Experiments

Git Metrics MCP Server

MCP server for analyzing git repository metrics and understanding team health. Published on npm as @jonmatum/git-metrics-mcp-server.

evergreen#mcp#git#metrics#typescript#npm#team-health#devtools

What it is

An MCP server that extracts meaningful metrics from Git repositories to help teams understand their development patterns, identify risks early, and have better conversations about code quality and team health.

Published on npm as @jonmatum/git-metrics-mcp-server and available as open source.

This is a mirror, not a microscope. Use it to reflect on team health and process quality, not to surveil individual behavior.

Available tools

The server exposes 11 MCP tools that can be invoked with natural language:

  • get_commit_stats — Overall commit statistics for a time period
  • get_author_metrics — Detailed per-contributor metrics
  • get_file_churn — Files with the most changes (complexity indicators)
  • get_team_summary — Comprehensive team performance report
  • get_commit_patterns — Analyze when people commit (burnout detection)
  • get_code_ownership — Bus factor and knowledge distribution
  • get_velocity_trends — Week/month velocity tracking
  • get_collaboration_metrics — Team interaction patterns
  • get_quality_metrics — Code quality indicators
  • get_technical_debt — Stale files and complexity hotspots
  • get_conventional_commits — Conventional commit analysis and release frequency

Production features

  • Input sanitization against command injection
  • Structured JSON logging with timestamps
  • Configurable timeouts via GIT_TIMEOUT environment variable
  • Error boundaries with detailed logging
  • CI/CD with GitHub Actions

Why it matters

Git metrics are an underutilized source of signals about team health. This server lets you ask natural language questions like "what's our bus factor?" or "are people committing late at night?" and get structured answers without manually parsing logs.

References

Experiments