# Adversarial Quality Gate

> Decide whether an article deserves publication through hard checks and skeptical comparison.

- Skill: `lionelndong/adversarial-quality-gate` (Agent Skill)
- Install (CLI): `npx skillmds add lionelndong/adversarial-quality-gate`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lionelndong/adversarial-quality-gate/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: lionelndong (https://skillmd.com/u/lionelndong)
- Updated: 2026-08-19
- Page: https://skillmd.com/skills/lionelndong/adversarial-quality-gate

---


# Adversarial Quality Gate

## Mechanical checks

Assess depth against the Beat Spec, consensus coverage, AI-tell patterns, evidence, and structure. Score is diagnostic only.

## Editorial verdict

Review the article beside the leading result and strongest plausible alternative through three lenses: substance/competitiveness, voice/readability, and reader-intent satisfaction. Ask: **which page would the reader clearly keep?** A competitor win, a tie, or material uncertainty fails.

## Pass conditions

- score is at least 85;
- no critical issue or failed hard gate;
- product proof, author fit, citations, internal links, and visual proof pass;
- no unsupported, deceptive, unsafe, or off-intent framing.

## Failure handling

Return an exact punch list and restart stage. One targeted rewrite is allowed. After two failures at the same producing gate, quarantine rather than iterating indefinitely.