Agent-Ready Repo Auditor

Repository readiness for AI agents is usually discussed in vague terms. This tool makes the question concrete by showing what documentation and workflow evidence it can actually find. The resulting score is useful as a checklist, but it does not measure code quality, security, compliance, or whether a repository is safe for autonomous changes.

The tool runs as a CLI or GitHub Action. It reads a bounded set of public GitHub data and creates an agent-ready-audit.md report.

GitHub Actions users also get a workflow summary plus score, band, report-path, and revision outputs for later steps. The tool does not clone or execute repository code. Missing, private, inaccessible, or truncated repositories fail closed instead of receiving a partial score.

Maintainers who want a repeatable first check of their repository instructions, contributor intake, and verification evidence.

It can point out documentation gaps before people or agents start making changes. A human maintainer still has to decide what the findings mean and what to fix.

This is a practical checklist for a problem that often gets hand-waved. It shows whether the basic written signals are present, then gets out of the way.

Treat the score as a prompt for maintenance work, not as a quality grade or permission to let an agent run unattended.