Agent Playbooks — AI skill for Claude Code

Agent Playbooks

[](https://github.com/chiragmangaldev3112/agent-playbooks/releases) [](LICENSE) [](https://github.com/chiragmangaldev3112/agent-playbooks/stargazers)

A portable rulebook for AI coding agents — how to fix bugs, build features, review code, test frontend/backend work, run security reviews, and (if you want it) operate autonomously against a standing mission. Works with Claude Code, Antigravity, Cursor, GitHub Copilot/Codex, or any agent that can read a text file.

Why

Two separate problems, both real:

1. Every tool wants its own file. Different AI coding tools each expect instructions in their own format and location. This project has one generic instruction set instead, so you write the rules once and point every tool at the same source of truth.
2. An agent left to its own judgment fails in the same few ways, every time. It guesses at an ambiguous spec instead of asking. It grades its own work instead of checking it independently. It forgets a project's own conventions the moment a new session starts. It occasionally runs something destructive because nothing stopped it.

This is that process, written down once, usable everywhere — and every playbook in it has been run against a real, throwaway test case at least once, not just written and published.

**[▶ Watch the demo in your browser](https://chiragmangaldev3112.github.io/agent-playbooks/demo.html)** (or [download the file directly](https://github.com/chiragmangaldev3112/agent-playbooks/releases/download/v1.17.0/agent-playbooks-demo.mp4)) — 9 minutes, narrating and actually running all 35 playbooks, one real terminal demo each, not a slideshow.

Install

curl -fsSL https://raw.githubusercontent.com/chiragma