Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
.claude/skills/asymmetric-al-find-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 122% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 126% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 146% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 69% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 78% | 0% |
This skill helps you discover and install skills from the open agent skills ecosystem.
Use this skill when the user:
The Skills CLI (npx skills) is the package manager for the open agent skills ecosystem. Skills are modular packages that extend agent capabilities with specialized knowledge, workflows, and tools.
Key commands:
npx skills find [query] - Search for skills interactively or by keywordnpx skills add <package> - Install a skill from GitHub or other sourcesnpx skills update / npx skills upgrade - Update installed skillsnpx skills experimental_install - Restore project skills from skills-lock.jsonDo not use npx skills check as a read-only update check in this repo. With skills@1.5.7, that subcommand is not listed in npx skills --help and was observed to rewrite .agents/skills/* and skills-lock.json. Treat it like update: run it only when you intentionally want to refresh installed skills and are prepared to review and revert the full diff.
Browse skills at: https://skills.sh/
Before searching the public index, check docs/ai/skills/ and AGENTS.md → Skill Routing. Canonical repo-owned skills live under docs/ai/skills/ and are mirrored to .agents/skills/, .cursor/skills/, and .claude/skills/ with bun run skills:sync (CI: bun run skills:verify).
Use the canonical docs path when routing or documenting repo-critical skills. Treat .agents/skills/, .cursor/skills/, and .claude/skills/ as runtime mirrors for tool compatibility, not the primary authoring location.
Some extra ecosystem or tool-specific skills may exist only in .cursor/skills/ or .agents/skills/. Those are optional unless the repo explicitly promotes them into docs/ai/skills/.
Example — Supabase: platform work is covered by docs/ai/skills/supabase/SKILL.md and Postgres tuning by docs/ai/skills/supabase-postgres-best-practices/SKILL.md. Install or refresh from supabase/agent-skills with npx skills add supabase/agent-skills -y, then bun run skills:refresh-upstream, bun run skills:sync, and bun run skills:verify. skills-lock.json records CLI pins; use targeted npx skills add <source> -y refreshes before falling back to npx skills experimental_install -y, because the latter rewrites all lockfile entries.
Example — Tiptap: rich text work is covered by docs/ai/skills/tiptap/SKILL.md. To compare against upstream: npx skills add ueberdosis/tiptap (see also Tiptap agent skill).
Example — npm dependency cleanup: reducing unused or heavy JS dependencies across npm, pnpm, Yarn, or Bun is covered by docs/ai/skills/npm-deps-cleanup/SKILL.md. Install or refresh from anthonyshew/dotfiles with npx skills add anthonyshew/dotfiles -y, then bun run skills:refresh-upstream, bun run skills:sync, and bun run skills:verify.
Example — Resend CLI: terminal, automation, and CI work is covered by docs/ai/skills/resend-cli/SKILL.md. It is vendored from resend/resend-cli (skills/resend-cli/ at tag v2.0.0+); maintainer refresh notes are in docs/ai/skills/resend-cli/references/upstream.md.
Example — Resend platform skills: SDK, webhooks, inbound, and send/receive platform workflows are a separate concern. Optional ecosystem install resend/resend-skills may live under .agents/skills/resend/ when added with the Skills CLI.
Example — Resend app integration: app-level Resend integration in this monorepo is documented in docs/guides/features/resend-integration.md.
Example — bendc frontend guidelines: semantic HTML, CSS discipline, and vanilla JS readability patterns from bendc/frontend-guidelines are covered by docs/ai/skills/bendc-frontend-guidelines/SKILL.md (vendored README.md under references/); maintainer refresh notes are in docs/ai/skills/bendc-frontend-guidelines/references/upstream.md. Use docs/ai/rules/frontend.md first for apps/* and packages/ui work.
Example — Emil Kowalski design-engineering pack: the five skills from emilkowalski/skills are canonical under docs/ai/skills/{animation-vocabulary,apple-design,emil-design-eng,improve-animations,review-animations}/. Refresh with npx --yes skills@latest add emilkowalski/skills -y, then bun run skills:refresh-emilkowalski, bun run skills:sync, and bun run skills:verify. Follow the narrow routes and Core precedence in AGENTS.md rather than treating the installer output as authoritative.
Example — Grill for Unknowns: the complete nicobailon/grill-for-unknowns plugin is canonical under docs/ai/skills/grill-for-unknowns/. Refresh with npx --yes skills@latest add nicobailon/grill-for-unknowns -y, then bun run skills:refresh-grill-for-unknowns, bun run skills:sync, and bun run skills:verify. It is explicit-only in Core: use it for a requested map-vs-territory, blindspot, unknown-known prototype, or launch-packet workflow; generic grilling stays with grilling or grill-with-docs.
Example — Payload CMS:
docs/ai/skills/payloadcms-payload/SKILL.md for collections, fields, hooks, access, queries, adapters, plugins, and Payload app behavior.docs/ai/skills/payloadcms-cms-migration/SKILL.md for CMS to Payload content-model migration.payloadcms/skills (skills/payload/, skills/cms-migration/).docs/ai/skills/payloadcms-payload/references/upstream.md and docs/ai/skills/payloadcms-cms-migration/references/upstream.md.npx skills add payloadcms/skills.docs/ai/rules/backend.md for server and data boundaries.Example — Inngest agent tooling: durable workflow and agent integration work is covered by official skills vendored under docs/ai/skills/inngest-*, with docs/ai/skills/inngest/SKILL.md as the router. Use docs/ai/skills/inngest-brownfield-audit/SKILL.md before changing existing app workflows, and use docs/ai/skills/inngest-setup/SKILL.md only when explicitly adding product runtime Inngest. Refresh with bun run skills:refresh-inngest, then bun run skills:sync and bun run skills:verify.
When a user asks for help with something, identify:
Run the find command with a relevant query:
bashnpx skills find [query]
For example:
npx skills find react performancenpx skills find pr reviewnpx skills find changelogThe command will return results like:
Install with npx skills add <owner/repo@skill>
vercel-labs/agent-skills@vercel-react-best-practices
└ https://skills.sh/vercel-labs/agent-skills/vercel-react-best-practicesWhen you find relevant skills, present them to the user with:
Example response:
I found a skill that might help! The "vercel-react-best-practices" skill provides
React and Next.js performance optimization guidelines from Vercel Engineering.
To install it:
npx skills add vercel-labs/agent-skills@vercel-react-best-practices
Learn more: https://skills.sh/vercel-labs/agent-skills/vercel-react-best-practicesIf the user wants to proceed, you can install the skill for them:
bashnpx skills add <owner/repo@skill> -g -y
The -g flag installs globally (user-level) and -y skips confirmation prompts.
When searching, consider these common categories:
| Category | Example Queries | | --------------- | ---------------------------------------- | | Web Development | react, nextjs, typescript, css, tailwind | | Testing | testing, jest, playwright, e2e | | DevOps | deploy, docker, kubernetes, ci-cd | | Documentation | docs, readme, changelog, api-docs | | Code Quality | review, lint, refactor, best-practices | | Design | ui, ux, design-system, accessibility | | Productivity | workflow, automation, git |
vercel-labs/agent-skills or ComposioHQ/awesome-claude-skillsIf no relevant skills exist:
npx skills initExample:
I searched for skills related to "xyz" but didn't find any matches.
I can still help you with this task directly! Would you like me to proceed?
If this is something you do often, you could create your own skill:
npx skills init my-xyz-skilldocs/ai/skills/ and Skill Routing in AGENTS.md before suggesting npx skills finddocs/ai/skills/ and bun run skills:sync when mirrors matter| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 10,079 | 5,989 | -41% | 1 | 1 | 0% | 1,876 | 4,170 | +122% | 0 | 0 | — |
case-08 | fail→pass | 7,864 | 2,460 | -69% | 1 | 1 | 0% | 1,505 | 3,397 | +126% | 0 | 0 | — |
case-01 | fail→fail | 13,658 | 5,806 | -57% | 1 | 1 | 0% | 2,130 | 3,220 | +51% | 0 | 0 | — |
case-02 | fail→fail | 15,777 | 4,449 | -72% | 1 | 1 | 0% | 2,741 | 3,081 | +12% | 0 | 0 | — |
case-03 | fail→fail | 14,746 | 4,961 | -66% | 1 | 1 | 0% | 2,648 | 3,122 | +18% | 0 | 0 | — |
case-04 | pass→pass | 6,823 | 6,305 | -8% | 1 | 1 | 0% | 1,014 | 3,851 | +280% | 0 | 0 | — |
case-05 | pass→pass | 11,617 | 8,627 | -26% | 1 | 1 | 0% | 2,543 | 4,635 | +82% | 0 | 0 | — |
case-06 | pass→pass | 7,509 | 8,274 | +10% | 1 | 1 | 0% | 1,426 | 4,254 | +198% | 0 | 0 | — |
case-09 | fail→pass | 8,274 | 4,581 | -45% | 1 | 1 | 0% | 1,573 | 3,876 | +146% | 0 | 0 | — |
case-10 | pass→pass | 7,838 | 3,912 | -50% | 1 | 1 | 0% | 1,374 | 3,697 | +169% | 0 | 0 | — |
case-11 | pass→pass | 15,635 | 2,984 | -81% | 1 | 1 | 0% | 1,420 | 3,351 | +136% | 0 | 0 | — |
case-12 | fail→pass | 12,275 | 3,659 | -70% | 1 | 1 | 0% | 2,159 | 3,649 | +69% | 0 | 0 | — |
case-13 | fail→pass | 11,311 | 3,560 | -69% | 1 | 1 | 0% | 1,980 | 3,517 | +78% | 0 | 0 | — |
case-14 | fail→pass | 8,448 | 2,121 | -75% | 1 | 1 | 0% | 1,583 | 3,299 | +108% | 0 | 0 | — |
case-15 | fail→pass | 8,593 | 3,259 | -62% | 1 | 1 | 0% | 1,322 | 3,529 | +167% | 0 | 0 | — |
case-16 | fail→pass | 11,421 | 4,617 | -60% | 1 | 1 | 0% | 1,975 | 3,812 | +93% | 0 | 0 | — |
case-17 | fail→pass | 9,301 | 3,556 | -62% | 1 | 1 | 0% | 1,518 | 3,539 | +133% | 0 | 0 | — |
case-18 | fail→pass | 12,471 | 1,774 | -86% | 1 | 1 | 0% | 2,508 | 3,189 | +27% | 0 | 0 | — |
case-19 | fail→pass | 10,781 | 3,053 | -72% | 1 | 1 | 0% | 1,747 | 3,401 | +95% | 0 | 0 | — |
case-20 | pass→pass | 3,943 | 2,383 | -40% | 1 | 1 | 0% | 673 | 3,286 | +388% | 0 | 0 | — |
case-21 | fail→pass | 9,299 | 5,385 | -42% | 1 | 1 | 0% | 1,576 | 3,782 | +140% | 0 | 0 | — |
case-22 | fail→fail | 6,683 | 1,857 | -72% | 1 | 1 | 0% | 1,024 | 3,185 | +211% | 0 | 0 | — |
case-23 | fail→pass | 5,323 | 2,871 | -46% | 1 | 1 | 0% | 860 | 3,360 | +291% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 20 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +57 percentage points is the difference between those two pass rates over the 20 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.