Your build starts here
Replace bracketed details with your own goal, context, and checks. Keep the first build small enough to inspect.
Review [change, branch, or files] against this original brief: [requirements and acceptance cases]. Treat the implementation author's summary as a claim to check. Read the relevant change and follow the affected user journey where the environment supports it. Keep the review focused on this scope.
For each finding, give the observed trigger, actual and expected behavior, affected location, and practical consequence. Distinguish a demonstrated defect from a question, preference, or check you could not run. Do not create speculative defects merely to fill a quota. If a requirement is ambiguous, identify the decision that needs clarification instead of assuming the strictest possible interpretation.
Return a short prioritized list and the evidence behind it. Do not edit code during this review. If no actionable problem is found, say what was checked and state the remaining verification limits. Suggest one targeted follow-up only when it would resolve a concrete uncertainty, and do not imply that an AI review certifies the whole application.Make this prompt yours
Useful as a second perspective on a bounded change. Provide the original requirement, not just the builder’s account of what it finished.
You might also hear
- AI code review prompt
- independent evaluator
- acceptance review
The docs behind it
These references explain the underlying concepts and APIs. The prompt templates and project plans are original examples to adapt to your own setup.
Research sources reviewed .
Editorial review · October 28, 2026
Supported: Original, unexecuted review template. Criteria-based evaluation is supported by Anthropic's experiment; GitHub documents missed issues and false positives. The prompt requests evidence and does not certify reviewer independence or coverage.