Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
This collection provides the tools to build iterated coding agents that operate as closed-loop systems. By defining sensors, controllers, and actuators tailored to a specific repository, developers can automate complex maintenance and development tasks through structured agentic workflows.
The skills guide the transition from initial design interviews to the implementation of locally runnable components and scheduled GitHub Actions workflows, ensuring agents can respond to disturbances in the codebase with precision.
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
A function or entry point and its relevant source code, or access to the repository containing it.
Use the show-me skill to explain the runtime control flow of the function I provide. Show a concise call tree with only the calls and boundaries needed to understand it, and keep the accompanying prose brief.
The CLAUDE.md file to improve, supplied as text or an accessible file.
Use the improve-claude-md skill to rewrite the CLAUDE.md I provide. Leave foundational project context as plain markdown at the top, and wrap task-specific guidance in <important if="condition"> blocks with specific, targeted conditions.
Access to its prop definitions and repository call sites, including supporting stories, tests, or mocks where available.
Use the narrow-react-prop-types skill to inspect one React component I identify. Find its live call sites, separate them from stories, tests, and mocks, and propose which props should be required, optional, or removed. Explain any required nullable props using the observed runtime states.