use-component
Skill1 starsChanged 4 months agoMITView on GitHub
Recommend and wire up an existing component from the user's component library. Use whenever the user asks which component fits a need, what props a component takes, or how to connect one to data/an API — e.g. "which component for a confirm dialog?", "how do I wire CommonTable to data?", "what props does CommonButton take?", "is there a toast in the library?". Read-only: it answers strictly from the design-system brain and never invents components or props.
In this repository4 other files it gives its agents.
- audit-librarySkill
- create-componentSkill
- delete-componentSkill
- edit-componentSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Always read these parts first, in order:
- Distinctive: line 45
Prefer the component whose documented purpose and props fit the need with the least glue.
- Very specific: line 61
If no component matches, say so plainly.
- Distinctive: line 68
Never invent props, defaults, or import paths.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Forge — Use Component
Prerequisite: forge.config.json exists at the repo root (run forge init once).
Advise on the existing library: pick the right component, give exact props, and hand back a working wiring snippet. Answer only from the brain so you never recommend a component or prop that doesn't exist.
The brain
Single source of truth: ${CLAUDE_PLUGIN_ROOT}/skills/_shared/DESIGN-SYSTEM.md.
Always read these parts first, in order:
- Foundations — tokens, theming (
data-theme), source layout. - Conventions — house rules (
classNameoverride hook, variant unions, sizes). - Accumulated guidance (in the brain) and
${CLAUDE_PLUGIN_ROOT}/skills/_shared/references/standing-guidance.md— these are corrections the team gave the engine; they override defaults. - The Component index table, then the matching
## <Name>section(s).
Workflow
1. Understand the need
Restate what the user is building in one line (e.g. "a destructive confirm dialog", "a data grid bound to an API"). If the ask is ambiguous between two components, ask one short question — otherwise proceed.
2. Match against the brain
- Scan the Component index for candidates by name, category, and import symbol.
- Open the candidate
## <Name>section(s). Read the Props table and Usage. - Prefer the component whose documented purpose and props fit the need with the least glue. If two fit, name both and say when to pick each.
3. Answer
Give, concisely:
- Component + import — the exact symbol and import line from the section.
- Props to set — only the props relevant to the need, with types and defaults copied from the table (don't paraphrase types).
- Wiring snippet — a
tsxblock adapted from the section's Usage, filled in for the user's case (data source, handlers). Reference design tokens by their canonical names; don't inline raw hex/px. - Caveats — if a prop's section is tagged
_(inferred — confirm against source)_, surface that: the prop is read from docs/examples, not verified source, so it may differ. Never present inferred props as guaranteed.
4. When nothing fits
If no component matches, say so plainly. Do not invent one. Offer the
create-component skill to scaffold it, or edit-component if an existing
component is close but missing a prop.
Guardrails
- Answer only from the brain. If it isn't documented there, it doesn't exist for this skill — say so.
- Never invent props, defaults, or import paths. Copy them from the section.
- Standardize imports on the canonical bare package name from
forge.config.json(the brain notes legacy/placeholder paths exist — don't reproduce them). - Surface
_(inferred)_caveats every time they apply.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

