โ† AI library

Fixture-verified workflow kit โ€” reviewed September 9, 2026

Review an AI-authored code change with three bounded roles

Read the Agency Agents source review: โ†’

For teams using coding agents in repositories with executable acceptance checks.

Stack
Agency Agents: Minimal Change Engineer โ†’ Project checks โ†’ Agency Agents: Code Reviewer โ†’ Agency Agents: AI-Generated Code Security Auditor โ†’ Human approver

Run this recipe

Work through it here

Progress stays in this browser. The downloaded Markdown kit works in any notes app or repository.

0/6 complete

Is this recipe useful?

Procedure

  1. Pin the Agency Agents revision and install only the three named roles into a disposable or project-scoped path; record the harness, model, instructions, tools, and permissions.
  2. Give the Minimal Change Engineer one outcome, explicit exclusions, allowed files, and an executable acceptance command; it may implement but not deploy.
  3. Run the repository check and return exact failures at most twice, stopping for a person when the same failure repeats.
  4. Give the Code Reviewer the task, diff, and check log with read-only access; require separate blockers, suggestions, and nits rather than edits.
  5. Give the AI-Generated Code Security Auditor the same evidence plus relevant trust boundaries; keep scans local and read-only and exclude credentials from prompts and logs.
  6. Have a person resolve or accept every blocker, rerun the same checks, capture fresh runtime proof, and record the ship or stop decision.

Acceptance artifact

A scope contract, pinned role files, reviewed diff, verifier logs, separate correctness and security findings, runtime proof, and human decision

Do not use it blindly

Role prompts do not isolate permissions or prove correctness; reviewers should not receive write, deploy, or production-secret access, and a larger roster adds selection and context overhead.

Evidence and setup