wu-uxr-002-card-studio

Execute one slice of WU-UXR-002, the Card Studio UX remediation from the 2026-10-11 field study (bidi-safe text, truthful feedback, no lost cards, legible credits, mobile reachability, consistent publ

SKILL.md

GitHub

WU-UXR-002 — Card Studio UX Remediation

The canonical execution skill is:

skills/wu-uxr-002-card-studio/SKILL.md

Read that file in full before coding. Then follow its mandatory read order, gates and slice discipline.

Sources

specs/WU-UXR-002-card-studio-ux-remediation.md
specs/WU-UXR-002-IMPLEMENTATION-CHECKLIST.md
docs/audits/urdu-card-studio-ux-2026-10-11/   (findings CARD-UX-001…039, evidence, harness)

Non-negotiable rules (summary)

  • One slice per branch (claude/wu-uxr-002<letter>-<slug>). Stop after the slice.
  • Write failing tests first. Register them in scripts/run-contract-tests.js, playwright.config.js and .github/workflows/quality.yml.
  • Decision-line slices (H, L, M, N) and decision-gated slices (I: D-UX2-2, F homepage: D-UX2-3, O: D-UX2-1) need the decision recorded first. If it isn't, ask the user.
  • Slice O (mobile hierarchy) requires production evidence. Never implement it speculatively.
  • Every new string in English and natural Urdu. Run npm run locale:generate after HTML changes.
  • Bump sw.js CACHE_NAME and its test pins when APP_SHELL files change.
  • No user-uploaded images in remix payloads, no publishing without consent, content-free telemetry only.
  • Keep data-* hooks stable, or update all specs and the audit harness together.