sherlock-review

Investigate code-change claims against evidence from diffs, tests, and logs.

436|78|Updated Sep 11, 2025
One-click install
npx skills add https://github.com/proffesor-for-testing/agentic-qe --skill sherlock-review
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sherlock-review
Source: https://github.com/proffesor-for-testing/agentic-qe/tree/main/.claude/skills/sherlock-review
Command: npx skills add https://github.com/proffesor-for-testing/agentic-qe --skill sherlock-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill performs rigorous, evidence-based code reviews to determine what actually happened versus what was claimed, aiding root-cause analysis and verifiable fixes.

Core Features & Use Cases

  • Observe & Deduce: Gather code, tests, history, and runtime behavior to form conclusions.
  • Document with Proof: Provide findings rooted in reproducible evidence.
  • Agent Collaboration: Integrates code-review, security, and performance perspectives.

Quick Start

Run an investigative review on a recent PR to verify the claims against the observed behavior.

Frequently Asked Questions about sherlock-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify that a bug fix actually solves the root cause?

Evidence-based code review examines the claimed fix against diffs, tests, logs, and historical changes to confirm the root cause was identified and addressed. Compare the fix to observed behavior in tests and reproduction steps to ensure the claim matches reproducible proof.

What's the best way to investigate code changes across a codebase and tests?

Gather evidence by collecting diffs, test results, runtime logs, and change history. Eliminate impossible explanations systematically, then document findings rooted in reproducible proof that connects claims to observed results.

Can I use investigative code review to verify performance claims?

Yes. Investigative review applies to performance claims by gathering evidence from benchmarks, logs, and behavioral changes, then comparing claimed improvements against measured results to confirm the fix is valid.

How do I document code review findings with reproducible proof?

Root your conclusions in gathered evidence: diffs showing what changed, tests demonstrating the fix works, and logs confirming behavior. Structure findings to eliminate impossible explanations and make the proof reproducible for others.

When should I use deduction to analyze what actually happened in a codebase?

Use deduction when claims about bug fixes, features, or root causes don't align with observed behavior. Gather code, test, and historical evidence to form conclusions by eliminating impossible explanations and confirming what the evidence shows.

Does investigative review help with security and performance analysis in code?

Yes. The skill integrates code-review, security, and performance perspectives to investigate claims across all three domains, using evidence from diffs, tests, and behavior to verify fixes and root causes.