メニュー
Design direction, UX workflows, component architecture, WCAG 2.2 accessibility, live-environment design review. Use when designing new interfaces, reviewing PR UI changes, or auditing accessibility.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SOC 職業分類に基づく
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/bpmforge/attest-claude --skill ux-expertコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Gauntlet loop — multi-agent quality harness: a lead sets a real reference bar (a named product, test suite, or baseline to match or beat), splits the goal into gradeable units, builders produce artifacts in clean context, and blind fresh-per-round critics grade each artifact against the bar with evidence; failures loop until every unit passes, two rounds stall, or budget runs out. NOT /challenge (verifies factual claims); NOT /review (one-pass verdict) — this rebuilds until the work beats something real.
Claude-Design-style visual iteration loop — render the running UI, screenshot at 375/768/1440, critique against docs/design/tokens.json and design principles, apply fixes, re-capture until clean (cap 3 iterations). Also extracts a token baseline from an existing codebase (--sync) and audits real logged-in browsers (--real). NOT a one-pass review — use /ux --review for findings-only; NOT functional conformance — use /ui-verify.
Project intelligence steward — audits CLAUDE.md / AGENTS.md alignment with actual codebase, captures session learnings, updates project docs. Use after major sessions or when docs feel stale.
| name | UX Expert |
| trigger | /ux |
| description | Design direction, UX workflows, component architecture, WCAG 2.2 accessibility, live-environment design review. Use when designing new interfaces, reviewing PR UI changes, or auditing accessibility. |
| agent | ux-engineer |
| arguments | [{"name":"task","description":"What to design or review (e.g., \"redesign settings page\", \"review PR 234\", \"audit dashboard\")","required":false},{"name":"--design","description":"Greenfield mode — produce DESIGN_PRINCIPLES.md + STYLE_GUIDE.md + UX_SPEC.md","required":false},{"name":"--review","description":"Live design review — 7-phase methodology with triage matrix (Blocker/High/Medium/Nit)","required":false},{"name":"--audit","description":"WCAG 2.2 Level AA accessibility audit only","required":false},{"name":"--flows","description":"User workflow diagrams only (no code, no style guide)","required":false}] |
Triggers the ux-engineer subagent in a forked context.
Senior UX engineer that combines Nielsen Norman methodology, Anthropic's frontend-design aesthetic principles ("no AI slop"), and Silicon-Valley-style live design review (Stripe/Airbnb/Linear standards).
Three modes:
/ux --design (greenfield) — pick a bold aesthetic direction, write design principles + style guide + UX spec. Invoked by sdlc-lead during Phase 3 for UI-bearing projects./ux --review (PR / existing UI) — 7-phase live review: Interaction → Responsive → Visual → A11y → Robustness → Code Health → Content. Uses Playwright MCP if available, falls back to script or static. Blocker/High/Medium/Nit triage./ux --audit (WCAG) — focused accessibility audit against WCAG 2.2 Level AA.Live Environment First: the agent prefers the running interface over static source. It will use Playwright MCP if installed, or write a Playwright script, or fall back to static analysis (and explicitly note the downgrade).
Anti-AI-slop rules: never Inter/Roboto/Arial, never purple-gradient-on-white, pick an extreme aesthetic direction (not middle-ground). Reference: references/design-review-checklist.md.
Outputs:
--design → docs/design/DESIGN_PRINCIPLES.md, docs/design/STYLE_GUIDE.md, docs/design/UX_SPEC.md--review → docs/UX_REVIEW.md + screenshots--audit → docs/ACCESSIBILITY_AUDIT.md