Skip to main content

MCP server exposing debugger tools for AI clients (Node.js and Python)

Project description

debugger-mcp

MCP server that exposes debugger tools to AI clients (Cursor, Claude Code, etc.): launch, breakpoints, continue, step, evaluate, stack, variables, probe, stop. Supports Node.js (CDP) and Python (DAP). The debugger runtime (top-coder-ai-skills-debugger) is included as a dependency — no separate installation.

Install

One install is enough (the debugger library is pulled in automatically):

pip install debugger-mcp
# or
uv add debugger-mcp
# or
poetry add debugger-mcp

Requires: Python ≥3.11. Nothing else to install.

Global

Use without adding the package to a project:

  • pipx (install once, run from anywhere):

    pipx install debugger-mcp
    debugger-mcp
    

    Then in MCP config use "command": "debugger-mcp" (pipx puts it on PATH).

  • uvx (run on demand, no install — like npx):

    uvx debugger-mcp
    

    In MCP config: "command": "uvx", "args": ["debugger-mcp"].

Run

Stdio transport (e.g. for Cursor MCP). Configure in your MCP client (e.g. Cursor) using one of these:

Option A — pipx (after pipx install debugger-mcp; use if debugger-mcp is on your PATH):

{
  "mcpServers": {
    "debugger": {
      "command": "debugger-mcp"
    }
  }
}

Option B — uvx (no install; runs on demand, like npx):

{
  "mcpServers": {
    "debugger": {
      "command": "uvx",
      "args": ["debugger-mcp"]
    }
  }
}

Option C — in-project (if you added the package to the project with uv add / poetry add / pip install):

{
  "mcpServers": {
    "debugger": {
      "command": "python3",
      "args": ["-m", "debugger_mcp"]
    }
  }
}

If using Option A and the app doesn’t see debugger-mcp on PATH, use the full path (e.g. ~/.local/bin/debugger-mcp on Linux/macOS) or use Option B (uvx) instead.

Tools

Tool Description
debug_launch Start debugging a program (program, language)
debug_breakpoint Set breakpoint (file, line)
debug_continue Resume until next breakpoint or exit
debug_step Step over or step in
debug_evaluate Evaluate expression in current frame
debug_stack Get call stack
debug_variables Get local variables
debug_probe One-shot: run to line, dump state, stop
debug_stop End the debug session

See the repo and ARCHITECTURE.md for protocol and usage details.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

debugger_mcp-0.1.1.tar.gz (10.0 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

debugger_mcp-0.1.1-py3-none-any.whl (5.2 kB view details)

Uploaded Python 3

File details

Details for the file debugger_mcp-0.1.1.tar.gz.

File metadata

  • Download URL: debugger_mcp-0.1.1.tar.gz
  • Upload date:
  • Size: 10.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.6.13

File hashes

Hashes for debugger_mcp-0.1.1.tar.gz
Algorithm Hash digest
SHA256 c0b7aac654def32dba2be95cac7c167843c247068c46a74daae699a74c52f0a1
MD5 93f254122ec6682500607ed75e9043e4
BLAKE2b-256 6c2f9dd7317e799e684077ed232b6bd64e2d32a84a353d3dbcdf4400ccac6cad

See more details on using hashes here.

File details

Details for the file debugger_mcp-0.1.1-py3-none-any.whl.

File metadata

File hashes

Hashes for debugger_mcp-0.1.1-py3-none-any.whl
Algorithm Hash digest
SHA256 53aea2d2403c6790f5026a6ac47bd28647343a7f9fbf2756f51e5c561f48faf0
MD5 9ca6baf71bc071c482c887309abc75aa
BLAKE2b-256 475535d3d3305a1604fa5883fb66e8b4c854691cb6d8d50dd68709daa1f7467b

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page