Bake judgment into the agent — briefs, viewports, logos, harness, ghosts, launch checks.
Install once, then invoke by name. Judgment for briefs, viewports, logos, harness, ghosts, and launch checks — without re-explaining context every session.
Agents don't have design judgment. They ship a page at 1280 and call it done while it breaks at 390. They "clean up" an optically-aligned logo row into a grid that looks worse. And every new session, you re-type the same context you typed yesterday.
These skills bake that judgment — and your context — into the agent.
npx skills@latest add iruhdam1/skillsThen invoke a skill by name. Source on GitHub.
Bake your context into a brief once. Never re-explain it again.
Mobile, tablet, and desktop screenshots plus a pass/warn/fail layout report.
ESLint for design. A judged QA pass with a fixed check catalog.
Align logo rows optically, not mathematically.
Score harness readiness, pick tasks, set up only what's missing. Also on /harness.
Skeleton UI with --ghost-* roles. Twin of the Tiny Design Shop tool.
Pre-launch SEO, OpenGraph, agent readiness, AEO. Twin of the launch checklist tool.
More skills land here as companions ship. Pack source: iruhdam1/skills.