Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

Adversarial Code Reviewer

Runs your diff through three hostile reviewer personas so you never get a lazy "LGTM" before merging.

Dev & CodingIntermediate★ 24,151⑂ 3,405AI score 9/10Last updated: Aug 9, 2026

What it does

When an AI reviews code it just wrote, it shares the author's mental model and tends to conclude "looks good to me." This skill forces a real perspective shift using three adversarial personas:

  • The Saboteur — "I'm trying to break this in production": unvalidated input, race conditions, swallowed errors, resource leaks
  • The New Hire — "I must modify this in 6 months with zero context": vague names, magic numbers, implicit knowledge, missing tests
  • The Security Auditor — OWASP-informed checklist: injection, broken access control (IDOR), hardcoded secrets, insecure defaults

Every persona must report at least one finding, and any issue flagged by 2+ personas is promoted one severity level (NOTE→WARNING→CRITICAL). Output is a structured report ending in a BLOCK / CONCERNS / CLEAN verdict.

Who it's for

  • Solo developers and small teams merging PRs with no human reviewer
  • Anyone uneasy about how easily AI approves their code
  • Engineers touching auth, payments, data access, or public API endpoints
  • Developers wanting a blind-spot check after a long coding session

Examples

  1. Pre-merge PR gate: /adversarial-review --diff main...HEAD reviews the whole branch and returns a BLOCK/CONCERNS/CLEAN verdict.
  2. Recent-commit audit: /adversarial-review --diff HEAD~3 surfaces null dereferences and missing error handling in three rushed commits.
  3. Deep dive on a critical file: /adversarial-review --file src/auth.ts reads the full file and checks it against the OWASP table for IDOR, token leakage, and missing role checks.

· · · Install guide · · ·

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the .gemini/skills/adversarial-reviewer folder from the GitHub repo alirezarezvani/claude-skills into my ~/.claude/skills/adversarial-reviewer/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/alirezarezvani/claude-skills.git /tmp/claude-skills && mkdir -p ~/.claude/skills && cp -r /tmp/claude-skills/.gemini/skills/adversarial-reviewer ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open your terminal.
  2. Clone the repo into a temp folder: git clone https://github.com/alirezarezvani/claude-skills.git /tmp/claude-skills
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy just this skill: cp -r /tmp/claude-skills/.gemini/skills/adversarial-reviewer ~/.claude/skills/
  5. Verify with ls ~/.claude/skills/adversarial-reviewer — you should see SKILL.md.
  6. Restart Claude Code, open a git repository, and run /adversarial-review (or ask "give me an adversarial review of my recent changes").
  7. Narrow the scope with --diff main...HEAD or --file <path> as needed.
View source on GitHub ↗License: MIT