Enforce adversarial cross-model code review with mutation-based test verification.
npx skills add https://github.com/broomva/skills --skill cross-review-broomva
Please help me install this Agent Skill. Skill: cross-review Source: https://github.com/broomva/skills/tree/main/skills/governance/cross-review Command: npx skills add https://github.com/broomva/skills --skill cross-review-broomva
This skill solves the problem of single-model echo chambers where an AI model fails to identify its own blind spots during the planning, implementation, and review phases of software development.
Run the cross-review pre-push command to validate your current changes against the adversarial gate before submitting your pull request.
High-intent search queries and answers about installing and using this skill.
Adversarial code review prevents AI blind spots by enforcing a mandatory review through a different model or a fresh-context subagent. This cross-vendor verification ensures code quality meets a strict anti-slop rubric by actively identifying systematic blind spots that single-model echo chambers miss.
To validate your test suite before a pull request, use mutation-based test verification. This technique neutering the target code to verify the test fails, applying multi-strata evaluation to ensure your tests actively discriminate against code changes rather than passing silently.
Cross-vendor verification can utilize a fresh-context subagent as an alternative to the optional codex CLI. The skill requires git and bash for core adversarial review operations, while the codex CLI specifically enhances cross-vendor validation but remains an optional dependency.
The best way to enforce an anti-slop rubric for AI-generated code is triggering an adversarial review gate before pushing substantive pull requests. This gate scores the implementation against a 10-point rubric, ensuring only high-quality, well-tested code passes multi-strata evaluation.
You should avoid adversarial cross-model review for non-substantive pull requests. This process is designed specifically for high-integrity validation of substantive changes and design plans, requiring multi-strata evaluation and mutation-based test verification that may be excessive for trivial modifications.
naj2rCreates git commits with auto-generated messages and optional push.
yocxy2Merge pull requests and verify production deployment health with canary checks.
Kushal9889Sets up Husky git hooks to automate code quality checks and testing before commits.
konglong87Apply Charlie Munger's mental models to identify decision problems and surface biases.
