# ezyang/codemcp

Coding assistant MCP for Claude Desktop

Repository: https://github.com/ezyang/codemcp
Canonical: https://ross.abutalabs.com/products/codemcp
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2025-12-25T02:10:23+00:00

## Health v2 (maintenance only)
Score: 36/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 58, release rhythm 8, longevity 38
- inputs: {"age_days": 542, "days_push": 252, "days_rel": 500, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1608, forks 140 (observed 2026-08-28T04:05:10.726973+00:00)

## What it is
codemcp is an MCP server that turns Claude Desktop into a pair programming assistant that can directly edit files, run tests, and implement features on a local codebase. It emphasizes auto-accept agent workflows, predeclared shell commands, and fine-grained Git versioning of all LLM edits.

## Use cases
- have Claude implement features directly in my codebase
- let an AI agent fix bugs and run tests locally
- roll back AI code edits with fine-grained git versioning
- use Claude Pro subscription for agentic coding with no per-token cost
- pair program with Claude without copying code in and out of chat
- use an AI coding agent that works with any IDE like vim

## When to avoid
- the author considers it fully obsolete now that Claude Code is included with subscriptions
- you want a modern actively developed agentic coding CLI
- you need an IDE-integrated experience like Cursor or VSCode extensions

## Facets
- artifact type: service
- maturity: maintenance
- function: mcp, agent-framework, developer-tools
- domain: developer-tools, large-language-models
- platform: windows, python
- tags: claude-desktop, coding-assistant, pair-programming, git-versioning, auto-accept, obsolete, ai-agents, automation, macos, linux

## Member repositories
- ezyang/codemcp (main) score 36

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:10.726973+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:51:25.564563+00:00, confidence not recorded.
  - readme: https://github.com/ezyang/codemcp (fetched 2026-08-28T04:05:10.726973+00:00, sha 8ec64da81199)
- Data as of 2026-08-30T08:39:29.467469+00:00.