nx-mcp

Exposes Nx's internal task graph and project dependency metadata through the Model Context Protocol, allowing AI clients to query project structure, task definitions, and dependency relationships without direct filesystem access. Implements MCP resource handlers that serialize Nx's graph data structures into JSON-RPC responses, enabling stateless queries of monorepo topology.

Implements MCP tools that allow AI clients to trigger Nx task execution (build, test, lint, etc.) with automatic context injection about affected projects and dependencies. Wraps nx exec/run commands through MCP tool handlers that capture task output, exit codes, and logs, returning structured results to the AI client for decision-making.

Provides MCP resources that return filtered, project-specific source code and configuration files to AI clients, implementing smart context windowing based on project boundaries and dependency relationships. Uses Nx's project metadata to determine file inclusion/exclusion, reducing irrelevant context sent to LLMs and improving token efficiency.

Exposes Nx's affected project detection algorithm through MCP tools, allowing AI clients to query which projects are impacted by code changes in specific files or branches. Implements handlers that call nx affected with various filters and return structured lists of affected projects, enabling AI to make informed decisions about what to test or rebuild.

Provides MCP resources that expose Nx workspace configuration (nx.json, project.json files, plugin settings) and installed plugin metadata to AI clients. Serializes Nx's configuration objects and plugin registry into JSON-RPC responses, enabling AI to understand workspace-level settings, executor configurations, and available generators.

Implements MCP tools that allow AI clients to invoke Nx generators (schematics) with specified options, enabling autonomous code scaffolding and project creation. Wraps nx generate commands through tool handlers that accept generator names and option objects, execute the generator, and return results including created/modified files.

Exposes Nx's computed dependency graph through MCP resources in multiple formats (adjacency lists, edge lists, visual descriptions), enabling AI clients to reason about project relationships and identify circular dependencies or architectural issues. Implements graph serialization handlers that convert Nx's internal graph data structures into formats suitable for LLM analysis.

Provides MCP resources that expose ESLint, Nx lint rules, and other code quality tool configurations to AI clients, including rule definitions, severity levels, and fix suggestions. Implements handlers that parse lint configuration files and return structured rule metadata, enabling AI to understand what violations to fix and how.