# Adversarial Review · KZvilla

> [skill, loads itself] Background knowledge for hostile audits; /lagrange/audit is the explicit trigger. Runs a skeptical, concise, evidence-based audit of either (a) an already-built implementation, checked against what the plan/spec/ticket actually required, or (b) a proposed plan/design, checked against the flows, business rules, data model, tests, and conventions that already exist in the project, including an explicit check for over-engineering. Use proactively whenever the user asks to audit, review, do a hostile/adversarial code review, verify whether the agent followed the plan, check if this plan fits the current system, spot over-engineering, or hands over a PR/diff/commit alongside the original plan, or a plan/RFC alongside repo access, expecting the two to be checked against each other. This is not a style review or a friendly summary: it is a skeptical audit that actively looks for evidence that should block approval.

## Facts
- Page: https://tashan.sh/capability/skill-kzvilla-adversarial-review
- tashan id: skill:KZvilla/adversarial-review
- Source: https://github.com/KZvilla/claude-plugin-antigravity
- Type: skill
- Category: devtools
- tashan score: not scored (catalogued only — too little public evidence)
- Adoption: 14.0
- Upkeep: 64.0
- Freshness: 100.0
- Evidence coverage: 84% of the inputs this score can use
- Health: active
- Instruction depth: not yet graded
- License: MIT
- Official: no

## Install

```sh
cp -r adversarial-review ~/.claude/skills/
```

## Security audit
Not scanned. We audit npm-published capabilities; this one has no npm package we can resolve, or has not reached the queue. This is not a clean bill of health.

---
Measured 2026-09-13 by tashan (https://tashan.sh) from public evidence. Scorer s5.