# Builder Spec Judge Governance

> Internal governance (not user-invocable): the Builder bar for react-component, spec-judge face. Grading whether the .feature COVERS each discipline's concern. Loaded at the spec gate.

- Skill: `cyberuni/builder-spec-judge-governance` (Agent Skill)
- Install (CLI): `npx skillmds@latest add cyberuni/builder-spec-judge-governance`
- Raw SKILL.md: https://api.skillmd.com/api/skills/cyberuni/builder-spec-judge-governance/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Web & Frontend
- Author: cyberuni (https://skillmd.com/u/cyberuni)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/cyberuni/builder-spec-judge-governance

---


# builder · spec · judge — grade the suite's coverage, by discipline

The **Builder** bar at the **spec gate**, **backward** face: grade the `.feature` for *coverage* of
each discipline and boolean testability. A concern with no scenario is a gap → fail. Same sections as
the spec-producer, read as **checks**.

## engineer

- Each scenario asserts an observable outcome (no "renders correctly"); typed surface is exercised.

## designer

- Every visual state has a scenario; none unspecced.

## a11y

- Role, keyboard, accessible name, and `axe` each have a scenario; no a11y concern unscenarioed.

## security

- Each untrusted-input path has a sanitization/inert scenario.

## qa

- Empty/error/boundary scenarios present; happy-path-only is a coverage gap.