Install one skill or module. Your agent gets Nuxt best practices, module APIs, and project-specific guidance before it changes your code.
npx nuxi module add nuxt-skill-hub
ℹ Resolved 3 module skill sources in 412ms.
✔ Generated nuxt-my-app skill for claude-code, codex
.claude/skills/nuxt-my-app/
SKILL.mdrouter
references/
nuxt/rules/data-fetching-ssr.md, plugins.md, …
vue/Vue best practices
modules/@nuxt/ui, @nuxt/content, …
7 packs routed3 module skills0 files to commit
How it works
nuxt-{package.name}Doctor
Doctor scans your project before review and flags framework bugs agents often miss.
SSR drift
public config
report ready
$pnpm dlx vite-doctor . --rules nuxt/hydration
loaded
nuxt.config.ts
scanned
38 Vue components
matched
15 Nuxt rules
nuxt/hydration/no-client-conditional-in-template
app/components/UserMenu.vue:18:5
Browser-only state changes the server-rendered markup.
nuxt/runtime/no-secret-in-public-config
nuxt.config.ts:42:13
Move private keys out of public runtime config.
suggested agent action
Apply the narrow fix, rerun Doctor, then discharge the PR with the report attached.
What's inside
General agents can write Vue, but they still miss Nuxt conventions, module APIs, and common pitfalls. Skills package that context into files your agent can read before it edits code.
A single SKILL.md sends each task to the smallest relevant guide, so agents read one pack instead of everything
Before editing, agents check whether Nuxt, Nuxt UI, or Nuxt Content already owns the surface instead of hand-rolling it
Data fetching, hydration, page meta, errors, Nitro, plugins, and performance. Each rule says when it applies and how to verify it
A sibling Vue skill covers SFC structure, reactivity, props/emits, and composables once the Nuxt decision is settled
Installed modules add their own skills from the package, its repository, or its docs site, scoped to that module
When a module ships no skill, a small generated router points agents to its official docs and source
Playground
Toggle modules to see how the skill changes as your stack changes.
Modules
2 activeConfiguration
Most projects work with the defaults. Configure skillHub in nuxt.config.ts when you need a custom skill name, explicit agent targets, extra module-author guidance, a different generation policy, or remote lookup control.
autoCustom skill name. Defaults to nuxt-{package.name}.[]Optional explicit agent targets. Leave empty to auto-detect installed agents.falseAdds module-author guidance on top of the default Nuxt app skill.prepare`prepare` writes the skill on `nuxt prepare`. `manual` writes nothing.enabledControls remote module skill lookup, concurrency, cache refresh, timing logs, timeout, and .well-known context budgets.export default defineNuxtConfig({ modules: ['nuxt-skill-hub'], skillHub: { targets: ['claude-code'], generationMode: 'prepare', remote: { refresh: false, timeoutMs: 1500, concurrency: 8, githubHeuristics: false, timings: false, }, } })
Pipeline
On every nuxi prepare, nuxt-skill-hub scans your project, resolves the best available sources, and writes agent-ready files.
Find agent folders
Claude, Cursor, Gemini, and Codex targets
Read your stack
Nuxt modules, layers, and extra packages
Check .well-known paths
Standard discovery for docs-published skills
Prefer packaged skills
Use skill files shipped with the package
Fall back to GitHub
Load the module repository when needed
Use curated mappings
Last-resort community skill map
Write agent-ready files
One entry skill plus module guidance per agent