adversary

Adversarial red-teaming on a pull request systematically applies targeted attack lenses to identify hidden coupling, unhandled edge cases, and logic flaws. You can execute this skeptical review by providing a GitHub PR number as the artifact identifier.

Mandatory self-refutation in code review requires the reviewer to attempt to refute their own candidate findings before reporting them. This protocol ensures high-quality findings by filtering out false positives during adversarial artifact analysis.

Adversarial review can be used on architecture plans and specifications to systematically attempt to break them. It forces a rigorous search for omissions and failure modes that traditional review processes often miss.

You need the gh CLI to resolve targets via GitHub PR numbers when performing an adversarial review. The skill depends on gh to fetch the pull request artifacts and execute the skeptical-by-default analysis.

To identify missing requirements in a software specification, apply artifact-specific attack lenses using an adversarial review approach. This forces a rigorous search for omissions and failure modes by assuming the specification is broken or incomplete.

Standard code reviews have blind spots and often miss omissions and failure modes. Adversarial artifact review overcomes this by assuming artifacts are broken or incomplete, forcing a rigorous search for hidden coupling and missing requirements.