# Gate Visual

> The self-critique loop that separates working from first-grade — screenshots every page light and dark via the pixel-qa subagent, delegates scoring to the design-judge subagent against design/DIRECTION.md and the taste rubric (hierarchy, typography, spacing, color, distinctiveness, craft — 1–10 each), fixes the worst-ranked defect, re-shoots, and repeats — minimum 2 full rounds, stopping only on a SHIP verdict. Invoke in Phase 11 of the ultraweb pipeline after gate-code and gate-responsive are green, after any major visual change to an ultraweb site, or when the user says "is it actually good", "critique the design", "judge the site", or "run the visual gate". Writes per-round scores, defects fixed, and screenshot paths to design/QA.md.

## Facts
- Page: https://tashan.sh/capability/skill-blyatiful1-gate-visual
- tashan id: skill:blyatiful1/gate-visual
- Source: https://github.com/blyatiful1/ultraweb
- Type: skill
- Category: design
- tashan score: not scored (catalogued only — too little public evidence)
- Adoption: 9.0
- Upkeep: 99.0
- Freshness: 97.0
- 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 gate-visual ~/.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 2026-08-23 by tashan (https://tashan.sh) from public evidence. Scorer s5.