What problem does it solve? Repositories often lack a concise, agent-oriented onboarding document, so AI coding agents waste context rediscovering build commands, conventions, and hard rules on every session. This Skill inspects the repo and writes a tight, reference-heavy AGENTS.md that a fresh agent can read once and stay unblocked. ## Core Features & Use Cases - Create and update flows: Drafts a new AGENTS.md from repo discovery, or surgically updates an existing one by diffing repo state against the file's last-touched commit. - Repo-level and directory-level scope: Produces a root onboarding guide or a focused per-directory guide (e.g. src/api/AGENTS.md) based on where it is invoked. - Quality-gated output: Enforces a 200-400 word budget, checkable rules only, @-references instead of code snippets, and critical-rules-first ordering. - Use Case: After restructuring a monorepo's scripts and CI, run the update flow to refresh AGENTS.md so it reflects renamed scripts, new lint configs, and current commit conventions without rewriting hand-edited sections. ## Quick Start Ask the agent to create an AGENTS.md onboarding document for this repository, optionally passing a target directory or file path.
