Agent♥︎Age
Catalog

patchwork — Codebase Conventions for AI Agents

Official

by SaiNarayana-B · Python

Scan any codebase, generate CONVENTIONS.md, expose conventions as MCP tools for AI agents.

Patchwork scans codebases to generate CONVENTIONS.md and exposes these conventions as MCP tools for AI agents.

🛠️ Key Features

  • Scans a codebase to mine conventions and style guidance
  • Generates a CONVENTIONS.md document from findings
  • Exposes conventions as MCP tools for AI agents
  • Open-source MIT-licensed with Python 3.9+ compatibility indicators

🚀 Use Cases

  • Establish and enforce project-wide coding standards
  • Provide agents with reliable style guidance during tasks
  • Automate documentation of codebase conventions
  • Reuse generated conventions across multiple repositories

⚡ Developer Benefits

  • Clear, machine-accessible conventions for MCP agents
  • Reduces risk of agents inventing or deviating from style
  • Lightweight, Python-based tooling suitable for CI workflows
  • Clear readme excerpt and metadata to facilitate integration

⚠️ Limitations

  • Source data excerpt indicates partial readme availability; behavior may depend on repository structure
  • Generated conventions rely on codebase presence and quality of in-repo patterns
patchwork — Codebase Conventions for AI Agents - agentage MCP Catalog