skills.rest
DojoGenesis
DojoGenesisCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

adversarial-reviewer

Review code changes with three adversarial personas and issue a BLOCK, CONCERNS, or CLEAN verdict.

2|Updated Apr 8, 2026
Search Tags:#software-quality#code-review#adversarial#security-audit#pull-request#peer-review
One-click install
npx skills add https://github.com/DojoGenesis/gateway --skill adversarial-reviewer-dojogenesis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: adversarial-reviewer
Source: https://github.com/DojoGenesis/gateway/tree/main/plugins/strategic-thinking/skills/adversarial-reviewer
Command: npx skills add https://github.com/DojoGenesis/gateway --skill adversarial-reviewer-dojogenesis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Adversarial code review that breaks the self-review monoculture. Use when you want a genuinely critical review of recent changes, before merging a PR, or when you suspect Claude is being too agreeable about code quality. Forces perspective shifts through hostile reviewer personas that catch blind spots the author's mental model shares with the reviewer.

Core Features & Use Cases

  • Three adversarial personas — Saboteur, New Hire, Security Auditor
  • Mandatory findings — Each persona must surface at least one issue
  • Severity promotion — Findings caught by multiple personas are promoted to the next severity level
  • Self-review trap breaker — Concrete techniques to disrupt shared mental-model blind spots
  • Structured verdicts — BLOCK / CONCERNS / CLEAN with explicit merge guidance

Quick Start

Run adversarial-review on the desired changes, for example /adversarial-review for staged/unstaged changes, or use --diff <ref> or --file <path> to target a specific set of changes.

Frequently Asked Questions about adversarial-reviewer

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

FAQPage Schema
How do I run an adversarial code review on a pull request?▼

Run adversarial-review on staged changes or target specific diffs using --diff <ref> or --file <path> to conduct a multi-persona peer review and output a BLOCK/CONCERNS/CLEAN verdict.

What is an adversarial peer review and how does it find security issues?▼

Adversarial peer review uses three hostile reviewer personas—Saboteur, New Hire, and Security Auditor—to force perspective shifts and catch security or quality blind spots that standard self-reviews miss.

Can I use this code review tool on specific file diffs across a repository?▼

Yes, you can target specific file diffs across a repository by providing the --file <path> argument to review individual files, or use --diff <ref> to review specific commit changes.

How does severity promotion work for security audit findings?▼

Severity promotion elevates findings caught by multiple reviewer personas to the next severity level, ensuring that overlapping security or quality concerns are prioritized in the final structured report.

What is the best way to break the self-review trap before merging a PR?▼

Use adversarial-review to break the self-review monoculture by applying concrete techniques that disrupt shared mental-model blind spots, requiring mandatory findings from each reviewer persona before merging.

Related Skills

Explore More Software Engineering Skills→
BSteffaniakBSteffaniak

commit-message-staged-write

Generate Git commit messages from staged changes and write them for commit tools.

Community
Intermediate
devCharuzudevCharuzu

adversarial-reviewer

Simulate three hostile personas to identify adversarial findings in code reviews.

Community
Advanced
FaisalAlqarniFaisalAlqarni
5

springboot-verification

Automate build, static analysis, testing, and security scanning for Spring Boot projects.

Community
Advanced
outcomeengoutcomeeng
1

reviewing-python

Review Python codebases adversarially and issue APPROVED or REJECTED verdicts.

Official
Advanced