challenge-review

Reviews code changes adversarially by mutating implementations and detecting test-gaming patterns.

4.7k|238|Updated Jan 25, 2016

Search Tags:#verification #mutation testing #second opinion #adversarial review #Code review #test falsification #check-gaming

One-click install

npx skills add https://github.com/werf/werf --skill challenge-review

Or copy as Structured Prompt for Agent▼

Please help me install this Agent Skill.
Skill: challenge-review
Source: https://github.com/werf/werf/tree/main/.agents/skills/challenge-review
Command: npx skills add https://github.com/werf/werf --skill challenge-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Passing tests, high coverage, and readable code do not prove an implementation is correct. This Skill performs an independent adversarial review pass that attempts to disprove implementation claims rather than confirm them, catching weak tests, check-gaming, and unverified assumptions before merge.

Core Features & Use Cases

Contract Recovery: Identifies intended behavior and acceptance criteria first, flagging unclear requirements as the initial finding.
Test Falsification: Mutates the implementation (inverting conditions, removing validation, suppressing errors) to verify that tests actually fail when the code is wrong.
Check-Gaming Detection: Flags weakened assertions, unexplained golden file updates, skipped tests, lowered thresholds, and hardcoded fixture answers.
Use Case: Before merging a high-risk change touching auth, migrations, or billing, invoke this Skill to get a second-opinion review that reports only actionable findings with concrete corrections.

Quick Start

Run a challenge review on my current diff to try to disprove that the implementation and its tests are correct.

Frequently Asked Questions about challenge-review

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

FAQPage Schema

Can I review my own code adversarially instead of getting a second reviewer?▼

Self-review catches localized bugs and weak tests but inherits your own design assumptions, making it a poor substitute for independent review of overall approach. For anything beyond mechanical changes, escalate to a fresh reviewer before merging.