Your team’s know‑how, installed in the AI you already use.
Reusable skills for Claude, ChatGPT, Codex, Cursor and Gemini in Antigravity. Start with the free library. Buy a specialist pack when you need one. Give your team one shared library.
14-day trial for up to 5 people. No card to start. Then $15 per seat per month.
An illustration: Claude Code, ChatGPT, Codex, Antigravity, Cursor and claude.ai each ask for a task, and each loads the matching skill from the team's Skill Me library before answering.
Illustrative stand-ins for each app. Each connects to Skill Me through MCP; prompts and answers are examples.
Skills published by the teams who build your tools.
Published in their own public repositories. Listing is not an endorsement.
- Public skills
- 2,300+
- Clients supported
- 6
- Claude Code, Cursor, Codex, claude.ai, ChatGPT, Antigravity
- Companies publishing their own skills
- 39
- Each from its own public repository. Listing is not an endorsement.
A skill is your procedure, written down.
It doesn’t make the model smarter. It tells it how your team does the work: what to check, in what order, and what a finished answer looks like.
- “Review this PR,” without a skill
- A capable review in whatever order things were noticed. Its shape changes from one review, and one reviewer, to the next.
- With Code Review Checklist
- Six passes, correctness first. Every finding labeled blocking, should-fix or nit, with a location and a fix. Then a verdict.
An illustration of the instructions’ effect, not a benchmark. Read the full skill
Team library
Example
- Code Review Checklist
- PR Description Writer
- Release checklist
- Git Commit Writer
code-review-checklist/SKILL.md
- # Code Review Checklist
- ## Operating procedure
- Run the passes in this order. Correctness comes first because a beautifully designed wrong answer is worthless…Sets the order of work, so every review starts with what the change gets wrong.
- ### Pass 1 - Correctness
- ### Pass 2 - Design and readability
- ### Pass 3 - Security
- ### Pass 4 - Performance
- ### Pass 5 - Tests
- ### Pass 6 - Report
- State an explicit verdict: approve, approve-with-nits, or request changes.Fixes the shape of the result: severity-ranked findings and a verdict.
In Claude Code: “Review this PR”
Illustrative output
Verdict: request changes. 1 blocking, 1 should-fix, 1 nit.
Blocking · session.ts:42 compares with < where the ticket says expiry is inclusive. Suggest <= and a boundary test.
Should-fix · getUser() also writes the cache. Rename it or split the write out.
Nit · unused import in session.test.ts.
Premium packs · $99 once each
Deep packs for one job, paid once.
First-party packs built to be sold, never moved from the free catalog. Each page lists every skill inside before you buy. Every pack is included with a Team plan or trial.
Creative OS
The creative system for software companies: set your brand standards, direct product photography and CGI, prompt image and video models, ship launch stills and films.
$99 onceSee the packProcedural Game Character Pipeline
Turn a written spec into rigged, LOD'd game characters and creatures in headless Blender, with every step checked against your engine budgets and a real Godot import before you hand off to an artist.
$99 onceSee the packProduction Readiness Review
Run a go/no-go launch review against your own repo: vulnerable dependencies, leaked secrets, lock-taking migrations rehearsed on a scratch Postgres, rollback, observability, config drift and a staging load test.
$99 onceSee the packFounder Finance & Metrics
Your SaaS metrics from your own Stripe and bank data: exports or a read-only key, checked complete before anything is computed, reconciled payout by payout, then MRR movements, cohorts, NRR, CAC payback.
$99 onceSee the packLLM App Evals & Guardrails
A release gate for your LLM app, run on your real traffic: a frozen eval set from your traces with PII scrubbed, tested graders, an LLM judge calibrated against your reviewers, a paired regression test that runs in CI.
$99 onceSee the packTechnical SEO & Content Engine
A technical SEO audit and content plan run against your own site and exports: a polite crawl with robots.txt obeyed, structured data checked against Google's pages, internal links, Core Web Vitals per template.
$99 onceSee the pack
One library for the whole team, instead of everyone’s private prompts.
- Code Review
- PR Writer
- Incident Write-upprivate
- Commit Messages
- Release Notesprivate
- Code Reviewpinned · 4452bca7
- PR Writerpinned · b1227f40
- Incident Write-upprivate · members only
- Commit Messagespinned earlier · newer version available
- Release Notesprivate · members only
An example team library. Hashes are the live catalog versions.
A shared team library
Skills you add load for every member whose AI is connected and signed in to Skill Me.
Pinned versions
A pinned skill stays exactly as pinned until an admin re-pins it. Admins see when a newer catalog version exists.
Private team skills
Write your own. Only members can load them; they never appear in the public catalog.
Roles and invite links
Owner, admin and member. Invite by link, see open links, revoke them.
A usage view
Owners and admins see which team skills loaded, and for how many members, over 30 days.
Premium packs included
Starting with Creative OS, $99 on its own, included while your plan or trial is active.
How it works
Connect.
Add Skill Me as a connector in Claude Code, Cursor, Codex, claude.ai, ChatGPT or Antigravity. The setup page has exact steps for each, and a check that it worked.
Ask.
When a task matches a skill in your library, your AI can load it and follow it. Or ask for one by name: “use the code review checklist.”
Keep.
Saved skills stay in your library across conversations and clients. On a team, the shared library is there for everyone, at the version you pinned.
Technical details: MCP, SKILL.md and the CLI
Skill Me is a remote MCP (Model Context Protocol) server. Clients that support remote MCP connect to https://skillme.dev/api/mcp and get tools to find, load, save and manage skills. Signing in links the connection to your account and your team.
Skills use the open SKILL.md format: YAML frontmatter (name and description) followed by markdown instructions. Public skills link to their source.
Prefer files in the repo? The CLI writes skills into .claude, .cursor or .agents folders. Read the docs.
Three ways to use it.
Start with the free library. Buy a specialist pack when you need one. Give your team one shared library.
Free library
$0
For anyone, on every supported client.
- The free public catalog. Browse, read and use every free skill and pack. Premium packs are marked with their price.
- The connector in every supported client. Claude Code, Cursor, Codex, claude.ai, ChatGPT and Antigravity.
- A personal library. Skills you save follow your account into each connected client.
- The CLI. Write skills into a project’s .claude, .cursor or .agents folder.
Premium packs
$99once
Creative OS, Procedural Game Character Pipeline, Production Readiness Review, Founder Finance & Metrics, LLM App Evals & Guardrails, Technical SEO & Content Engine. Each is bought on its own.
- Paid once. A one-time price per pack, in US dollars. No subscription.
- Yours on your account. The full instructions load in any AI connected and signed in to the account that bought it.
- Included with Teams. Members of a team with an active plan or trial get every premium pack at no extra cost.
- Sales are final. Access starts as soon as payment is confirmed, so purchases are not refundable.
Team
$15per seat / month
Billed monthly. 14-day trial for up to 5 people. No card to start. Then $15 per seat per month.
- Everything in Free, every premium pack while active, plus:
- A shared team library. Skills you add load for every member whose AI is connected and signed in to Skill Me.
- Pinned versions. A pinned skill stays exactly as pinned until an admin re-pins it. Admins see when a newer catalog version exists.
- Private team skills. Write your own. Only members can load them; they never appear in the public catalog.
- Roles and invite links. Owner, admin and member. Invite by link, see open links, revoke them.
- A usage view. Owners and admins see which team skills loaded, and for how many members, over 30 days.
From the catalog
Browse 2,300+ skills- TDD ExpertDrive development with strict Red-Green-Refactor discipline – write one failing test, write the minimum code to pass it, refactor on green – including test-list planning and a worked kata cycle.Skill Me
- PR Description WriterWrite pull request titles and descriptions from a branch diff – problem and evidence, approach, risk and blast radius, real test evidence, rollout and rollback, a reviewer guide, before/after screenshots for UI, and correct closing keywords.Skill Me
- Security AuditRuns a whole-application security audit – threat-model-lite scoping, dependencies and supply chain, secrets, authentication and authorization, input handling, transport and headers, logging, and infrastructure config.Skill Me
- Next.js App RouterBuilds and reviews Next.js App Router code – server/client component boundaries, data fetching, caching and revalidation choices, streaming with Suspense, and Server Action mutations – and delivers routes where every dynamic-vs-cached decision is explicit.Skill Me
- Frontend DesignOfficial Anthropic skill. Guidance for distinctive, intentional visual design when building new UI or reshaping an existing one, so the result does not read as a templated default.Anthropic
- Deep ResearchRuns a thorough multi-source research process – decompose the question, gather independent sources, cross-check, synthesize – and delivers a structured, cited brief with confidence levels and gaps.Skill Me
Questions people ask first
Do we have to change AI tools?
No. Skill Me connects to the clients you already use. The connector is free; your AI subscription is separate, and that provider’s plan rules still apply.
What exactly is a skill?
A plain-text SKILL.md file: a description that tells the AI when it applies, and instructions for doing the work. It is an open format, and you can read every free skill in full before you use it.
What is free and what is paid?
The public catalog, the connector and a personal library are free. Premium packs are sold once per pack; their pages show the price and a preview, and the full instructions load only for the account that has access. The Team plan is billed per seat and includes every premium pack while it, or its trial, is active.
Will everyone on the team get the same answer?
No tool can promise that: models vary. What a team library fixes is the instructions. Everyone’s AI loads the same pinned text, so the approach, structure and checks stay consistent.
Does Skill Me see our conversations or code?
It never reads your conversation, files or code. It receives only what your AI sends in tool calls, such as a skill name, a search term or a short task description for recommendations. Private team skills are stored by Skill Me and load only for members of that team.
Billing, trials and members: see the pricing questions.
“Every good prompt I ever wrote died in a chat window. Skill Me is the place they live instead.”
Start your team’s library.
Create the team, pin a few skills, invite a teammate. No card to start.