SECURITY REVIEW
Not yet assessed
Review the original instructions and requested permissions before installing.
No security review is available for this catalog entry yet.
Frontend · Design / AGENT SKILL
ui
0
Review a specified UI scenario or provide one bounded improvement proposal.
Review the named UI scenario or emit one bounded improvement packet.
Understand the trade-offs.
SKILL QUALITY
Not yet assessed
How clearly the skill guides your agent, how complete its workflow is, and how you can check the outcome.
No quality assessment is available for this catalog entry yet.
The full skill.
Original instructions from the publisher’s SKILL.md
# UI Read `docs/procedures/visual-review.md` before acting; it owns commands, hard-gate authority, critic bounds, packet rules, providers, and owner handoff. Also follow `docs/procedures/MODEL-CARD.md`. When the registered review needs a screenshot/video/HTML proof bundle or an iterative evidence handoff, read and follow [`visual-report-bundle.md`](visual-report-bundle.md); it is an internal reference of this skill, not a separately invocable skill. Accept only explicit `review <registered-spec>` or `improve <registered-spec>`. Resolve the exact name through the repository registry; reject URLs, paths, configs, commands, and unknown names. - `review` is read-only and reusable by reviewer agents for changed frontend components or behavior; it leaves a bounded spec-defined report. - `improve` creates one checkout-bound execution packet; validate it, then invoke `/skill:exec <run>/execution-packet.json` exactly once. Never apply packet fixes here, recurse into `improve`, auto-run provider hooks, or treat advisory critic scores as hard gates.