{"id":"jonrad-lsp-mcp","name":"lsp-mcp","homepage":null,"repo_url":"https://github.com/jonrad/lsp-mcp","category":"devtools","subcategories":[],"tags":["mcp","lsp","language-server-protocol","typescript","ai-agents","code-intelligence","developer-tools"],"what_it_does":"Provides an MCP server that exposes language-aware code intelligence by wrapping one or more LSP (Language Server Protocol) servers (e.g., typescript-language-server) so LLM/agent clients can request LSP features over MCP.","use_cases":["Enable LLM/agents to use code-aware LSP capabilities (symbols, go-to-definition, references, diagnostics, etc.) for one or more languages in a repo","Editor/agent context enrichment for programming assistants (e.g., Claude Desktop, Cursor, MCP CLI clients)","Rapidly adding LSP-backed tooling to agent workflows via Docker or npx","Dynamic LSP method exposure driven by an LSP JSON Schema/config"],"not_for":["Production-grade, security-sensitive deployments without reviewing the runtime permissions and sandboxing","Environments requiring a strongly specified REST/SDK integration contract","Use cases needing fine-grained, per-user authorization and auditability built into the service","High-reliability, long-lived sessions without additional operational hardening"],"best_when":"You can run the MCP server locally/in your infrastructure (often alongside Docker) and want LLMs to query LSP features for the files you provide to the runtime.","avoid_when":"You need centralized managed hosting with documented SLAs, strong auth controls, or you cannot safely mount/share code into the server/container environment.","alternatives":["Use MCP filesystem + direct LSP server connections (where supported by your MCP client)","Use existing editor/IDE integration APIs to retrieve LSP context","Stand up your own LSP server and proxy it behind your own MCP wrapper (or use a maintained MCP-LSP adapter)","Use multilspy/microsoft LSP clients directly in your agent runtime (bypassing MCP)"],"af_score":50.5,"security_score":29.2,"reliability_score":26.2,"package_type":"mcp_server","discovery_source":["github"],"priority":"high","status":"evaluated","version_evaluated":null,"last_evaluated":"2026-03-30T13:39:15.427651+00:00","interface":{"has_rest_api":false,"has_graphql":false,"has_grpc":false,"has_mcp_server":true,"mcp_server_url":null,"has_sdk":false,"sdk_languages":[],"openapi_spec_url":null,"webhooks":false},"auth":{"methods":[],"oauth":false,"scopes":false,"notes":"No authentication/authorization is described in the provided README. MCP clients generally connect over stdio (e.g., Claude Desktop) or local execution, so auth may be delegated to deployment environment rather than built into the service."},"pricing":{"model":null,"free_tier_exists":false,"free_tier_limits":null,"paid_tiers":[],"requires_credit_card":false,"estimated_workload_costs":null,"notes":"No pricing information provided (appears to be open-source)."},"requirements":{"requires_signup":false,"requires_credit_card":false,"domain_verification":false,"data_residency":[],"compliance":[],"min_contract":null},"agent_readiness":{"af_score":50.5,"security_score":29.2,"reliability_score":26.2,"mcp_server_quality":70.0,"documentation_accuracy":60.0,"error_message_quality":0.0,"error_message_notes":null,"auth_complexity":90.0,"rate_limit_clarity":0.0,"tls_enforcement":20.0,"auth_strength":20.0,"scope_granularity":10.0,"dependency_hygiene":55.0,"secret_handling":50.0,"security_notes":"No explicit TLS/auth/scope model is described; the likely stdio/local execution pattern reduces network exposure but increases risk if run with broad filesystem access. The README suggests mounting local directories into a container, which can expose sensitive code/assets; sandboxing and least-privilege mounting are important. Security hygiene beyond dependency selection is not documented in the provided materials.","uptime_documented":0.0,"version_stability":45.0,"breaking_changes_history":35.0,"error_recovery":25.0,"idempotency_support":"false","idempotency_notes":null,"pagination_style":"none","retry_guidance_documented":false,"known_agent_gotchas":["README notes this is a POC state, implying rough edges and incomplete spec coverage.","Claude Desktop via npx is described as finicky (tools may still work even when it reports failure).","Multiple LSPs at the same time are mentioned but later qualified as not yet supported in some npx workflow, so config expectations may vary by integration path."]}}