adversarial-reviewer

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

Updated Apr 16, 2026

Search Tags:#quality-assurance #code-review #adversarial #security-audit #pull-request #review-tool #hostile-personas

Ready to run

npx skills add https://github.com/devCharuzu/philfida-taskmanage --skill adversarial-reviewer-devcharuzu

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: adversarial-reviewer
Source: https://github.com/devCharuzu/philfida-taskmanage/tree/main/.windsurf/skills/adversarial-reviewer
Command: npx skills add https://github.com/devCharuzu/philfida-taskmanage --skill adversarial-reviewer-devcharuzu

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Adversarial code review that forces genuine perspective shifts through three hostile reviewer personas (Saboteur, New Hire, Security Auditor). Each persona MUST find at least one issue — no "LGTM" escapes. Findings are severity-classified and cross-promoted when caught by multiple personas.

Core Features & Use Cases

Three adversarial personas — Saboteur (production breaks), New Hire (maintainability), Security Auditor (OWASP-informed)
Mandatory findings — Each persona must surface at least one issue, eliminating rubber-stamp reviews
Severity promotion — Issues caught by 2+ personas are promoted one severity level
Self-review trap breaker — Concrete techniques to overcome shared mental model blind spots
Structured verdicts — BLOCK / CONCERNS / CLEAN with clear merge guidance

Quick Start

Run the adversarial-review command to generate a structured, three-persona review of the current changes.

Frequently Asked Questions about adversarial-reviewer

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

FAQPage Schema

What is adversarial code review and how does it prevent rubber-stamp approvals?▼

Adversarial code review prevents rubber-stamp approvals by simulating hostile reviewer personas that must each find at least one issue. This ensures genuine perspective shifts and thorough scrutiny before merge, eliminating superficial "LGTM" escapes.

How do I run a multi-persona pull request review to catch production breaks?▼

Run the adversarial-review command on your pull request to generate a structured, three-persona review. It applies the Saboteur, New Hire, and Security Auditor personas to your code changes, enforcing mandatory findings and a deterministic workflow with severity classification.

Can I use this security audit approach for self-review on my own code changes?▼

Yes, the adversarial review is designed to break the self-review trap. It uses concrete techniques to overcome shared mental model blind spots by forcing you to view your code through hostile personas like the OWASP-informed Security Auditor.

Does the code review workflow provide structured merge guidance for pull requests?▼

The code review workflow provides structured merge guidance by outputting a BLOCK, CONCERNS, or CLEAN verdict. Issues caught by two or more personas are promoted one severity level, ensuring clear actionable outcomes before merge.

What are the limitations of mandatory findings in a deterministic code review workflow?▼

The mandatory findings limitation means each persona must surface at least one issue, which may produce minor findings on already clean code. This deterministic approach prioritizes thorough scrutiny over quick approvals, potentially slowing down trivial pull requests.