Adversarial Review

This spawns code reviewers on the opposite AI model to challenge your work from adversarial angles. If you're using Claude, it fires up Codex reviewers via CLI. If you're Codex, it runs Claude. The reviewers attack from distinct lenses like Skeptic, Architect, and Minimalist, grounded in principles from your brain directory. You get a synthesized verdict with findings, then apply your own judgment about what to accept or reject. It's designed for post-cook sessions with large diffs or after planning phases. The hard part is the cross-model requirement, which means actual CLI calls, not subagents. Worth running when you want genuine friction, not rubber stamp validation.

skills.sh installs: 1.6K
GitHub stars: 607
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add poteto/noodle --skill adversarial-review --agent claude-code

Installs into .claude/skills of the current project.

Files

Adversarial Review

Spawn reviewers on the opposite model to challenge work. Reviewers attack from distinct lenses grounded in brain principles. The deliverable is a synthesized verdict — do NOT make changes.

Hard constraint: Reviewers MUST run via the opposite model's CLI (codex exec or claude -p). Do NOT use subagents, the Agent tool, or any internal delegation mechanism as reviewers — those run on your own model, which defeats the purpose.

Step 1 — Load Principles

Read brain/principles.md. Follow every [[wikilink]] and read each linked principle file. These govern reviewer judgments.

Step 2 — Determine Scope and Intent

Identify what to review from context (recent diffs, referenced plans, user message).

Determine the intent — what the author is trying to achieve. This is critical: reviewers challenge whether the work achieves the intent well, not whether the intent is correct. State the intent explicitly before proceeding.

Assess change size:

SKILL.md on GitHub