# Review Adversarial

> High-stakes adversarial code review after normal review. Challenges assumptions with distinct reviewer lenses and optional configured external model/CLI reviewers.

- Skill: `garyld1962/review-adversarial-4` (Agent Skill, multi-file: 5 files)
- Install (CLI): `npx skillmds@latest add garyld1962/review-adversarial-4`
- Raw SKILL.md: https://api.skillmd.com/api/skills/garyld1962/review-adversarial-4/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: garyld1962 (https://skillmd.com/u/garyld1962)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/garyld1962/review-adversarial-4

---


# Review Adversarial

Use for auth, payments, migrations, security-sensitive diffs, or large changes.

Load Codex-native references as needed:

- `references/workflow.md` for trigger, model, lens, and synthesis rules.
- `references/reviewer-lenses.md` for reviewer perspectives.
- `scripts/cli_probe.py` to check available external reviewer CLIs.

`references/legacy/` is archival only. Do not load it during normal review.

Prefer a different external model or CLI. If subagents are authorized as a fallback, clearly label the result as same-model adversarial review. Otherwise run lenses sequentially and clearly label each perspective.