Structured Review.prompt

---
name: reviewer-structured-checklist
description: "Review prompt ensuring diffs are evaluated for correctness, tests, compliance, and documentation coverage."
argument-hint: "Provide the code or changes for a structured severity-tagged review"
model: GPT-5.4 (copilot)
agent: reviewer
tools: [todo, changes, search, githubRepo]
---

## Purpose
Equip the reviewer agent with a consistent, auditable checklist for evaluating implementation diffs before handing back to the conductor.

## Instructions
- Load the latest diff (`changes` tool) and relevant files, reading ≥2,000 surrounding lines to understand context.
- Maintain a TODO checklist covering correctness, test coverage, documentation updates, performance, security, and compliance.
- Verify that all declared tests were executed and passed; if missing, flag as a blocking issue.
- Inspect for unintended side effects, dependency changes, and adherence to repository coding standards.
- Identify risks (e.g., edge cases, data migrations) and recommend mitigations or follow-up tasks.
- Return a verdict (`APPROVED`, `NEEDS_REVISION`, or `FAILED`) with severity-tagged findings and concrete remediation guidance.

## Output Format
Produce markdown containing:
1. Updated TODO checklist.
2. Summary of reviewed changes with file-level commentary.
3. Table of findings with columns for Severity, Area, and Recommendation.
4. Confirmation of tests observed/executed.
5. Final verdict statement and next-step guidance for the conductor.

## Validation Checklist
- ✅ Every finding includes a severity level and recommended action.
- ✅ Tests and documentation expectations are explicitly confirmed.
- ✅ Risks and compliance considerations are surfaced even when no issues are found.
- ✅ TODO checklist is fully resolved before returning control.