AI agent readiness

Is my code repo AI ready?

A repo is AI ready when a coding agent can find the commands, boundaries, and verification standard without guessing. The fastest check is the instruction surface the agent reads first: `AGENTS.md`.

Run the checker View example AGENTS.md

Quick repo-readiness check

Setup Install command, environment file, package manager, and local services are named.

Commands Test, lint, type-check, and build commands are exact enough to copy and run.

Repo map Important folders, generated files, ownership boundaries, and monorepo scopes are explicit.

Safety Secrets, production data, migrations, destructive commands, and approval rules are visible.

Review PR summary, verification evidence, screenshots, and test expectations are stated.

Maintenance The file says when to update instructions as scripts, architecture, and workflows change.

What AGENTS.md should do

`AGENTS.md` should act as a concise repo index, not a long policy dump. Put the commands and rules that apply to every change near the top, then link to deeper docs only when agents need more detail.

The goal is not to document every implementation detail. The goal is to remove the decisions that cause agents to guess: how to set up the repo, which checks count, which paths are sensitive, and what evidence belongs in a pull request.

Score the current draft

The free checker scores pasted `AGENTS.md` content locally in the browser. It checks setup, commands, repo map, conventions, safety, review expectations, scope, and maintenance cues, then returns a copyable Markdown fix plan.

Check an AGENTS.md draft