AgentReady

import { Badge } from ‘@astrojs/starlight/components’;

AgentReady is an external CLI for checking whether a repository is prepared for AI-assisted development. Use it before connecting important repos to ACP, or feed its report into a session as context.

Why it helps ACP

Agents perform better when repositories contain:

• clear build and test commands.
• current documentation.
• repository instructions such as CLAUDE.md.
• predictable structure.
• CI configuration.
• security and contribution guidance.

AgentReady gives you a report you can use to improve those inputs.

Run it locally

uvx agentready assess .
agentready assess /path/to/repo --format markdown --output agentready-report.md
agentready assess /path/to/repo --format json --output agentready-report.json

Use the report in ACP

Attach or paste the report into a session:

Read artifacts/agentready-report.md and identify the top five improvements that
would make this repository easier for ACP agents to work on. Do not change files yet.

Then run focused follow-up sessions for documentation, test command cleanup, or repository instruction updates.

CI pattern

Run AgentReady in CI when you want a readiness floor:

- name: Check agent readiness
  run: |
    uvx agentready assess . --format json --output agentready-report.json
    python - <<'PY'
    import json
    score = json.load(open("agentready-report.json"))["score"]
    if score < 75:
        raise SystemExit(f"AgentReady score {score} is below threshold")
    PY