
- Updated all 18 locale files with the missing keys
- Fixes the "errors.refresh_after_disable" display issue in MCP server management
* fix: remove fallback English strings from MCP error translations
As requested by @mrubens, removed the || operator and fallback English strings from:
- t("mcp:errors.disconnect_servers_partial")
- t("mcp:errors.refresh_after_disable")
- t("mcp:errors.refresh_after_enable")
The translation system will now rely solely on the localized strings without fallbacks.
* fix: translate MCP error messages to all locales
- Translated "refresh_after_disable" error message
- Translated "refresh_after_enable" error message
- Translated "disconnect_servers_partial" error message with count placeholder
- All 17 non-English locales now have proper translations instead of English text
---------
Co-authored-by: Roo Code <roomote@roocode.com>
* fix: handle current directory path "." correctly in codebase_search tool
- Fix path filtering logic in QdrantVectorStore.search() to properly handle current directory representations
- When directoryPrefix is ".", "./", "", or similar, set filter to undefined to search entire workspace
- Add comprehensive tests covering various current directory path formats including cross-platform support
- Resolves issue where codebase_search with path="." returned no results
Fixes#6514
* fix: normalize directory prefix handling in Qdrant vector store
* fix: normalize paths starting with './' and fix OS-dependency issue
- Use forward slash for splitting after toPosix() conversion
- Remove leading './' from paths like './src' to normalize them to 'src'
- Update test expectations to match correct behavior
* refactor: use path.posix.normalize instead of custom toPosix method
- Replaced directoryPrefix.toPosix() with path.posix.normalize()
- Added proper handling of backslashes before normalization
- Updated test mock to include posix.normalize method
- All tests passing (381 tests in code-index service)
* refactor: address review comments - improve path normalization
- Keep check for './' after normalization as path.posix.normalize('./') returns './'
- Use actual Node.js path.posix implementation in tests instead of custom mock
- Apply path.posix.normalize to cleanedPrefix for consistency
All 381 code-index tests pass
* fix: apply path.posix.normalize when cleaning prefix to avoid redundant normalization
Addresses review comment from @mrubens to normalize the path at line 385 instead of normalizing twice
* fix: correct current directory detection logic
The issue was that the condition checked for an empty string after normalization,
but path.posix.normalize('') actually returns '.', not ''. This caused the
current directory check to fail when an empty string was passed.
Removed the redundant empty string check since normalize('') returns '.' which
is already handled by the first condition.
---------
Co-authored-by: Roo Code <roomote@roocode.com>
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
Co-authored-by: hannesrudolph <hrudolph@gmail.com>
* fix: recover from error state when Qdrant becomes available
- Add recoverFromError method to CodeIndexManager to clear error state and reset internal services
- Update startIndexing handler to check for error state and recover before initialization
- Add comprehensive tests for error recovery functionality
Fixes#6660
* fix: address PR review comments for code indexing error recovery
- Add race condition protection for multiple rapid clicks on Start Indexing button
- Add error handling for setSystemState in recoverFromError method
- Enhance JSDoc documentation for recoverFromError method
- Add test cases for recoverFromError idempotency and error handling
* refactor: move error recovery logic into startIndexing method
- Moved error recovery from webviewMessageHandler into CodeIndexManager.startIndexing()
- This ensures error recovery happens whenever indexing is started, not just from UI
- Added race condition prevention flag within CodeIndexManager
- Simplified webviewMessageHandler by removing error state checking
- The startIndexing method now automatically recovers from error state before proceeding
* fix: remove await from startIndexing calls and update JSDoc
- startIndexing should never be awaited as it's a long-running background process
- Added JSDoc warning to never await this method
- Updated webviewMessageHandler to not await startIndexing calls
* fix: use platform-agnostic paths in code-index manager tests for Windows compatibility
---------
Co-authored-by: Roo Code <roomote@roocode.com>
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
