Agent skills I use for design work with Claude Code and other coding agents. Each one is a small folder: a SKILL.md the agent reads, plus the files it needs to do the job well.

Two routes. Pick one: installing both gives you every skill twice.

Claude Code plugin. Installs the set as one managed bundle that updates when this repo does.

claude plugin marketplace add SkylarKitchen/skills
claude plugin install skylarkitchen-skills@skylarkitchen

Or, from inside a session:

/plugin marketplace add SkylarKitchen/skills
/plugin install skylarkitchen-skills@skylarkitchen

skills.sh, for any agent. Copies the skill files into your project as ordinary files you own and can edit.

npx skills@latest add SkylarKitchen/skills

Model-invoked skills run when you type them or when the agent sees a task that fits. User-invoked skills run only when you type them. Everything here is model-invoked so far.

Skills for visual work: drawings, diagrams, and interactive pages.

Model-invoked

• 3d-assembly-manual: Build a phone-first, step-by-step 3D build guide (LEGO-manual pacing, BIG-diagram look) as a single Three.js HTML file, from a starter engine you only configure.
• launch-film: Make a short, frame-exact product film for X or a portfolio from the real app, site or Three.js scene: an authored camera over the real surface, rendered on a canvas engine instead of cut from a screen recording.
• room-from-photos: Turn a few phone photos of an existing room into a near-photoreal 3D model and plan, with furniture layouts to compare, a library of pieces to add, fit checks, and a gizmo to drag pieces around, as a single Three.js HTML file.

MIT