Enhance skills, search capabilities, and MCP integration features - #26
Merged
Merged
Conversation
…hangelogs - Introduced skill to draft conventional commit messages based on git context. - Added skill for generating pull request summaries from branch commits and diffs. - Created skill to curate changelog entries from commits since the last tag. - Updated skill tags for better clarity. - Refactored tests and documentation for improved readability and organization. - Removed legacy purge script and added a new script for verifying extension launch configurations.
- Implement BraveSearchProvider for querying Brave's search API. - Implement SearxngSearchProvider for querying SearXNG instances. - Implement TavilySearchProvider for querying Tavily's search API. - Introduce SearchProviderChain to allow fallback between multiple search providers. - Add configuration resolution for search providers from environment variables. - Create tests for each search provider and the configuration resolution logic. - Implement URL safety checks for outbound retrieval. - Add safety limits for search requests and responses. - Update vitest configuration to include search-kit tests. - Ensure search-kit is free of dependencies on v8/sdk/host packages.
…unctionality - Introduced tests for enterprise techniques including fetch continuation windows, sequential thinking, time tools, and git argument safety. - Added FetchTreeAndReadWindows tests to validate fetch policy and directory tree functionality. - Implemented Git and Knowledge Graph tests to ensure proper interaction with git logs, branches, and memory graph operations. - Enhanced path containment tests to handle NFC-equivalent Unicode path components. - Developed KnowledgeGraphManager and InMemoryKnowledgeGraphStore for managing entities and relations. - Created contracts for knowledge graph entities and relations with validation schemas. - Updated memory module to include knowledge graph functionalities and types. - Documented knowledge graph features in the memory module README.
…efForPrompt functions feat(decision-policy): update DecisionPolicyPipeline to include policyFactsFirst feat(decision-policy): introduce factsFirstMinConfidence and factsFirstMinMargin thresholds test(decision-policy): add tests for compileDecisionBrief and policyFactsFirst routing feat(prompt-construction): enhance buildSystemInstructions to accept decisionBriefText feat(prompt-construction): update PromptConstructionPipeline to handle decisionBriefText feat(request-understanding): implement applyClarificationFactPatch for structured fact patches feat(request-understanding): add intersectRecommendedSkillTags for skill tag management feat(request-understanding): enhance SuperIntent to handle ambiguous slots in clarifications test(request-understanding): add tests for intersectRecommendedSkillTags and applyClarificationFactPatch
- Added method to for undoing workspace mutations to a specified RestorePoint. - Introduced function to handle the restoration logic, including validation and file restoration. - Implemented method to retrieve durable restore points for a given run. - Enhanced and to include requestId in their parameters. - Created to persist RestorePoints after successful mutations. - Added tests for restore functionality, ensuring correct restoration of files and interaction modes. - Updated various internal types and schemas to support the new restore feature.
- Added adversary evaluation logic to restrict tool execution based on defined policies. - Implemented describe_tool for progressive disclosure of tool schemas. - Normalized common tool arguments for better compatibility with existing schemas. - Enhanced preflight checks to include adversary evaluations and handle their decisions. - Updated README and documentation to reflect new features and usage. - Added tests for adversary behavior and describe_tool functionality to ensure reliability.
- Added support for SearXNG base URL configuration in the VSCode client. - Introduced SearchSettingsSnapshot interface for managing search settings. - Updated sidebar and webview components to handle search API key settings. - Enhanced decision policy to allow web search for product compatibility queries. - Implemented logic to merge network hosts from web search results into tool grants. - Added tests to verify the new decision-making behavior regarding web search. - Created a smoke test script for validating SearXNG integration.
fix: update version to 2.9.39 in search-kit package.json
…g loading in createCliClient
- Introduced a new navigation element in the log viewer UI for switching between Logs and Benchmark views. - Updated server script to support a new benchmark root argument for loading benchmark data. - Implemented handling for benchmark requests in the server, serving a new benchmark HTML page. - Enhanced server startup logs to include benchmark runner information.
…build scripts to use it
… sync-versions script for staging
…ation for duplicates - Implemented retrieval of layered L1/L2/L3 partitions in MemoryPipeline when mode is set to "layered". - Added functionality to consolidate whitespace-normalized duplicates in MemoryPipeline. feat(hybrid-retrieval): introduce importance-based candidate boosting - Added boostCandidatesByRepoMapImportance function to boost retrieval candidates based on RepoMap importance. - Implemented importanceByPathFromRepoMapEntries to create a mapping of importance scores from RepoMap entries. - Updated HybridRetriever to apply importance boosts to candidates if RepoMap entries are provided. docs(hybrid-retrieval): update README with new boosting features - Documented the optional boosting of fused scores from published RepoMap importance in the Hybrid Retrieval README. test(hybrid-retrieval): add tests for importance boosting functionality - Created unit tests for boostCandidatesByRepoMapImportance to validate boosting logic and behavior. fix: enforce stricter TypeScript settings - Updated tsconfig.json to enable noEmitOnError and specify lib options. chore: update pnpm workspace and dependencies - Added MCP package to pnpm workspace and updated dependencies in pnpm-lock.yaml. - Updated sync-versions script to include MCP package.json. refactor: improve test case documentation - Updated README files in tests and benchmark directories to reflect the current number of test cases and their categories.
…o CLI and VS Code settings
… architecture handling
- Updated to prioritize explicitly configured VS Code settings over defaults, improving the visibility of disk configurations. - Added logging for MCP server connection results, including error handling for failed servers. - Introduced to restore in-memory publish from disk after client invalidation or Extension Host restart, optimizing indexing processes. - Enhanced MCP server configuration to include Excalidraw as a built-in option, expanding the toolset available to users. - Implemented logic to handle workspace write requirements for MCP tools, allowing read access for certain tools under specific grants. - Updated tests to cover new functionality and ensure proper behavior under various grant conditions.
- Introduced new artifacts directory for storing Excalidraw diagrams. - Enhanced protocol with McpAppViewPayload to handle diagram metadata. - Implemented persistExcalidrawFromToolResult to save diagrams and related files. - Updated sidebar to display Excalidraw diagrams in the activity timeline. - Created McpAppCard component for rendering diagram details in the UI. - Added tests for Excalidraw artifact persistence and conversation carry. - Improved styling for MCP App cards and timeline entries.
- Added support for attaching MCP server IDs to tool grants, allowing for more granular control over which MCP tools are accessible based on the specified server IDs. - Implemented functions for parsing and merging MCP mentions from user messages. - Updated various components to utilize the new MCP server scoping, including tool filtering and grant validation. - Introduced tests for the new MCP attachment functionality to ensure correctness. - Added documentation for the MCP attach module detailing its usage and flow.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request updates documentation, build scripts, and configuration to reflect the addition of the
@mitii/search-kitpackage, clarifies project structure, and improves consistency across docs and scripts. It also updates VS Code and version requirements, and corrects several minor issues in documentation and automation scripts.Key changes:
Project structure and documentation updates:
CONTRIBUTING.mdto clarify the canonical repository layout, explicitly include@mitii/search-kit, and describe bundled skills and their formats. Also improved instructions and replaced ambiguous arrows with hyphens for consistency. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]README.mdto require VS Code 1.124+, reflect the new version (2.9.50), add@mitii/search-kitand web retrieval to feature lists, and clarify CLI and SDK usage. [1] [2] [3] [4] [5]Build and automation scripts:
.githooks/pre-committo usesync-versions.mjs --stage, ensuring all product packages (including new ones likesearch-kit) are staged automatically, and simplified the list of files to stage..github/workflows/npm-publish.ymlto include@mitii/search-kitin the publish order..vscode/tasks.jsonfor@mitii/hostandapps/vscodeto use correct filters.Minor documentation and formatting fixes:
CONTRIBUTING.mdandREADME.md. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]