Language

This MCP server exposes language server protocol (LSP) features like definitions, references, and diagnostics to LLMs, enhancing code navigation and understanding.

Connections & Capabilities

Connects To

GitHub

Capabilities

readwriteexec

Quickstart

Config

{
  "mcpServers": {
    "language-server": {
      "command": "/full/path/to/your/clone/mcp-language-server/mcp-language-server",
      "args": [
        "--workspace",
        "/path/to/workspace",
        "--lsp",
        "language-server-executable"
      ],
      "env": {
        "LOG_LEVEL": "DEBUG"
      }
    }
  }
}

Exposed MCP Tools (6)

safe

definition

Retrieves the source code definition of a symbol.

Read-only operation; no side effects.

safe

references

Locates all usages of a symbol throughout the codebase.

Read-only operation; no side effects.

safe

diagnostics

Provides diagnostic information (warnings, errors) for a file.

Read-only operation; no side effects.

safe

hover

Displays documentation or type hints for a given location.

Read-only operation; no side effects.

high

rename_symbol

Renames a symbol across a project.

Modifies code structure; potential for breaking changes.

high

edit_file

Applies multiple text edits to a file based on line numbers.

Modifies file content; potential for introducing errors.

Safety Assessment

This server is relatively safe for read-only operations. However, the ability to execute code indirectly through language server features introduces some risk. Ensure the underlying language server is up-to-date and the workspace is properly secured.

Relies on established language server implementations (gopls, rust-analyzer, etc.)
Primarily exposes read-oriented operations like definition lookup and finding references.
Sandboxed by the underlying language server's capabilities.
No direct filesystem access outside the workspace.

Can execute code indirectly through language server features (e.g., code completion, refactoring).
Potential for information leakage if the language server exposes sensitive data in diagnostics or hover information.
Vulnerable to exploits in the underlying language server if not properly updated.
Workspace access grants read permissions to the codebase.