{"description":"Agent skills listed on CookiesForAI, with the facts we checked and the date we checked them.","license":"Our text and data: CC BY 4.0, credit cookiesforai.app. Each listed item keeps its own license.","generated":"2026-10-03","sources":[{"repo":"anthropics/skills","owner":"anthropics","publisher":"Anthropic","publisherUrl":"https://www.anthropic.com","official":true,"trust":"Anthropic's own repository of Agent Skills, the skill format it created for Claude.","stars":179326,"license":"Apache-2.0 for most skills; docx, pdf, pptx and xlsx are proprietary","lastPush":"2026-09-29","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"antonbabenko/terraform-skill","owner":"antonbabenko","publisher":"Anton Babenko","publisherUrl":"https://github.com/antonbabenko","official":false,"trust":"Listed under trust rule C: on 2026-10-01 Gemini, Grok and Claude each named this repository by its owner in most of their runs of our question about reviewing Terraform plans. It is Anton Babenko's personal project, not HashiCorp's.","stars":2394,"license":"GitHub reports NOASSERTION; the LICENSE file and the SKILL.md say Apache-2.0","lastPush":"2026-07-03","archived":false,"aiNamed":{"question":"Is there an agent skill that helps my coding agent review Terraform plans?","date":"2026-10-01","makers":[{"maker":"Google","surface":"Gemini app, Gemini 3.6 Flash, temporary chats","runsTotal":3,"runsNamed":3,"quotes":["names: terraform-skill (by Anton Babenko) \"the standard open-source skill\"","names: terraform-skill (by Anton Babenko), terraform-infrastructure-as-code / HashiCorp Agent Skills","names: terraform-skill (by Anton Babenko), custom .claude/skills / SKILL.md protocol skill"]},{"maker":"xAI","surface":"Grok, Fast model, private chats","runsTotal":3,"runsNamed":3,"quotes":["antonbabenko/terraform-skill, daymade/claude-code-skills terraform-skill","names: HashiCorp agent-skills terraform-style-guide (official, reviews HCL not plans), antonbabenko/terraform-skill","antonbabenko/terraform-skill, (further code-review skills in table, truncated)"]},{"maker":"Anthropic","surface":"Claude Code CLI: Opus 5.5, Sonnet 5.5 and Fable 5.1, one run from memory and two with web search each","runsTotal":9,"runsNamed":6,"quotes":["**`antonbabenko/terraform-skill`** is a community skill","**terraform-skill** by Anton Babenko (Apache 2.0)","Anton Babenko's `terraform-skill`","[antonbabenko/terraform-skill](https://explainx.ai/skills/antonbabenko/terraform-skill/terraform-skill) is a general Terraform best-practices skill","**[antonbabenko/terraform-skill](https://explainx.ai/skills/antonbabenko/terraform-skill/terraform-skill)** covers Terraform best practices in general","[terraform-skill by Anton Babenko](https://agenticskills.io/skills/terraform-skill) works with Claude Code"]}]},"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T02:06:05+02:00"},{"repo":"bradautomates/claude-video","owner":"bradautomates","publisher":"bradautomates","publisherUrl":"https://github.com/bradautomates","official":false,"trust":"Open-source video skill for Claude Code and other agents.","stars":17914,"license":"MIT","lastPush":"2026-09-25","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"cloudflare/skills","owner":"cloudflare","publisher":"Cloudflare","publisherUrl":"https://www.cloudflare.com","official":true,"trust":"Cloudflare's own skills, published from the cloudflare GitHub organization and bundled with Cloudflare's MCP server in its Claude Code and Codex plugins.","stars":2965,"license":"Apache-2.0","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"coreyhaines31/marketingskills","owner":"coreyhaines31","publisher":"Corey Haines","publisherUrl":"https://corey.co","official":false,"trust":"Marketing skills by SaaS marketer Corey Haines, with 52,097 GitHub stars on 2026-10-01.","stars":52171,"license":"MIT","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"expo/skills","owner":"expo","publisher":"Expo","publisherUrl":"https://expo.dev","official":true,"trust":"The Expo team's own agent skills, linked from Expo's documentation at docs.expo.dev/skills.","stars":2649,"license":"MIT","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"firebase/agent-skills","owner":"firebase","publisher":"Firebase (Google)","publisherUrl":"https://firebase.google.com","official":true,"trust":"The Firebase team's own skills, published from the firebase GitHub organization and also packaged as a Gemini CLI extension.","stars":462,"license":"Apache-2.0","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"firecrawl/cli","owner":"firecrawl","publisher":"Firecrawl","publisherUrl":"https://www.firecrawl.dev","official":true,"trust":"Firecrawl's own command-line tool and its skills, published from the company's GitHub account and documented at docs.firecrawl.dev/cli.","stars":643,"license":"GitHub reports none and the repo has no license file; package.json declares ISC","lastPush":"2026-10-02","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"getsentry/skills","owner":"getsentry","publisher":"Sentry","publisherUrl":"https://sentry.io","official":true,"trust":"Sentry's own engineering skills, published from the getsentry GitHub organization; its README says they are written for Sentry employees.","stars":1031,"license":"Apache-2.0 (the security-review skill folder carries its own CC-BY-SA-4.0 LICENSE for OWASP-derived material)","lastPush":"2026-09-29","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"googleworkspace/cli","owner":"googleworkspace","publisher":"Google Workspace","publisherUrl":"https://developers.google.com/workspace","official":true,"trust":"Published from Google Workspace's GitHub account, with 31,223 stars on 2026-10-01. Its README says it is not an officially supported Google product.","stars":31231,"license":"Apache-2.0","lastPush":"2026-09-24","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"hashicorp/agent-skills","owner":"hashicorp","publisher":"HashiCorp","publisherUrl":"https://www.hashicorp.com","official":true,"trust":"HashiCorp's own Terraform and Packer skills, published from the hashicorp GitHub organization; its README says only HashiCorp-internal contributors may contribute.","stars":881,"license":"MPL-2.0","lastPush":"2026-09-28","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"heygen-com/hyperframes","owner":"heygen-com","publisher":"HeyGen","publisherUrl":"https://hyperframes.heygen.com","official":true,"trust":"HeyGen's own open-source video framework and its skills, published from HeyGen's GitHub account, with 55,056 stars on 2026-10-01.","stars":55348,"license":"Apache-2.0","lastPush":"2026-10-02","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"huggingface/skills","owner":"huggingface","publisher":"Hugging Face","publisherUrl":"https://huggingface.co","official":true,"trust":"Hugging Face's own agent skills, published from its GitHub account; the hf-cli skill is also listed in Cursor's marketplace and the Codex plugins directory.","stars":11120,"license":"Apache-2.0","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"kepano/obsidian-skills","owner":"kepano","publisher":"Steph Ango","publisherUrl":"https://stephango.com","official":false,"trust":"Skills for Obsidian files by Steph Ango (kepano), whose GitHub profile lists Obsidian as his company, published from his own GitHub account, with 49,059 GitHub stars on 2026-10-01.","stars":49077,"license":"MIT","lastPush":"2026-09-15","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:06:34+02:00"},{"repo":"Leonxlnx/taste-skill","owner":"leonxlnx","publisher":"Leon Lin","publisherUrl":"https://tasteskill.dev","official":false,"trust":"Open-source frontend design skills with 91,679 GitHub stars on 2026-10-01.","stars":91799,"license":"MIT","lastPush":"2026-09-26","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"mattpocock/skills","owner":"mattpocock","publisher":"Matt Pocock","publisherUrl":"https://www.aihero.dev/skills","official":false,"trust":"Personal skills of TypeScript educator Matt Pocock, with 273,468 GitHub stars on 2026-10-01 and a plugin in Claude Code's official marketplace.","stars":273901,"license":"MIT","lastPush":"2026-09-29","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"microsoft/azure-skills","owner":"microsoft","publisher":"Microsoft","publisherUrl":"https://azure.microsoft.com","official":true,"trust":"Microsoft's Azure Skills Plugin, published from the microsoft GitHub organization and also installable from Anthropic's official Claude Code plugin marketplace as azure@claude-plugins-official.","stars":1522,"license":"MIT","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"microsoft/playwright-cli","owner":"microsoft","publisher":"Microsoft (Playwright team)","publisherUrl":"https://playwright.dev","official":true,"trust":"Published by Microsoft from its own GitHub account; Playwright CLI is the command-line companion to Microsoft's Playwright testing framework.","stars":13721,"license":"Apache-2.0","lastPush":"2026-09-28","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"neondatabase/agent-skills","owner":"neondatabase","publisher":"Neon","publisherUrl":"https://neon.com","official":true,"trust":"Neon's own skills, published from the neondatabase GitHub organization alongside its Claude Code, Cursor and Kimi Code plugins.","stars":98,"license":"Apache-2.0","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"nextlevelbuilder/ui-ux-pro-max-skill","owner":"nextlevelbuilder","publisher":"Next Level Builder","publisherUrl":"https://www.uupm.cc","official":false,"trust":"Open-source UI and UX design skill with 132,210 GitHub stars on 2026-10-01 and its own npm installer, ui-ux-pro-max-cli.","stars":132343,"license":"MIT","lastPush":"2026-09-27","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"obra/superpowers","owner":"obra","publisher":"Jesse Vincent","publisherUrl":"https://github.com/obra","official":false,"trust":"Open-source skills framework, also listed in Anthropic's official Claude Code plugin marketplace.","stars":293976,"license":"MIT","lastPush":"2026-09-27","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"openai/skills","owner":"openai","publisher":"OpenAI","publisherUrl":"https://openai.com","official":true,"trust":"OpenAI's own skills catalog for Codex, published from its GitHub account. Its README marks the repository deprecated since 2026-06-22 in favour of openai/plugins.","stars":27839,"license":"GitHub reports none; each skill folder carries its own LICENSE.txt (Apache-2.0 for pdf, gh-fix-ci and gh-address-comments)","lastPush":"2026-09-08","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"pbakaus/impeccable","owner":"pbakaus","publisher":"Paul Bakaus","publisherUrl":"https://impeccable.style","official":false,"trust":"Open-source design skill by Paul Bakaus, creator of jQuery UI, with 73,322 GitHub stars on 2026-10-01.","stars":73679,"license":"Apache-2.0","lastPush":"2026-10-02","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"prisma/skills","owner":"prisma","publisher":"Prisma","publisherUrl":"https://www.prisma.io","official":true,"trust":"Prisma's own skills for Prisma ORM, published from the prisma GitHub organization.","stars":68,"license":"MIT","lastPush":"2026-09-30","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-01T15:29:34+02:00"},{"repo":"remotion-dev/skills","owner":"remotion-dev","publisher":"Remotion","publisherUrl":"https://www.remotion.dev","official":true,"trust":"Remotion's own agent skills, linked from its documentation at remotion.dev/docs/ai/skills.","stars":4803,"license":"GitHub reports none and the repo has no license file; Remotion itself is under the Remotion License (free for individuals and small companies, paid company license otherwise)","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"shadcn-ui/ui","owner":"shadcn-ui","publisher":"shadcn/ui","publisherUrl":"https://ui.shadcn.com","official":true,"trust":"The shadcn/ui project's own repository; its documentation at ui.shadcn.com/docs/skills tells readers to install this skill.","stars":124973,"license":"MIT","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"stripe/ai","owner":"stripe","publisher":"Stripe","publisherUrl":"https://stripe.com","official":true,"trust":"Stripe's own repository for AI tooling, published from its GitHub account and linked from docs.stripe.com/agents.","stars":1852,"license":"MIT","lastPush":"2026-10-02","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"supabase/agent-skills","owner":"supabase","publisher":"Supabase","publisherUrl":"https://supabase.com","official":true,"trust":"Supabase's own skills, linked from its documentation.","stars":2671,"license":"MIT","lastPush":"2026-09-28","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-01T15:18:36+02:00"},{"repo":"trailofbits/skills","owner":"trailofbits","publisher":"Trail of Bits","publisherUrl":"https://www.trailofbits.com","official":true,"trust":"Security research firm Trail of Bits publishes these skills from its own GitHub organization as a Claude Code plugin marketplace.","stars":7328,"license":"CC-BY-SA-4.0","lastPush":"2026-09-28","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"vercel-labs/agent-browser","owner":"vercel-labs","publisher":"Vercel","publisherUrl":"https://vercel.com","official":true,"trust":"Vercel Labs' browser automation CLI for agents, with 43,427 GitHub stars on 2026-10-01. Vercel also makes the skills CLI.","stars":43444,"license":"Apache-2.0","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"vercel-labs/agent-skills","owner":"vercel-labs","publisher":"Vercel","publisherUrl":"https://vercel.com","official":true,"trust":"Vercel's official collection of agent skills. Vercel also makes the skills CLI and skills.sh.","stars":31813,"license":"MIT, stated in the README (no license file)","lastPush":"2026-08-28","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"repo":"wshobson/agents","owner":"wshobson","publisher":"Seth Hobson","publisherUrl":"https://github.com/wshobson","official":false,"trust":"Independent plugin marketplace with 184 skills and 40,127 GitHub stars on 2026-10-01, above the 5,000 star bar.","stars":40142,"license":"MIT","lastPush":"2026-10-01","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"}],"skills":[{"url":"https://cookiesforai.app/skills/anthropics/academy-guide","owner":"anthropics","slug":"academy-guide","name":"Academy Guide","skillName":"academy-guide","title":"academy-guide skill: Claude Academy courses for learning Claude","summary":"Academy Guide is Anthropic's agent skill that ends Claude's answers about using Claude with one or two matching Claude Academy courses or tutorials, and only when the match is strong.","category":"learning-ai-assistants","source":"anthropics/skills","path":"skills/academy-guide","license":"Apache-2.0","worksWith":["Claude Code","Claude API"],"agentNote":"Academy Guide is about Claude itself: it recommends Claude Academy material for questions about Claude products, so it only makes sense in Claude. It needs web access to fetch the Academy catalog.","cliChecked":"2026-10-01","whatItDoes":["Academy Guide is Anthropic's skill for pointing people to Claude Academy (academy.claude.com), Anthropic's learning hub of courses, tutorials and use cases. It loads when someone asks how to use Claude or a Claude product (how do I, getting started with, what can Claude do), asks about artifacts, projects, skills, plugins, connectors or MCP, or wants training or onboarding material for a team, class or organization.","With Academy Guide loaded, Claude answers the question first, then adds at most two Academy items, usually one, each with a one-sentence description. It recommends only on a strong match, meaning the person wants to learn a feature, not finish a task in progress. When the match is weak, it says nothing about Academy. Items that need an Academy sign-in are marked as such.","Academy Guide holds no list of courses. Claude fetches the Academy catalog, a JSON file on academy.claude.com, once per conversation and copies item links from it exactly, treating the file as data rather than instructions. If the catalog cannot be fetched or is out of date, Claude names no specific course and points to a product hub (Claude, Claude Code, Claude Cowork, AI Fluency or the developer platform) or the resources library instead."],"useItWhen":["You ask Claude how a feature such as projects, skills or connectors works and want a course to go deeper.","You are rolling Claude out to a team or a class and need onboarding material.","You want Claude's learning links to point to real Academy pages, never made-up ones."],"notFor":["You use a different assistant, such as ChatGPT or Gemini: every link points to courses about Claude products."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the academy-guide plugin","command":"/plugin install academy-guide@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill academy-guide"}],"sourceUpdated":"2026-08-18","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/agent-browser","owner":"vercel-labs","slug":"agent-browser","name":"agent-browser","skillName":"agent-browser","title":"agent-browser skill: browser automation CLI for AI agents","summary":"agent-browser is the agent skill for Vercel Labs' browser automation CLI: a native Rust tool that lets an agent open pages, read accessibility snapshots and click, fill and screenshot by element reference.","category":"browser-and-web","source":"vercel-labs/agent-browser","path":"skills/agent-browser","license":"Apache-2.0","worksWith":["Claude Code","Codex","Cursor","Gemini CLI","GitHub Copilot","Goose","OpenCode","Windsurf"],"agentNote":"The SKILL.md is a stub: it tells the agent to run agent-browser skills get core and load the real instructions from the installed CLI, so the agent needs shell access and the CLI on its path.","cliChecked":"2026-10-01","whatItDoes":["agent-browser is the skill that ships with Vercel Labs' browser CLI of the same name. The CLI drives Chrome or Chromium over the DevTools protocol without Playwright or Puppeteer. The agent opens a page, takes an accessibility-tree snapshot, and acts on short references such as @e2 to click, fill, read text or take screenshots, then snapshots again after the page changes.","The SKILL.md in the repository is deliberately thin. It points the agent to agent-browser skills get core, which prints workflows and troubleshooting from the installed CLI, so the instructions always match the version on your machine. Specialized guides load the same way for Electron desktop apps such as VS Code or Slack, Slack workspaces, exploratory QA, deriving an API client from recorded traffic, Vercel Sandbox and AWS Bedrock AgentCore browsers.","The CLI itself covers sessions, a vault for saved logins, persistent profiles, state encryption, network inspection, accessibility audits, React and Web Vitals checks, and video recording. An observability dashboard on port 4848 shows running sessions, and an MCP server mode exists for clients that prefer tools over shell commands."],"useItWhen":["You want your agent to log in to a web app and test a flow end to end.","You need data pulled from a page that only renders with JavaScript.","You want an agent to click through an Electron app such as Slack or VS Code.","You want browser instructions that stay in step with the installed tool version."],"notFor":["Writing Playwright test files for your repo: Microsoft's Playwright CLI skill generates Playwright tests."],"requirements":["The agent-browser CLI (npm, Homebrew or Cargo)","Chrome, downloaded by agent-browser install or detected if already present"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add vercel-labs/agent-browser --skill agent-browser"},{"label":"Install the agent-browser CLI, which the skill drives","command":"npm install -g agent-browser"},{"label":"Download Chrome for Testing the first time","command":"agent-browser install"}],"sourceUpdated":"2026-08-24","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/coreyhaines31/ai-seo","owner":"coreyhaines31","slug":"ai-seo","name":"AI SEO","skillName":"ai-seo","title":"ai-seo skill: get cited in ChatGPT, Perplexity and AI Overviews","summary":"ai-seo is Corey Haines's agent skill for answer engine optimization: the agent audits whether AI assistants cite your site, then restructures content and site files so AI search can extract and cite it.","category":"marketing-and-seo","source":"coreyhaines31/marketingskills","path":"skills/ai-seo","license":"MIT","worksWith":["Claude Code","Codex","Cursor","Windsurf","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["AI SEO is one of Corey Haines's marketing skills, for getting content cited by AI assistants and AI search: Google AI Overviews and AI Mode, ChatGPT, Perplexity, Claude, Gemini and others. It loads for terms such as AEO, GEO, LLM optimization, llms.txt or agent readiness. Like the rest of the set, it first reads your product marketing context file if one exists.","The audit has four steps. The agent tests 10 to 20 of your key queries across platforms and records who gets cited, compares how cited competitors structure their pages, checks whether your content can be extracted as standalone passages, and checks that robots.txt lets AI search crawlers in. The strategy then works on three pillars: structure, authority and presence on the third-party sites AI systems read.","Structure rules include leading each section with a direct answer, short answer passages, headings phrased like real questions, and tables for comparisons. The skill also covers machine-readable files such as llms.txt, schema markup, which content types get cited most, monitoring tools, and a list of things not to do drawn from Google's guidance, such as writing separate content just for AI or blocking the crawlers you want citations from."],"useItWhen":["You want to know whether ChatGPT or Perplexity mention your product for the questions buyers ask.","A competitor gets cited in AI answers and you do not.","You want pages rewritten so AI search can quote them.","You are deciding which AI crawlers to allow in robots.txt."],"notFor":["Classic ranking problems such as indexing or Core Web Vitals: use seo-audit from the same set.","Treating its numbers as fixed rules: citation patterns change as AI engines update, so retest its advice on the engines you care about."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add coreyhaines31/marketingskills --skill ai-seo"},{"label":"Or, in Claude Code, add the publisher's plugin marketplace","command":"/plugin marketplace add coreyhaines31/marketingskills"},{"label":"Then install all the marketing skills","command":"/plugin install marketing-skills"}],"sourceUpdated":"2026-09-05","version":"2.5.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/algorithmic-art","owner":"anthropics","slug":"algorithmic-art","name":"Algorithmic Art","skillName":"algorithmic-art","title":"algorithmic-art skill: generative art in p5.js with seeds","summary":"Algorithmic Art is Anthropic's open-source agent skill for generative art: the agent writes a short art philosophy, then builds an interactive p5.js piece with seeds and parameter sliders.","category":"video-and-media","source":"anthropics/skills","path":"skills/algorithmic-art","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"agentNote":"The viewer template is built to show as an interactive artifact in Claude.ai; in any other agent you open the HTML file in a browser.","cliChecked":"2026-10-01","whatItDoes":["Algorithmic Art is an Anthropic skill for making art with code. It loads when you ask for generative art, algorithmic art, flow fields or particle systems, and it produces two things: a written philosophy as a Markdown file and an interactive HTML piece. Its instructions tell the agent to create original work rather than copy existing artists, to avoid copyright problems.","With Algorithmic Art, the agent first names a generative art movement and writes its philosophy in four to six paragraphs. It also picks a subtle reference to your request to weave into the algorithm. Then it expresses that philosophy as a p5.js sketch with seeded randomness, so the same seed always draws the same image, and with parameters such as quantities, scales, angles and thresholds that you can tune.","The Algorithmic Art HTML file is self-contained and runs in Claude.ai or any browser, loading p5.js from a CDN. The agent builds it from the bundled viewer template, which keeps Anthropic's branding and a fixed sidebar: seed controls (previous, next, random, jump to a seed), sliders for the parameters, optional color pickers, and regenerate, reset and download PNG buttons. A second template shows how to structure p5.js code."],"useItWhen":["You want generative art such as a flow field or particle system without writing p5.js yourself.","You want to explore many variations of one piece by changing a seed and a few sliders.","You need a single HTML file of interactive art that opens in any browser."],"notFor":["You want a static poster or print file instead of an interactive sketch: use Canvas Design.","You want a short animated GIF: use Slack GIF Creator."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill algorithmic-art"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/microsoft/azure-deploy","owner":"microsoft","slug":"azure-deploy","name":"Azure deploy","skillName":"azure-deploy","title":"azure-deploy skill: run azd up and terraform apply with an agent","summary":"azure-deploy is Microsoft's agent skill for deploying an already prepared and validated app to Azure with azd, Bicep or Terraform, with error recovery, role checks and verified https endpoints at the end.","category":"cloud-and-deployment","source":"microsoft/azure-skills","path":"skills/azure-deploy","license":"MIT","worksWith":["GitHub Copilot (VS Code, Copilot CLI, IntelliJ IDEA)","Claude Code","Codex","Cursor","Gemini CLI","OpenCode"],"agentNote":"The skill calls tools from the Azure MCP Server (names starting mcp_azure_mcp). The plugin routes install that server; with the skills CLI alone the agent falls back to the Azure CLI commands the skill also lists.","cliChecked":"2026-10-01","whatItDoes":["azure-deploy is an agent skill from Microsoft that runs the deployment step for an app that is already prepared for Azure. It owns the commands that change cloud resources: azd up, azd deploy, terraform apply and az deployment. It refuses to start unless .azure/deployment-plan.md exists with the status Validated and a filled-in validation proof, and it forbids the agent from setting that status itself.","The azure-deploy workflow has ten steps. The agent reads the plan, works through a pre-deploy checklist, loads the recipe for the plan's deployment type, and for Container Apps with a registry checks that the AcrPull role has propagated. It then deploys, applies post-deploy steps such as SQL managed identity and EF migrations, handles errors from the recipe's error guide, and confirms the endpoints respond.","After deployment the skill has the agent query Azure to confirm the assigned roles are correct and sufficient, then report every endpoint as a full https link. Destructive actions need the user's confirmation. Reference files cover troubleshooting, azd and Azure Identity in Python, .NET, TypeScript and Java."],"useItWhen":["Your app is prepared for Azure and you want to run azd up or azd deploy.","You want terraform apply or a Bicep deployment run with checks before and after.","You are pushing an update to an app that already runs on Azure.","A Container Apps deployment fails because the image cannot be pulled from the registry."],"notFor":["Creating a new app or writing its infrastructure: the skill sends that to azure-prepare, then azure-validate, both in the same repository."],"requirements":["An Azure subscription, Azure CLI signed in with az login, and Azure Developer CLI signed in with azd auth login","A .azure/deployment-plan.md marked Validated by the azure-validate skill"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add microsoft/azure-skills --skill azure-deploy"},{"label":"Or, in Claude Code, install Microsoft's Azure plugin from Anthropic's official marketplace","command":"/plugin install azure@claude-plugins-official"},{"label":"Or, in Gemini CLI, install the Azure extension","command":"gemini extensions install https://github.com/microsoft/azure-skills"},{"label":"Or, in GitHub Copilot CLI, add Microsoft's marketplace","command":"/plugin marketplace add microsoft/azure-skills"},{"label":"Then install the Azure plugin in Copilot CLI","command":"/plugin install azure@azure-skills"}],"installNote":"The plugin and extension routes install all of Microsoft's Azure skills together with the Azure MCP Server and Foundry MCP. The skills CLI line installs only this skill's instructions, not the MCP servers it calls. In Codex, run codex plugin marketplace add microsoft/azure-skills, then install azure from /plugins.","sourceUpdated":"2026-08-10","version":"1.2.1","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/microsoft/azure-diagnostics","owner":"microsoft","slug":"azure-diagnostics","name":"Azure diagnostics","skillName":"azure-diagnostics","title":"azure-diagnostics skill: troubleshoot Azure production issues","summary":"azure-diagnostics is Microsoft's agent skill for debugging Azure production problems with AppLens, Azure Monitor and resource health, with guides for App Service, Functions, Container Apps, AKS, VMs and messaging.","category":"cloud-and-deployment","source":"microsoft/azure-skills","path":"skills/azure-diagnostics","license":"MIT","worksWith":["GitHub Copilot (VS Code, Copilot CLI, IntelliJ IDEA)","Claude Code","Codex","Cursor","Gemini CLI","OpenCode"],"agentNote":"The skill calls tools from the Azure MCP Server (names starting mcp_azure_mcp). The plugin routes install that server; with the skills CLI alone the agent falls back to the Azure CLI commands the skill also lists.","cliChecked":"2026-10-01","whatItDoes":["azure-diagnostics is an agent skill from Microsoft for finding out why something on Azure is failing. It gives the agent a fixed order: name the symptom, check Azure resource health, read the logs, look at metrics, then look for recent changes. Where the Azure MCP Server is available, the agent uses AppLens for automated detection and root cause hints, and Azure Monitor to run KQL queries against Log Analytics.","The azure-diagnostics skill routes each service to its own troubleshooting guide. Container Apps covers image pull failures, cold starts, health probes and port mismatches. App Service covers high CPU, failed deployments, crashes and TLS. Functions covers invocation failures, timeouts and missing settings. Separate guides handle AKS clusters and pods, VM RDP and SSH problems with firewalls and credential resets, and Event Hubs and Service Bus SDK errors.","The skill also includes a KQL query library, Azure Resource Graph queries and plain Azure CLI commands for health checks and activity logs, so diagnosis still works without the MCP tools. The agent is asked to record what it found and which fixes it tried."],"useItWhen":["Your App Service app is slow, crashing or stuck at high CPU.","A container app will not start because of image pull or health probe errors.","Pods in your AKS cluster are pending or in a crash loop.","You cannot connect to an Azure VM over RDP or SSH."],"notFor":["Deploying or redeploying an app: use azure-deploy from the same repository."],"requirements":["An Azure subscription and Azure CLI signed in with az login"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add microsoft/azure-skills --skill azure-diagnostics"},{"label":"Or, in Claude Code, install Microsoft's Azure plugin from Anthropic's official marketplace","command":"/plugin install azure@claude-plugins-official"},{"label":"Or, in Gemini CLI, install the Azure extension","command":"gemini extensions install https://github.com/microsoft/azure-skills"},{"label":"Or, in GitHub Copilot CLI, add Microsoft's marketplace","command":"/plugin marketplace add microsoft/azure-skills"},{"label":"Then install the Azure plugin in Copilot CLI","command":"/plugin install azure@azure-skills"}],"installNote":"The plugin and extension routes install all of Microsoft's Azure skills together with the Azure MCP Server and Foundry MCP. The skills CLI line installs only this skill's instructions, not the MCP servers it calls. In Codex, run codex plugin marketplace add microsoft/azure-skills, then install azure from /plugins.","sourceUpdated":"2026-08-19","version":"1.2.6","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/brainstorming","owner":"obra","slug":"brainstorming","name":"Brainstorming","skillName":"brainstorming","title":"brainstorming skill: agree the design before your agent codes","summary":"Brainstorming is an agent skill from Superpowers that makes your coding agent work out what you want and agree a design with you before it writes any code.","category":"planning-and-workflow","source":"obra/superpowers","path":"skills/brainstorming","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Brainstorming is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It runs before any creative work, such as a new feature, a component or a change in behavior, and turns an idea into a design through conversation.","In Brainstorming, the agent first works out what you want, who it is for and what success looks like. When something is missing, it asks one focused question about purpose before proposing features. It then writes its understanding back to you, keeping what you said apart from what it assumed, so you can correct it.","A hard gate in Brainstorming stops the agent from writing product code, scaffolding or installing dependencies until you approve the design: a written spec for larger work, or a short design in the chat for smaller work. The folder also holds a visual companion, a small local server with a page template."],"useItWhen":["You are about to ask an agent for a new feature and want to agree on what it builds first.","Your agent tends to start coding before it understands the request.","You want a written spec for bigger work, or a short design in the chat for small work."],"notFor":["You already have an approved spec: Writing Plans turns it into tasks."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill brainstorming"}],"faq":[{"q":"Does the Superpowers brainstorming skill work in Codex or Cursor?","a":"Yes. The Superpowers README documents installs for Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot CLI, OpenCode and nine more agents. Installing the whole plugin rather than this one skill keeps the skills it hands off to, such as Writing Plans."},{"q":"Will the brainstorming skill stop my agent from coding?","a":"Until you approve a design, yes: a hard gate blocks product code, scaffolding and dependency installs. A small change gets a short design in the chat, larger work a written spec."}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/brand-guidelines","owner":"anthropics","slug":"brand-guidelines","name":"Brand Guidelines","skillName":"brand-guidelines","title":"brand-guidelines skill: apply Anthropic's colors and fonts","summary":"Brand Guidelines is Anthropic's open-source agent skill that holds Anthropic's own brand colors and fonts, so an agent can give a document, slide deck or web artifact the Anthropic look.","category":"frontend-and-design","source":"anthropics/skills","path":"skills/brand-guidelines","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"cliChecked":"2026-10-01","whatItDoes":["Brand Guidelines holds Anthropic's official brand colors and typography, for any artifact that should have Anthropic's look and feel. It loads when brand colors, style guidelines, visual formatting or company design standards apply. It is one of Anthropic's example skills, so it also shows how a company can package its own palette and fonts as a skill.","The Brand Guidelines palette has four main colors (dark #141413, light #faf9f5, mid gray #b0aea5 and light gray #e8e6dc) and three accents (orange #d97757, blue #6a9bcc and green #788c5d). Headings of 24pt and larger use Poppins with an Arial fallback, and body text uses Lora with a Georgia fallback. Text color is chosen to suit the background.","In Brand Guidelines, shapes that are not text take the accent colors, cycling through orange, blue and green. Colors are applied as exact RGB values through the RGBColor class of python-pptx, the Python library for PowerPoint files. The skill installs no fonts: it uses Poppins and Lora when they are already on the system and falls back to Arial and Georgia otherwise."],"useItWhen":["You are making a slide deck or document that should match Anthropic's visual identity.","You need Anthropic's exact hex values and font pairing in one place.","You want a working example of a brand skill to adapt for your own company's colors and fonts."],"notFor":["You want your own company's look: the skill holds only Anthropic's palette and fonts. Theme Factory offers ten other themes or builds one from a description."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill brand-guidelines"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/canvas-design","owner":"anthropics","slug":"canvas-design","name":"Canvas Design","skillName":"canvas-design","title":"canvas-design skill: design posters and art as PNG or PDF","summary":"Canvas Design is Anthropic's open-source agent skill for static visual work: the agent writes a design philosophy, then makes a poster or art piece as a finished PNG or PDF file.","category":"video-and-media","source":"anthropics/skills","path":"skills/canvas-design","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"cliChecked":"2026-10-01","whatItDoes":["Canvas Design is an Anthropic skill for static visual pieces. It loads when you ask for a poster, a piece of art, a design or another static image. Its instructions tell the agent to make original designs and never copy existing artists' work, to avoid copyright problems. The output is a design philosophy as a Markdown file plus the finished piece as a PDF or PNG.","With Canvas Design, the agent first names an aesthetic movement and writes its philosophy in four to six paragraphs. It then expresses that philosophy on a single page that is mostly visual with minimal text, weaving in a subtle reference to your request. It checks that nothing overlaps or falls off the page, then takes a second pass that refines what is already there instead of adding more.","The Canvas Design folder bundles a canvas-fonts directory of typefaces, from serif and sans-serif families to monospace and pixel faces, and the agent picks type from them. When you ask for more pages, it makes them in the same style but each one distinct, bundled into one PDF or several PNGs."],"useItWhen":["You need a poster or cover image and want a finished PNG or PDF, not a mockup.","You want abstract art with a quiet reference to your topic built into it.","You want a short series of pages in one consistent visual style."],"notFor":["You want interactive or seeded generative art: use Algorithmic Art.","You want a whole web page or app screen designed: use Frontend Design."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill canvas-design"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/claude-api","owner":"anthropics","slug":"claude-api","name":"Claude API","skillName":"claude-api","title":"claude-api skill: build apps with the Anthropic SDK and Claude","summary":"Claude API is Anthropic's open-source agent skill for code that calls Claude: current model IDs and prices, guides for seven SDK languages plus cURL, and guided migrations to newer models.","category":"coding-and-frameworks","source":"anthropics/skills","path":"skills/claude-api","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude API"],"agentNote":"Its named workflows start as /claude-api <subcommand>, the slash command form Claude Code gives skills; the SKILL.md also accepts the bare subcommand as the request, which works in agents without slash commands. It asks for your language with Claude Code's AskUserQuestion tool and falls back to Python where that tool is missing.","cliChecked":"2026-10-01","whatItDoes":["Claude API is Anthropic's reference skill for building applications on Claude. It loads when a prompt names Claude, Anthropic or a Claude model, when you ask about model choice, pricing or limits, or when a task is LLM-shaped and no provider is named. It skips projects that use another provider such as OpenAI or Gemini, and it will not add Anthropic SDK calls to a file written for one.","With Claude API loaded, the agent detects the project language (Python, TypeScript, Java, Go, Ruby, C#, PHP or cURL) and reads that language's bundled guides. It uses the official Anthropic SDK, uses raw HTTP only when asked, and takes function names from documentation instead of guessing. The skill carries a dated table of model IDs with prices, a table of API shapes that changed in 2025 and 2026, and quick references for thinking, caching, streaming, tool use and Managed Agents.","The Claude API skill also runs named workflows through /claude-api subcommands: migrate (move code to a newer model), upgrade (the Python SDK from 0.x to 1.x), prompt-audit (find dated patterns in prompts and CLAUDE.md files), cost-optimize, build-eval, hillclimb (improve an app against an eval), preserved-thinking-migration and managed-agents-onboard. The cost and eval workflows ask for your approval before they spend money on model calls."],"useItWhen":["You are adding a Claude call to an app and want current model IDs and SDK code, not patterns from old training data.","You need to move existing code to a newer Claude model.","Your Claude API bill is high and you want the agent to look for savings such as caching and batching.","You are choosing between a single API call, tool use and Managed Agents for an agent you are building."],"notFor":["Your project calls OpenAI, Gemini or another provider: the skill stands down when it finds one.","You want to give an agent tools through MCP rather than call the API: use MCP Builder."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the claude-api plugin","command":"/plugin install claude-api@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill claude-api"}],"sourceUpdated":"2026-09-29","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/getsentry/code-review","owner":"getsentry","slug":"code-review","name":"Code review","skillName":"code-review","title":"code-review skill: review pull requests the Sentry way","summary":"code-review is Sentry's agent skill for reviewing pull requests by Sentry's engineering practices: runtime errors, performance, side effects, compatibility, security, design, tests and the tone of feedback.","category":"code-review-and-git","source":"getsentry/skills","path":"skills/code-review","license":"Apache-2.0","worksWith":["Claude Code","Cursor","Cline","GitHub Copilot"],"cliChecked":"2026-10-01","whatItDoes":["code-review is an agent skill from Sentry that turns Sentry's published code review guidelines into a checklist for an agent. The agent looks for runtime errors, performance traps such as quadratic loops and N+1 queries, unintended side effects, breaking API changes without a migration path, heavy Django ORM queries and security issues such as injection, XSS, access control gaps and exposed secrets.","Beyond bugs, the code-review skill asks whether the change fits the existing architecture, whether tests cover the real requirements at the right level, and whether the change needs a senior reviewer: schema changes, API contract changes, new libraries, hot code paths and security-sensitive code all qualify. Feedback should be polite, specific and phrased as a question when the reviewer is unsure.","The skill is short, about 370 words, and its examples are Python with Django and TypeScript with React. It tells the agent to approve once only small issues are left and not to block a pull request over style, since the goal is lower risk rather than perfect code."],"useItWhen":["You want an agent to review a pull request with a consistent checklist.","You want review comments that are specific and polite rather than vague.","You need to know which changes deserve a second, senior reviewer.","You work on a Django or React codebase and want common pitfalls flagged."],"notFor":["A security-focused review: use security-review from the same repository, or Trail of Bits' differential-review."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add getsentry/skills --skill code-review"},{"label":"Or, for Claude Code, add Sentry's plugin marketplace","command":"claude plugin marketplace add getsentry/skills"},{"label":"Then install the sentry-skills plugin, which holds all 28 skills","command":"claude plugin install sentry-skills@sentry-skills"}],"installNote":"Restart Claude Code after installing the plugin. Sentry's README says its skills were written for Sentry employees, so some carry Sentry conventions.","sourceUpdated":"2026-04-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/wshobson/database-migration","owner":"wshobson","slug":"database-migration","name":"Database migration","skillName":"database-migration","title":"database-migration skill: zero-downtime schema changes, rollbacks","summary":"database-migration is an agent skill from Seth Hobson's wshobson/agents marketplace for writing schema and data migrations in Sequelize, TypeORM and Prisma, with zero-downtime patterns and rollback strategies.","category":"databases","source":"wshobson/agents","path":"plugins/framework-migration/skills/database-migration","license":"MIT","worksWith":["Claude Code","Codex","Cursor","OpenCode","Antigravity CLI","GitHub Copilot","Pi"],"cliChecked":"2026-10-01","whatItDoes":["database-migration is an agent skill in wshobson/agents, an independent plugin marketplace with 40,127 GitHub stars on 2026-10-01. It gives an agent working templates for migrations in three ORMs: Sequelize migration files with up and down functions, TypeORM migration classes, and Prisma schema changes applied with prisma migrate. Each template shows the command to run it and the command to undo it.","The database-migration skill focuses on changes that must not take an app offline. Adding a column comes with a default value. Renaming a column happens in three steps: add the new column and copy the data, switch the application over, then drop the old column in a later migration. Changing a column type goes through a temporary column with a cast, and a data transformation example splits one address column into street, city and state, with a down step that rebuilds it.","For rollbacks the skill shows migrations wrapped in a transaction and a checkpoint approach that copies the table to a backup and restores it if a check fails. A reference file adds blue-green deployment for databases and a worked PostgreSQL to MySQL migration. The examples update rows one at a time and are written for JavaScript and TypeScript only; there is nothing for Django, Rails, Alembic or Go migration tools."],"useItWhen":["You need to rename a column in production without downtime.","You want a migration with a working down step so you can roll it back.","You need to split one column into several, such as an address into street, city and state.","You are switching a Node.js app from one ORM to another."],"notFor":["Prisma 6 to 7 upgrades: Prisma's own prisma-upgrade-v7 skill covers the breaking changes.","Python, Ruby or Go projects: the examples are only Sequelize, TypeORM and Prisma."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add wshobson/agents --skill database-migration"},{"label":"Or with GitHub's gh CLI","command":"gh skill install wshobson/agents database-migration"},{"label":"Or, in Claude Code, add the wshobson/agents marketplace","command":"/plugin marketplace add wshobson/agents"},{"label":"Then install the framework-migration plugin, which holds this skill","command":"/plugin install framework-migration"}],"sourceUpdated":"2026-05-22","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/deploy-to-vercel","owner":"vercel-labs","slug":"deploy-to-vercel","name":"Deploy to Vercel","skillName":"deploy-to-vercel","title":"deploy-to-vercel skill: deploy your app and get a preview link","summary":"Deploy to Vercel is Vercel's agent skill for getting a project live: the agent checks how your project is set up, picks a deploy method and returns a preview link unless you ask for production.","category":"cloud-and-deployment","source":"vercel-labs/agent-skills","path":"skills/deploy-to-vercel","license":"MIT, stated only in the repository README (no license file)","licenseNote":"The SKILL.md has no license key, there is no LICENSE file in the skill folder or the repository, and GitHub reports no license (checked 2026-10-01). The only grant is a License heading at the end of the repository README that says MIT, which covers the whole collection.","worksWith":["Claude Code and other terminal agents","Claude.ai","Claude Desktop","Codex"],"agentNote":"The SKILL.md has separate paths for Claude Code and other terminal agents, the claude.ai sandbox and the Codex sandbox. Its terminal fallback runs the script from ~/.claude/skills/, Claude Code's skills folder, and notes the path may differ where you installed it.","cliChecked":"2026-10-01","whatItDoes":["Deploy to Vercel is Vercel's skill for getting a project live from an agent session. It loads when you ask the agent to deploy your app, push it live, give you a link or create a preview deployment. It deploys as a preview unless you explicitly ask for production, and it aims to leave the project linked to Vercel so that every git push deploys.","With Deploy to Vercel, the agent first runs four checks: whether there is a git remote, whether the folder is already linked to a Vercel project, whether the Vercel CLI is installed and signed in, and which teams you belong to. If you are in several teams, it asks which one to use. An unlinked project is linked first, by repository when a remote exists. With a git remote the agent deploys by commit and push, only after you approve the push; without one it runs vercel deploy.","When the CLI cannot be installed or signed in, as in the claude.ai sandbox or Codex, a script bundled with Deploy to Vercel detects the framework from package.json, packages the project without node_modules, .git and .env files, and uploads it to a claimable deploy endpoint. You get a preview URL plus a claim URL that moves the deployment into your own Vercel account. The agent returns the link without fetching the deployed site."],"useItWhen":["You built something with an agent and want a live preview link to share.","You have a project that is not linked to Vercel yet and want every git push to deploy it.","You are working in claude.ai or Codex, where the Vercel CLI cannot sign in, and still want a deployment you can claim later.","You want the agent to deploy a preview and leave production alone until you ask."],"notFor":["You deploy from CI or a headless agent with a Vercel access token: use Vercel CLI with Tokens.","You want to cut cost or speed up a project already on Vercel: use Vercel Optimize."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill deploy-to-vercel"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-03-06","version":"3.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/leonxlnx/design-taste-frontend","owner":"leonxlnx","slug":"design-taste-frontend","name":"Design taste frontend","skillName":"design-taste-frontend","title":"design-taste-frontend skill: landing pages without AI clichés","summary":"design-taste-frontend, the main skill in Leon Lin's Taste Skill repository, makes an AI agent read the brief, pick a design direction and build landing pages, portfolios and redesigns that avoid templated defaults.","category":"frontend-and-design","source":"Leonxlnx/taste-skill","path":"skills/taste-skill","license":"MIT","worksWith":["Codex","Cursor","Claude Code","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["Design taste frontend is the default skill in the Taste Skill repository, kept in the taste-skill folder but installed by its frontmatter name. The README labels this version v2 and experimental. Before writing code, the agent states a one-line design read: the page type, the audience, the visual language and the design system it will lean on. If the brief is unclear, it asks exactly one question.","The design read sets three dials from 1 to 10: design variance, motion intensity and visual density, with a baseline of 8, 6 and 4 and presets for agency sites, developer portfolios, public-sector services and redesigns. When a brief matches a real design system, such as Fluent, Material, Carbon, Primer, GOV.UK Frontend or shadcn/ui, the agent is told to install the official package instead of copying its look by hand.","The default stack is React or Next.js with Tailwind v4 and the Motion library. Long rule sections cover typography, color, layout, dark mode, reduced motion, Core Web Vitals targets and a list of AI tells to avoid, such as purple gradients, three equal feature cards and placeholder names. Redesigns start with an audit of the existing site, and every build ends with a pre-flight check."],"useItWhen":["You want a landing page that does not look like a default AI template.","You are building a portfolio and want the agent to choose a style that fits the audience.","You are redesigning a site and want an audit before anything changes.","You want the agent to use an official design system package when the brief calls for one."],"notFor":["Dashboards, data tables and multi-step product UI: the skill rules them out of scope in its first lines.","Projects that need fixed behavior: v2 is marked experimental, and the publisher keeps design-taste-frontend-v1 for pinning."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add Leonxlnx/taste-skill --skill design-taste-frontend"},{"label":"Or pin the original v1, which the publisher keeps for projects that depend on it","command":"npx skills add Leonxlnx/taste-skill --skill design-taste-frontend-v1"}],"sourceUpdated":"2026-05-26","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/diagnosing-superpowers","owner":"obra","slug":"diagnosing-superpowers","name":"Diagnosing Superpowers","skillName":"diagnosing-superpowers","title":"diagnosing-superpowers skill: why an agent session went wrong","summary":"Diagnosing Superpowers is an agent skill from Superpowers that reads a session's transcripts when work went wrong and reports what happened, with every finding tied to a file and line.","category":"planning-and-workflow","source":"obra/superpowers","path":"skills/diagnosing-superpowers","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have. This skill needs a subagent tool: Claude Code's Agent tool, or in Codex the multi-agent feature switched on in config.toml (multi_agent = true).","cliChecked":"2026-10-01","whatItDoes":["Diagnosing Superpowers is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when a Superpowers session went wrong and you want to know why: repeated work, an ignored plan, a skill that never fired, a session that took too long or cost too much. It works on the current session or a past one named by id or path, and it can prepare a bug report for the Superpowers maintainers.","In Diagnosing Superpowers, the agent first asks you one question at a time until it can state the problem, then finds the session files on disk and fills a case file under ~/.superpowers/diagnosing-superpowers/. It dispatches seven analyst subagents in parallel, one per dimension: skill timeline, plan adherence, repeated work, stumbles, quality evidence, request conflicts, and cost and time. A finding without a file path and line is dropped, and session files are never modified.","The Diagnosing Superpowers result is a report filled from a bundled template. The skill reports and does not diagnose: it states whether Superpowers was involved, but never names a defect in a skill or proposes a change. When Superpowers may be involved, or when you ask, it searches GitHub issues and drafts a new one, created only after you approve the exact text. On request it builds a bundle of the session with emails, names, secrets and private paths replaced, and reminds you that scrubbing can miss things."],"useItWhen":["You want to know why an agent session took so long or cost so much.","Your agent ignored the plan, repeated work, or never used a skill you expected.","You want to check what a session that is still running is actually doing.","You want to file a bug report with the Superpowers maintainers, backed by evidence from the transcript."],"notFor":["You want to debug your own code rather than an agent session: use Systematic Debugging."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill diagnosing-superpowers"}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/trailofbits/differential-review","owner":"trailofbits","slug":"differential-review","name":"Differential review","skillName":"differential-review","title":"differential-review skill: security review of a PR or diff","summary":"differential-review is Trail of Bits' agent skill for security review of a pull request, commit or diff, using git history, blast radius counts and test coverage, and ending in a written markdown report.","category":"security","source":"trailofbits/skills","path":"plugins/differential-review/skills/differential-review","license":"CC-BY-SA-4.0","licenseNote":"CC-BY-SA-4.0 is a share-alike license. You can copy the skill into a project and change it, even for commercial work, if you credit Trail of Bits, link the license and say what you changed. If you share a changed version, for example in a public repository, it must stay under CC-BY-SA-4.0. It is a content license, not a software license, so ask your legal team before bundling it inside a closed product.","worksWith":["Claude Code","Codex","ChatGPT workspace marketplace"],"agentNote":"For high-risk changes the skill can hand attacker modeling to a Claude Code subagent, differential-review:adversarial-modeler, which ships in the plugin and not in the skill folder. Through the skills CLI, the agent follows the adversarial.md guide in the folder instead.","cliChecked":"2026-10-01","whatItDoes":["differential-review is an agent skill from security firm Trail of Bits for reviewing code changes rather than a whole codebase. It classifies each change by risk, not size: authentication, cryptography, external calls, value transfer and removed validation are high risk; business logic and new public APIs are medium; comments, tests, UI and logging are low. Review depth scales with codebase size, from reading everything under 20 files to critical paths only above 200.","The differential-review workflow runs in phases: triage, code analysis that runs git blame when security code is deleted, test coverage of the changed lines, blast radius by counting callers, deeper context on high-risk changes, adversarial modeling with concrete exploit scenarios, and a final report. Supporting files cover the methodology, attacker modeling, vulnerability patterns such as regressions, reentrancy and access control, and the report format.","The skill lists shortcuts it rejects, such as treating a small PR or a refactor as low risk by default, and always writes a markdown report file with line numbers and commits behind each finding. Its own principles ask it to state coverage limits and confidence honestly rather than imply a clean bill of health."],"useItWhen":["You want a security review of one pull request before it merges.","You suspect a change reintroduces a bug that was fixed before.","You want to know what else in the codebase a change could break.","You need a written review report with line references for an audit trail."],"notFor":["Auditing a whole codebase from scratch: Trail of Bits pairs this skill with its audit-context-building skill for that."],"requirements":["A git repository with history for the changed files"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add trailofbits/skills --skill differential-review"},{"label":"Or, in Claude Code, add Trail of Bits' plugin marketplace","command":"/plugin marketplace add trailofbits/skills"},{"label":"Or, in Codex, add the same marketplace from your terminal","command":"codex plugin marketplace add trailofbits/skills"},{"label":"Then install the differential-review plugin in Codex","command":"codex plugin add differential-review@trailofbits"}],"installNote":"In Claude Code, open /plugin menu after adding the marketplace and install the differential-review plugin. For a ChatGPT workspace, Trail of Bits' README says to import the repository's .claude-plugin/marketplace.json.","faq":[{"q":"Can I copy Trail of Bits' differential-review skill into my own repository?","a":"Yes, with conditions. The skills are CC-BY-SA-4.0. You may copy and adapt them, even commercially, if you credit Trail of Bits, link the license and note your changes. Any changed version you share must carry the same CC-BY-SA-4.0 license. Using the skill unchanged in your own work does not put your code under that license."},{"q":"Does differential-review work outside Claude Code?","a":"Yes. Trail of Bits documents Codex through its plugin marketplace and ships an agents/openai.yaml file in the skill folder for ChatGPT workspace imports. The optional adversarial-modeler subagent is a Claude Code plugin feature; other agents use the adversarial.md guide in the same folder."}],"sourceUpdated":"2026-09-16","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/discernment-nudge","owner":"anthropics","slug":"discernment-nudge","name":"Discernment Nudge","skillName":"discernment-nudge","title":"discernment-nudge skill: questions to fact-check AI answers","summary":"Discernment Nudge is Anthropic's agent skill that ends a substantive Claude answer with two or three specific follow-up questions, so you can check its facts, reasoning and assumptions.","category":"learning-ai-assistants","source":"anthropics/skills","path":"skills/discernment-nudge","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude API"],"agentNote":"Anthropic wrote Discernment Nudge for Claude's answers, from its AI Fluency course, but the instructions use no Claude-only tool.","cliChecked":"2026-10-01","whatItDoes":["Discernment Nudge is an Anthropic skill built on three habits from the AI Fluency framework: checking facts, questioning reasoning and noticing missing context. It loads after Claude gives an answer you may act on: advice or a recommendation, a drafted plan, pitch, proposal or email, estimates or projections, an interpretation of data, factual claims you may rely on, or a multi-step argument.","With Discernment Nudge, Claude answers in full first. It then adds the line \"A few things worth a second look:\" followed by two or three short questions you could send back. Each question points at something specific in the answer, such as a figure, a step in the reasoning or an assumption about your situation, and stays under about 120 characters. Nothing follows the list.","The Discernment Nudge appears at most once per conversation. Claude leaves it off for creative writing, casual chat, code you will run, simple lookups and purely educational explanations. It also stays silent when you already asked it to verify or cite, asked for the quick version, asked it to review your own work, gave it your own material to summarize, or asked for its opinion."],"useItWhen":["You use Claude for business, money, health or career advice and want a prompt to check it before you act.","You get cost or timeline estimates from Claude and want to know which numbers to verify.","You want to build the habit of questioning AI answers without a warning on every reply."],"notFor":["You want sources checked on every reply: the nudge appears at most once per conversation and skips code, lookups and creative writing."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the discernment-nudge plugin","command":"/plugin install discernment-nudge@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill discernment-nudge"}],"sourceUpdated":"2026-08-17","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/dispatching-parallel-agents","owner":"obra","slug":"dispatching-parallel-agents","name":"Dispatching Parallel Agents","skillName":"dispatching-parallel-agents","title":"dispatching-parallel-agents skill: fix independent bugs at once","summary":"Dispatching Parallel Agents is an agent skill from Superpowers: when several problems are independent, your agent sends a focused subagent to each at once, then checks the results together.","category":"planning-and-workflow","source":"obra/superpowers","path":"skills/dispatching-parallel-agents","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have. This skill needs a subagent tool: Claude Code's Agent tool, or in Codex the multi-agent feature switched on in config.toml (multi_agent = true).","cliChecked":"2026-10-01","whatItDoes":["Dispatching Parallel Agents is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when there are two or more independent tasks that share no state and do not depend on each other, such as test failures in different files or bugs in separate subsystems. Instead of investigating them one after another, the agent gives each problem its own subagent, and they work at the same time.","In Dispatching Parallel Agents, each subagent gets a prompt the agent writes for it, never the session's own history: one test file or subsystem as its scope, a clear goal, constraints such as not changing other code, and the summary it must return. The skill contrasts vague prompts (\"fix all the tests\") with focused ones that paste in the error messages and test names. All dispatches go out in the same response, which is what makes them run in parallel.","Dispatching Parallel Agents also says when not to use it: failures that are related, problems that need the whole system in view, exploratory debugging where the cause is still unknown, and work where agents would edit the same files. When the subagents return, the agent reads each summary, checks that the fixes do not conflict, runs the full test suite, and spot-checks the changes, since agents can make systematic errors."],"useItWhen":["You have several failing tests in different files that look unrelated.","You have separate parts of your project that broke at the same time and want them fixed in parallel, not one by one.","You want your main agent's context kept free for coordination while other agents do the digging."],"notFor":["The failures are related or the cause is unknown: use Systematic Debugging.","You are carrying out a written plan task by task: use Subagent-Driven Development."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill dispatching-parallel-agents"}],"sourceUpdated":"2026-07-24","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/doc-coauthoring","owner":"anthropics","slug":"doc-coauthoring","name":"Doc Co-Authoring","skillName":"doc-coauthoring","title":"doc-coauthoring skill: write specs, PRDs and design docs with AI","summary":"Doc Co-Authoring is Anthropic's agent skill that takes you and an AI agent through a document in three stages: gather context, draft section by section, then test it on a fresh reader.","category":"writing","source":"anthropics/skills","path":"skills/doc-coauthoring","license":"No license stated","licenseNote":"\"No license stated\" means exactly that: the doc-coauthoring folder has no LICENSE file, its SKILL.md has no license key, and the anthropics/skills repository has no root license (checked 2026-10-01). Without a license, copyright law reserves all rights to Anthropic. Anthropic distributes it in its example-skills plugin, but nothing grants you the right to copy, change or redistribute it.","worksWith":["Claude Code","Claude.ai","Claude API"],"agentNote":"Reader testing uses sub-agents where the agent has them, and the SKILL.md names Claude Code. Without sub-agents, as in Claude.ai, it gives you steps to run the test yourself in a fresh chat.","whatItDoes":["Doc Co-Authoring is an Anthropic skill for writing structured documents with an AI agent, such as technical specs, proposals, PRDs, design docs, decision docs and RFCs. It loads when you start a writing task like these. The agent first offers the workflow and explains its three stages: context gathering, refinement and structure, and reader testing. If you decline, it works freeform.","In the first Doc Co-Authoring stage the agent asks about the document type, audience, desired impact and any template, invites you to dump all your context, then asks 5 to 10 numbered questions about the gaps. It can pull context from connected tools such as Slack or Google Drive when they are available. In the second stage it builds the doc section by section: clarifying questions, 5 to 20 brainstormed options, your pick of what to keep, a draft, then targeted edits.","In Doc Co-Authoring's reader testing, the agent predicts 5 to 10 questions readers will ask and checks whether a fresh agent with only the document can answer them, also looking for ambiguity, hidden assumptions and contradictions. In Claude Code it runs this with sub-agents; in Claude.ai it gives you the steps to run it in a new chat. Sections with gaps go back to drafting."],"useItWhen":["You need to write a spec, proposal or decision doc and the context lives in your head, your notes or team threads.","You want to know whether a doc makes sense to someone who was not in the conversation.","You want to build a long document one section at a time instead of rewriting one big draft."],"notFor":["You want an existing page checked for style and tone: use Writing Guidelines.","You need a weekly status update or a company newsletter: use Internal Comms."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2025-12-04","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/docx","owner":"anthropics","slug":"docx","name":"DOCX","skillName":"docx","title":"docx skill: create, edit and redline Word files (Claude only)","summary":"DOCX is Anthropic's proprietary agent skill for creating, reading and editing Word files, including tracked changes, comments and tables of contents; its license limits it to Claude and Anthropic's own products.","category":"documents-and-office","source":"anthropics/skills","path":"skills/docx","license":"Proprietary: use only within Anthropic's services, no copying or redistribution","licenseNote":"The LICENSE.txt in the docx folder (read 2026-10-01) puts use under your agreement with Anthropic and forbids extracting the files from Anthropic's services, keeping copies outside them, copying them, making derivative works and redistributing them. Anthropic's README calls its four document skills source-available, not open source. You may read the files on GitHub, but you may not copy them into Codex, Cursor or any other agent, or publish a changed version.","worksWith":["Claude Code","Claude.ai","Claude API"],"whatItDoes":["The DOCX skill teaches Claude to work with Word documents (.docx) and Word templates (.dotx). It loads when you mention a Word file or ask for a report, memo, letter or template delivered as one. It covers creating, reading and editing documents, find-and-replace, images, tracked changes and comments. It is not for PDFs, spreadsheets or Google Docs.","DOCX writes new documents with the docx library for JavaScript, following a list of known pitfalls, such as A4 being the default page size and tables needing widths on the table and on every cell. Existing files are unzipped and their XML edited directly. Bundled Python scripts merge split text runs so phrases can be found, validate against the Office XML schemas, check that every redline edit is tracked, accept tracked changes and add anchored comments. After writing a file, Claude renders its pages as images and looks at them.","DOCX is one of Anthropic's four document skills, next to PDF, PowerPoint and Excel. Unlike most skills in Anthropic's repository it is not open source: the license allows use only within Anthropic's own services and forbids copying, changing or redistributing it. That is why this page lists only Anthropic's install routes."],"useItWhen":["You need a report, memo or letter as a Word file with headings, page numbers or a table of contents.","You want to redline a contract with tracked changes and comments.","You need to find and replace text or swap images in an existing Word document.","You want the text of a .docx file as Markdown."],"notFor":["You work in Codex, Cursor, Gemini CLI or another agent that is not Anthropic's: the docx license bars copying the skill there.","You want a PDF or a spreadsheet: the skill says it is not for those; Anthropic's PDF and XLSX skills cover them.","You want a Google Doc: the skill does not load for Google Docs."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the document-skills plugin, which holds the Word, PDF, PowerPoint and Excel skills","command":"/plugin install document-skills@anthropic-agent-skills"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","faq":[{"q":"Can I use Anthropic's docx skill in Codex, Cursor or ChatGPT?","a":"No. Its LICENSE.txt (read 2026-10-01) forbids keeping copies outside Anthropic's services, copying and redistribution, so it is meant for Claude.ai, Claude Code through Anthropic's plugin, and the Claude API. The open skills CLI does list it, but installing it into another agent copies the files, which the license forbids."},{"q":"Can the docx skill redline a contract with tracked changes?","a":"Yes. It edits the document XML directly, and bundled scripts check that every redline edit is tracked, add anchored comments and validate the file against the Office XML schemas. Claude then renders the pages as images and looks at them."}],"sourceUpdated":"2026-07-17","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/expo/eas-app-stores","owner":"expo","slug":"eas-app-stores","name":"EAS app stores","skillName":"eas-app-stores","title":"eas-app-stores skill: ship Expo apps to App Store and Google Play","summary":"eas-app-stores is the Expo team's agent skill for shipping mobile apps with EAS: build profiles, signing, version numbers, TestFlight and submissions to the App Store and Google Play.","category":"mobile","source":"expo/skills","path":"plugins/expo/skills/eas-app-stores","license":"MIT","licenseNote":"The skill is free under MIT, but the service it drives is not always free: Expo labels it an EAS paid-service skill, where cloud builds use EAS plan resources with free-tier limits.","worksWith":["Claude Code","Codex","Cursor","OpenCode","GitHub Copilot","Windsurf","Gemini CLI","Cline","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["EAS app stores is one of the official Expo skills, in the Services group that relies on Expo Application Services. It covers building iOS and Android apps in the cloud and sending them to TestFlight, the App Store or Google Play. It works for Expo and other React Native projects, and also for native apps that use EAS only for delivery.","The agent first picks a path: a SwiftUI or UIKit app with no React Native, a native Android app, an Expo or React Native app, or a native app gaining React Native screens. For Expo projects it initializes EAS, sets build profiles in eas.json with remote app versions and auto-increment, and runs production builds with optional auto-submit. Reference files cover TestFlight, App Store review, Play Store tracks, store metadata and CI workflows.","The skill is careful about what success means. It tells the agent that a finished build or a queued submission does not prove Apple accepted the build, that testers can reach it, or that it is live in the store, and it lists checks for processing and availability. Monitoring commands show recent builds and submission status."],"useItWhen":["You are publishing your first Expo app to TestFlight and Google Play.","You want build profiles, signing and version numbers set up once in eas.json.","You want builds that submit to the stores automatically after they finish.","You have a native Swift app and want EAS to handle builds and submissions."],"notFor":["Hosting an Expo website or API routes: Expo points to its eas-hosting skill.","Free-only setups: EAS builds count against plan limits, and store memberships cost extra."],"requirements":["An Expo account with EAS (free tier and paid plans, see expo.dev/pricing)","An Apple Developer membership for iOS and a Google Play Console account for Android","EAS CLI (npm install -g eas-cli)"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add expo/skills --skill eas-app-stores"},{"label":"Or, in Claude Code, install Expo's plugin from the official marketplace","command":"claude plugin install expo@claude-plugins-official"},{"label":"Or, in Codex, install Expo's plugin from the OpenAI-curated marketplace","command":"codex plugin add expo@openai-curated"}],"sourceUpdated":"2026-09-09","version":"1.1.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/executing-plans","owner":"obra","slug":"executing-plans","name":"Executing Plans","skillName":"executing-plans","title":"executing-plans skill: carry out an implementation plan","summary":"Executing Plans is an agent skill from Superpowers in which your agent works through a written plan itself, task by task with tests, then gets one fresh review of the whole branch.","category":"planning-and-workflow","source":"obra/superpowers","path":"skills/executing-plans","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Executing Plans is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when the agent carries out an implementation plan itself in the current session, because you chose inline execution or because no subagent tool is available. One context does every task, and one fresh reviewer checks the whole branch at the end.","Under Executing Plans, work happens in an isolated workspace from Using Git Worktrees, and Test-Driven Development governs every step: each test is watched failing before the code that makes it pass. When the code is wrong, the agent switches to Systematic Debugging. Progress goes into a ledger file that survives context compaction and is shared with Subagent-Driven Development. Bundled task-start and task-done scripts print each task's brief and mark a task complete only when its tests pass.","Under Executing Plans, the agent does not stop to check in between tasks. It decides plan conflicts itself and records each as a ruling, stopping only for destructive or security-sensitive actions, a merge, push or publish, or a plan too broken to follow. The final review uses the reviewer template from Requesting Code Review. The agent then lists every ruling and deferred minor finding for you and hands off to Finishing a Development Branch."],"useItWhen":["You have a finished implementation plan and want the agent to carry it out without asking after every task.","Your agent tool has no subagents, or you want the agent to do every task itself.","You want a written record of every decision the agent made on your behalf while following the plan."],"notFor":["Your plan has many independent tasks and you want each one reviewed before the next: use Subagent-Driven Development."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill executing-plans"}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/expo/expo-router","owner":"expo","slug":"expo-router","name":"Expo Router","skillName":"expo-router","title":"expo-router skill: navigation, tabs and modals in Expo apps","summary":"expo-router is the Expo team's agent skill for navigation in Expo apps: file-based routes, native stacks and tabs, links with previews and context menus, modals, form sheets, headers and search bars.","category":"mobile","source":"expo/skills","path":"plugins/expo/skills/expo-router","license":"MIT","worksWith":["Claude Code","Codex","Cursor","OpenCode","GitHub Copilot","Windsurf","Gemini CLI","Cline","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["Expo Router is one of the official Expo skills and sits in the free, open-source Framework group of the repository. It teaches an agent how navigation is built with Expo Router: routes live in the app folder as files, groups and dynamic segments organize them, and every app needs a route that matches the root path. Components, types and utilities are kept out of the app folder.","The skill favors native iOS behavior. Stacks are defined in _layout files using the native Stack, page titles are set with Stack.Title, and links get previews and long-press context menus where possible. It covers modals, form sheets, native tabs, toolbar buttons and header search bars, with reference files for tabs, headers, form sheets, search and Apple zoom transitions on iOS 18 and later.","Small conventions are fixed too: kebab-case file names, path aliases instead of deep relative imports, old route files removed when navigation moves, and, from SDK 56, imports from expo-router/react-navigation instead of the React Navigation packages directly. Styling and motion are left to the companion expo-native-ui and expo-animation skills."],"useItWhen":["You are setting up tabs, stacks and modals in a new Expo app.","You want navigation that feels native on iOS, with link previews and context menus.","Your agent mixes React Navigation imports into an Expo Router project.","You need a form sheet or a header search bar and want the current Expo Router API."],"notFor":["Screen styling, colors and visual effects: Expo points to its expo-native-ui skill.","Bare React Native apps without Expo Router."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add expo/skills --skill expo-router"},{"label":"Or, in Claude Code, install Expo's plugin from the official marketplace","command":"claude plugin install expo@claude-plugins-official"},{"label":"Or, in Codex, install Expo's plugin from the OpenAI-curated marketplace","command":"codex plugin add expo@openai-curated"}],"installNote":"The Expo plugins for Claude Code and Codex include every Expo skill and wire up the Expo MCP server, and update through their marketplaces.","sourceUpdated":"2026-09-08","version":"1.0.1","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/finishing-a-development-branch","owner":"obra","slug":"finishing-a-development-branch","name":"Finishing a Development Branch","skillName":"finishing-a-development-branch","title":"finishing-a-development-branch skill: merge, open a PR or keep it","summary":"Finishing a Development Branch is an agent skill from Superpowers: when work is done, your agent runs the tests, offers to merge, open a pull request or keep the branch, then cleans up.","category":"code-review-and-git","source":"obra/superpowers","path":"skills/finishing-a-development-branch","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Finishing a Development Branch is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when implementation is complete and the work needs to be integrated. The agent first runs the project's full test suite. If anything fails, it reports the failures and stops: the options appear only after a green run. Before any merge it also confirms which branch the work split from.","Finishing a Development Branch then checks whether the agent is in a normal repository or a worktree, and shows you a fixed menu: merge back locally, push the branch and open a pull request, or keep the branch as it is. From a detached HEAD there are two options, with no local merge. The integration decision is yours, and the agent waits for it. A local merge is tested again on the merged result before the branch is deleted.","Finishing a Development Branch removes a worktree only if Superpowers created it (under .worktrees/ or worktrees/) and leaves any other worktree to the tool that made it. If removal is refused because of files that were never committed, it lists them and asks you what to do instead of forcing it. Discarding the work happens only when you ask, and only after you type the word discard. It force-pushes only on your explicit request."],"useItWhen":["Your agent finished a feature and you want to choose between merging it and opening a pull request.","You want the tests run on the final tree before anything is merged.","You use git worktrees and want them cleaned up without losing uncommitted files."],"notFor":["You want a code review before deciding: use Requesting Code Review."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill finishing-a-development-branch"}],"sourceUpdated":"2026-08-12","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/firebase/firebase-security-rules-auditor","owner":"firebase","slug":"firebase-security-rules-auditor","name":"Firebase security rules auditor","skillName":"firebase-security-rules-auditor","title":"firebase-security-rules-auditor skill: red-team Firestore rules","summary":"firebase-security-rules-auditor is the Firebase team's agent skill for auditing Firestore and Cloud Storage security rules for privilege escalation, update bypasses and missing limits, returning a 1 to 5 score in JSON.","category":"security","source":"firebase/agent-skills","path":"skills/firebase-security-rules-auditor","license":"Apache-2.0","worksWith":["Gemini CLI","Claude Code","Codex","Kimi Code CLI","Cursor","Windsurf","GitHub Copilot"],"cliChecked":"2026-10-01","whatItDoes":["firebase-security-rules-auditor is an agent skill from the Firebase team at Google that has the agent act as a red-team auditor of Firebase security rules. Instead of assuming complex rules are safe, the agent tries to find a sequence of reads and writes that gets around them. The checklist and examples are written around Firestore, although the skill's description also names Cloud Storage rules.","The firebase-security-rules-auditor checklist has six items. Could a user write a valid document first and later edit it into an invalid state, for example by granting themselves a role? Do rules trust user-supplied fields like role, isAdmin or ownerId? Do the rules still let the app work? Are there size limits on strings and arrays? Are field types checked? Do hasOnly or diff checks also verify who is making the change?","The result is JSON with a score from 1, critical, to 5, secure, a summary and a list of findings with severity and a fix for each. A single hardcoded admin email does not lower the score if email verification is checked and no one can promote themselves. The skill is short, about 540 words, and has no scripts or emulator tests."],"useItWhen":["You wrote Firestore security rules and want them attacked before launch.","You suspect users can change their own role or another user's document.","You want a scored audit of your rules you can track over time.","An AI tool generated your Firebase rules and you want them checked."],"notFor":["Writing new rules from scratch: use firestore-rules-creation from the same repository.","Firebase CLI deploys, Auth setup or database queries: the skill says it does not cover those."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add firebase/agent-skills --skill firebase-security-rules-auditor"},{"label":"Or, in Gemini CLI, install the Firebase extension","command":"gemini extensions install https://github.com/firebase/skills"},{"label":"Or, for Claude Code, add the Firebase plugin marketplace","command":"claude plugin marketplace add firebase/skills"},{"label":"Then install the firebase plugin","command":"claude plugin install firebase@firebase"}],"installNote":"Firebase's README also documents Codex (codex plugin marketplace add firebase/skills, then codex plugin add firebase@firebase) and Kimi Code CLI. The README uses the older repository name firebase/skills, which GitHub redirects to firebase/agent-skills.","sourceUpdated":"2026-09-17","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/firecrawl/firecrawl","owner":"firecrawl","slug":"firecrawl","name":"Firecrawl","skillName":"firecrawl","title":"firecrawl skill: web search, scraping and crawling for agents","summary":"firecrawl is Firecrawl's agent skill for live web work through its CLI: the agent searches the web, scrapes pages to clean markdown, maps and crawls sites, and interacts with pages that need clicks or logins.","category":"browser-and-web","source":"firecrawl/cli","path":"skills/firecrawl","license":"ISC (package.json only, no license file)","licenseNote":"GitHub reports no license for firecrawl/cli on 2026-10-01 and the repository has no LICENSE file; only its package.json declares ISC. The skill is free to read, but most commands call Firecrawl's paid API and spend credits.","worksWith":["Claude Code","Codex","Cursor","Windsurf","OpenCode","OpenClaw","OpenHands","Hermes Agent"],"agentNote":"The SKILL.md frontmatter pre-approves only firecrawl and npx firecrawl-cli shell commands (allowed-tools).","cliChecked":"2026-10-01","whatItDoes":["Firecrawl is the main skill in Firecrawl's CLI repository and routes an agent's web tasks through the firecrawl command. It is meant for ordinary research as well as scraping: finding sources, reading pages, collecting links, extracting data in bulk, downloading a site, watching pages for changes, and pages that need clicks, forms or a login. Local files go to a separate parse skill.","The skill sets an escalation order. The agent starts with search, scrapes a known page directly, uses map to find the right subpage on a large site, crawls only when it needs a whole section, sets a monitor for recurring checks instead of repeating scrapes, and adds interact only after a scrape cannot reach the content. Results are saved under a .firecrawl folder so the agent can reuse them instead of fetching again.","Separate skills in the same repository handle research papers, developer questions about libraries and errors, and an AI agent mode for structured extraction. The main skill points to them, and to large-result recovery steps when output is too big for the agent's context."],"useItWhen":["You want your agent to research a topic on the live web and cite saved sources.","You need every page of a documentation site as clean markdown.","You want to be alerted when a competitor's pricing page changes.","You need content from a page behind a login or a load-more button."],"notFor":["Heavy use without an account: search, scrape and interact work keyless on a rate-limited free tier, but crawl, map, download, agent and monitor need a Firecrawl login."],"requirements":["The firecrawl-cli npm package","A Firecrawl account or API key for most commands and higher limits"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add firecrawl/cli --skill firecrawl"},{"label":"Or let Firecrawl install the CLI, log you in and add its skills to every detected agent","command":"npx -y firecrawl-cli@latest init -y --browser"}],"faq":[{"q":"Is the Firecrawl skill free to use?","a":"The skill is free to install, but it drives Firecrawl's API. On 2026-10-01 the skill says search, scrape and interact work without a key on a rate-limited free tier, while crawl, map, download, agent and monitor need a login and use credits."},{"q":"What license does the Firecrawl CLI skill have?","a":"There is no license file in firecrawl/cli and GitHub reports none on 2026-10-01. The package.json declares ISC, a permissive license, but no LICENSE text ships with the repository."}],"sourceUpdated":"2026-09-25","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/frontend-design","owner":"anthropics","slug":"frontend-design","name":"Frontend Design","skillName":"frontend-design","title":"frontend-design skill: UI that does not look AI-generated","summary":"Frontend Design is Anthropic's open-source agent skill that steers an AI agent toward a look made for your product and away from the templated pages AI tends to produce.","category":"frontend-and-design","source":"anthropics/skills","path":"skills/frontend-design","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"cliChecked":"2026-10-01","whatItDoes":["Frontend Design gives the agent the brief of a design lead at a studio known for giving every client its own visual identity. Before designing, the agent pins down the subject, the audience and the main job of the page, and draws its palette, typography and layout from that subject's world.","Frontend Design names the defaults that make generated pages look alike: a cream background with a serif headline and a terracotta accent, near-black with one acid-green accent, identical rounded cards with the same soft shadow, all-caps labels above every heading, and fade-in animations on every section. The agent treats these as defaults to avoid, unless your brief asks for one of them.","Frontend Design also sets rules for type (one or two clearly different typefaces, lines under 80 characters), for motion (one deliberate moment instead of effects everywhere) and for the words on the page. The agent works in four steps: plan, check the plan against the brief, build, then critique the result."],"useItWhen":["You are building a landing page or app screen and want it to look designed for your product, not generated.","You are reworking an interface that feels generic.","You want the agent to choose fonts, colors and layout with a reason behind each choice.","You need a first design direction to react to before you refine it yourself."],"notFor":["You want existing UI code checked for accessibility and UX rules: use Web Design Guidelines.","You want React or Next.js performance rules: use React Best Practices."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill frontend-design"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","faq":[{"q":"Does the frontend-design skill work in Codex or Cursor?","a":"Yes. It is Apache-2.0 and its instructions use no tool that only one agent has, so any agent that reads Agent Skills can load it; the open skills CLI listed it on 2026-10-01. Anthropic itself documents Claude Code, Claude.ai and the Claude API."},{"q":"Which design defaults does frontend-design tell the agent to avoid?","a":"The ones that make generated pages look alike: a cream background with a serif headline and a terracotta accent, near-black with one acid-green accent, identical rounded cards with the same soft shadow, all-caps labels above every heading, and fade-ins on every section, unless your brief asks for one."}],"sourceUpdated":"2026-09-03","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/getsentry/gha-security-review","owner":"getsentry","slug":"gha-security-review","name":"GitHub Actions security review","skillName":"gha-security-review","title":"gha-security-review skill: audit GitHub Actions workflows","summary":"gha-security-review is Sentry's agent skill for auditing GitHub Actions workflows for attacks an outsider can run, such as pwn requests, expression injection and credential theft, with a concrete exploit for each finding.","category":"security","source":"getsentry/skills","path":"skills/gha-security-review","license":"Apache-2.0","worksWith":["Claude Code","Cursor","Cline","GitHub Copilot"],"agentNote":"The SKILL.md lists Claude Code tool names in allowed-tools (Read, Grep, Glob, Bash, Task). Other agents ignore that line and use their own file and shell tools.","cliChecked":"2026-10-01","whatItDoes":["gha-security-review is an agent skill from Sentry that reviews GitHub Actions workflows, composite actions and the files they load. Its threat model is an outside attacker with no write access, who can open pull requests from forks, file issues and post comments. Anything that needs write access to exploit, such as workflow_dispatch inputs, is out of scope and not reported.","The gha-security-review skill checks eight classes of problem: pull_request_target workflows that run fork code, untrusted text such as PR titles or branch names expanded inside run steps, comment-triggered commands with no author check, long-lived tokens reachable by untrusted code, poisoned config files including AGENTS.md and CLAUDE.md read by AI agents in CI, unpinned third-party actions in privileged jobs, broad permissions, and self-hosted runners and caches.","Every high-confidence finding must name the entry point, the payload, how it runs, the impact and a short proof-of-concept sketch. If the agent cannot build all five, the finding is downgraded to needs verification, and theoretical issues are dropped. The attack patterns draw on a 2025 analysis of a real campaign against GitHub Actions."],"useItWhen":["You use pull_request_target and want to know whether a fork can run code with your secrets.","You run an AI agent in CI and worry a pull request could inject instructions into it.","You want your workflows audited before making a private repository public.","A bot command triggered by comments runs in your workflows and you want it checked."],"notFor":["Application code vulnerabilities: use security-review from the same repository."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add getsentry/skills --skill gha-security-review"},{"label":"Or, for Claude Code, add Sentry's plugin marketplace","command":"claude plugin marketplace add getsentry/skills"},{"label":"Then install the sentry-skills plugin, which holds all 28 skills","command":"claude plugin install sentry-skills@sentry-skills"}],"installNote":"Restart Claude Code after installing the plugin. Sentry's README says its skills were written for Sentry employees, so some carry Sentry conventions.","sourceUpdated":"2026-09-29","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/openai/gh-address-comments","owner":"openai","slug":"gh-address-comments","name":"GitHub address comments","skillName":"gh-address-comments","title":"gh-address-comments skill: work through PR review comments","summary":"gh-address-comments is OpenAI's agent skill for pull request review feedback: the agent fetches every comment and review thread on the open PR, numbers them, and fixes only the ones you pick.","category":"code-review-and-git","source":"openai/skills","path":"skills/.curated/gh-address-comments","license":"Apache-2.0","licenseNote":"The skill folder carries its own Apache-2.0 LICENSE.txt, so you may use, change and share it, including at work.","worksWith":["Codex","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"The instructions mention Codex's sandbox escalation (sandbox_permissions=require_escalated) for gh commands; other agents simply need network access for gh. OpenAI marked openai/skills deprecated on 2026-06-22 in favour of openai/plugins.","cliChecked":"2026-10-01","whatItDoes":["GitHub address comments is a short skill from OpenAI's skills catalog for answering review feedback on a pull request. It works on the open pull request for your current branch, using the GitHub CLI. The agent checks that gh is logged in first and asks you to run gh auth login if it is not.","A bundled script, fetch_comments.py, prints every comment and review thread on the pull request. The agent numbers them and writes a one-line summary of what each fix would take, then asks which numbers you want handled. Nothing is changed until you choose.","After you pick, the agent applies fixes for the selected comments only. If gh hits an authentication or rate limit problem partway through, the skill tells the agent to ask you to log in again and retry, rather than guess."],"useItWhen":["Your pull request has a long review and you want every comment listed in one place.","You want to choose which review comments the agent fixes and skip the rest.","You want a quick summary of what each requested change would involve."],"notFor":["Writing a code review of someone else's pull request: the skill only answers comments already on yours."],"requirements":["GitHub CLI (gh) logged in with repo and workflow scopes","Python, for the bundled comment script"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add openai/skills --skill gh-address-comments"},{"label":"Or, inside Codex, use the built-in skill installer","command":"$skill-installer gh-address-comments"}],"faq":[{"q":"Is openai/skills still the place to get this skill?","a":"OpenAI's README has marked openai/skills deprecated since 2026-06-22 and points to openai/plugins for current Codex skills. gh-address-comments still installs from openai/skills with the skills CLI as of 2026-10-01."},{"q":"What license does gh-address-comments use?","a":"Apache-2.0, from the LICENSE.txt inside the skill folder, read on 2026-10-01. You may use and adapt it outside Codex."}],"sourceUpdated":"2026-02-02","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/openai/gh-fix-ci","owner":"openai","slug":"gh-fix-ci","name":"GitHub fix CI","skillName":"gh-fix-ci","title":"gh-fix-ci skill: debug failing GitHub Actions checks on a PR","summary":"gh-fix-ci is OpenAI's agent skill for failing pull request checks: the agent uses the GitHub CLI to find failing GitHub Actions runs, pulls the logs, summarizes the failure and proposes a fix plan.","category":"code-review-and-git","source":"openai/skills","path":"skills/.curated/gh-fix-ci","license":"Apache-2.0","licenseNote":"The skill folder carries its own Apache-2.0 LICENSE.txt, so you may use, change and share it, including at work.","worksWith":["Codex","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"Step 6 tells the agent to draft its plan with a create-plan skill, which lives in the repository's experimental folder, not next to this skill; without it the agent writes the plan inline. OpenAI marked openai/skills deprecated on 2026-06-22 in favour of openai/plugins.","cliChecked":"2026-10-01","whatItDoes":["GitHub fix CI is a skill from OpenAI's skills catalog for a pull request whose checks have gone red. The agent first confirms the GitHub CLI is logged in, finds the pull request for the current branch or the one you name, and lists its checks. It works only on checks that run in GitHub Actions.","For each failing check, a bundled Python script, inspect_pr_checks.py, fetches the run, downloads the job log and cuts out the part around the failure. The script copes with changes in gh field names and falls back to job logs when a run is still in progress. The agent then reports each failing check by name, with its run link and a short log snippet, and says plainly when logs are missing.","The skill keeps a human gate before code changes. The agent drafts a fix plan and waits for your approval, then applies it, summarizes the diff and tests, and asks whether to open a pull request. Afterwards it suggests rerunning the tests and gh pr checks to confirm the fix."],"useItWhen":["Your pull request has failing checks and you want the cause found in the logs.","You want a summary of every red check with the relevant log lines.","You want the agent to propose a CI fix but not change code until you approve."],"notFor":["Checks that run outside GitHub Actions, such as Buildkite: the skill only reports their link."],"requirements":["GitHub CLI (gh) logged in with repo and workflow scopes","Python, for the bundled log script"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add openai/skills --skill gh-fix-ci"},{"label":"Or, inside Codex, use the built-in skill installer","command":"$skill-installer gh-fix-ci"}],"faq":[{"q":"Is OpenAI's gh-fix-ci skill deprecated?","a":"The repository it lives in, openai/skills, has been marked deprecated in its README since 2026-06-22, with openai/plugins named as the replacement. The skill itself still installs with the skills CLI as of 2026-10-01."},{"q":"Does gh-fix-ci change my code on its own?","a":"No. It reads checks and logs, writes a fix plan, and only implements after you approve that plan."}],"sourceUpdated":"2026-02-02","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/googleworkspace/gws-gmail","owner":"googleworkspace","slug":"gws-gmail","name":"Gmail (gws)","skillName":"gws-gmail","title":"gws-gmail skill: send, read and manage Gmail from an AI agent","summary":"gws-gmail is the Gmail agent skill in Google Workspace's gws CLI repository: it lets an agent send, read, reply to, forward and triage Gmail messages through the gws command line.","category":"writing","source":"googleworkspace/cli","path":"skills/gws-gmail","license":"Apache-2.0","licenseNote":"Apache-2.0 per the repository. The README says gws is not an officially supported Google product and expects breaking changes before version 1.0.","worksWith":["Gemini CLI","OpenClaw","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"gws-gmail tells the agent to read a companion skill, gws-shared, for login, global flags and safety rules first, so install both.","cliChecked":"2026-10-01","whatItDoes":["Gmail (gws) is one of more than 100 skills in the googleworkspace/cli repository, which builds a single gws command for every Google Workspace API from Google's Discovery Service. This skill covers Gmail. It gives the agent helper commands to send, reply, reply to all, forward, read a message's body or headers, show an unread inbox summary, and stream new mail as it arrives.","Beyond the helpers, the agent can reach the full Gmail API: drafts, labels, threads, messages, history and settings. Before calling a method, it is told to run gws schema on that method to see the required parameters, then pass them as JSON. Helper commands have their own small skills, such as gws-gmail-send and gws-gmail-triage, linked from this one.","The companion gws-shared skill sets the safety rules: never print tokens, always confirm with you before a write or delete, prefer a dry run for destructive actions, and optionally screen responses through Google's Model Armor. Output is JSON by default, with table, YAML and CSV available."],"useItWhen":["You want your agent to summarize unread email in your inbox.","You want a reply drafted and sent in the right thread after you approve it.","You want labels or filters managed from the terminal.","You want an agent to watch for new mail and act on it."],"notFor":["Quick setup on a personal Gmail account: you need your own Google Cloud OAuth app, and unverified apps are limited to about 25 scopes, so the broad scope preset fails."],"requirements":["The gws CLI (a release binary, or npm install -g @googleworkspace/cli)","A Google Cloud project with OAuth credentials, set up with gws auth setup or the Cloud Console"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add googleworkspace/cli --skill gws-gmail"},{"label":"Also install gws-shared, which holds login, flags and safety rules","command":"npx skills add googleworkspace/cli --skill gws-shared"},{"label":"Install the gws CLI, which the skill drives","command":"npm install -g @googleworkspace/cli"},{"label":"Or, in Gemini CLI, install the gws extension after running gws auth setup","command":"gemini extensions install https://github.com/googleworkspace/cli"}],"faq":[{"q":"Is the gws Gmail skill an official Google product?","a":"It is published from the googleworkspace GitHub account, but the README says on 2026-10-01 that gws is not an officially supported Google product and is under active development with breaking changes expected."},{"q":"Will the gws Gmail skill send email without asking me?","a":"The shared rules it depends on tell the agent to confirm with you before any write or delete command, and to prefer a dry run for destructive actions."}],"sourceUpdated":"2026-03-31","version":"0.22.5","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/mattpocock/grill-me","owner":"mattpocock","slug":"grill-me","name":"Grill me","skillName":"grill-me","title":"grill-me skill: stress-test a plan with an agent interview","summary":"grill-me is Matt Pocock's agent skill that turns your agent into a relentless interviewer, asking numbered rounds of questions about a plan or design until every decision is settled.","category":"planning-and-workflow","source":"mattpocock/skills","path":"skills/productivity/grill-me","license":"MIT","worksWith":["Claude Code","Codex","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"grill-me is a one-line entry point that only runs when you call it, then hands off to a second skill named grilling. It phrases that hand-off as a call to the Skill tool, so agents without one need grilling installed beside it and may need to be pointed at it.","cliChecked":"2026-10-01","whatItDoes":["Grill me is a skill in Matt Pocock's collection, which his README calls his most popular, for aligning with an agent before any work starts. You call it by name, for example /grill-me, and describe a plan, decision or idea. The agent then interviews you instead of building, so gaps in the plan show up as questions rather than as wrong code later.","The interview logic lives in the companion grilling skill. The agent maps your plan as a tree of decisions, where each choice opens the choices that depend on it. Each round it asks every question it can already ask, numbered, with its own recommended answer under each one, then waits for your replies before recomputing the next round.","Facts are the agent's job and decisions are yours. When a question needs a fact from the project, the agent looks it up, for example with a sub-agent, instead of asking you. The session ends only when no open branch is left, and the agent does not act until you confirm you share the same understanding."],"useItWhen":["You have a feature idea and want the agent to find the holes before it writes code.","You are about to make a decision and want it questioned from every side.","Your agent keeps building something other than what you meant.","You want a written list of decisions, each with a recommended answer you can accept or change."],"notFor":["Code changes where the plan should also be checked against project docs: the author points to grill-with-docs in the same repository."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add mattpocock/skills --skill grill-me"},{"label":"Also install grilling, which holds the interview steps grill-me hands off to","command":"npx skills add mattpocock/skills --skill grilling"},{"label":"Or, in Claude Code, install the whole collection from the official marketplace","command":"/plugin install mattpocock-skills"}],"sourceUpdated":"2026-08-15","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/huggingface/hf-cli","owner":"huggingface","slug":"hf-cli","name":"Hugging Face CLI","skillName":"hf-cli","title":"hf-cli skill: Hugging Face models, datasets and Hub jobs","summary":"hf-cli is Hugging Face's agent skill for its hf command-line tool: it lets an agent download and upload models and datasets, manage repos, Spaces and buckets, read papers and run jobs on the Hub.","category":"science-and-research","source":"huggingface/skills","path":"skills/hf-cli","license":"Apache-2.0","worksWith":["Claude Code","Codex","Gemini CLI","Cursor","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["Hugging Face CLI is the skill Hugging Face recommends installing first. It is generated from the hf tool itself, version 2.1.1 of huggingface_hub on 2026-10-01, and lists every hf command with its flags so the agent can work with the Hugging Face Hub without guessing. It also tells the agent that hf replaces the deprecated huggingface-cli command.","The command list covers downloading and uploading files, copying and syncing between local folders, repos and buckets, logging in and switching tokens, the local cache, collections, dataset queries, discussions and pull requests on Hub repos, Inference Endpoints, Spaces settings, webhooks and paper search. It also covers hf jobs for running scripts on Hugging Face hardware and an experimental sandbox command.","A short section explains hf-mount, a separate tool that mounts a Hub repo or bucket as a local folder and fetches files on demand. Tips at the end tell the agent to read each command's help, authenticate with the HF_TOKEN environment variable, and update the CLI with hf update."],"useItWhen":["You want an agent to download a model or dataset from the Hugging Face Hub.","You want trained weights or results uploaded to a Hub repo.","You need a Space's hardware, secrets or variables changed from the terminal.","You want to search Hugging Face papers or browse models for a task."],"notFor":["Writing training code: Hugging Face has separate skills such as huggingface-llm-trainer for that."],"requirements":["The hf CLI (curl -LsSf https://hf.co/cli/install.sh | bash -s)","A Hugging Face account and token for private repos, uploads and jobs"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add huggingface/skills --skill hf-cli"},{"label":"Or, in Claude Code, add Hugging Face's plugin marketplace","command":"/plugin marketplace add huggingface/skills"},{"label":"Then install the hf-cli skill","command":"/plugin install hf-cli@huggingface/skills"},{"label":"Or, in Gemini CLI, install Hugging Face's extension","command":"gemini extensions install https://github.com/huggingface/skills.git --consent"}],"sourceUpdated":"2026-10-01","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/huggingface/huggingface-llm-trainer","owner":"huggingface","slug":"huggingface-llm-trainer","name":"Hugging Face LLM trainer","skillName":"huggingface-llm-trainer","title":"huggingface-llm-trainer skill: fine-tune LLMs on cloud GPUs","summary":"huggingface-llm-trainer is Hugging Face's agent skill for fine-tuning language and vision models on its cloud GPUs, using TRL or Unsloth on Hugging Face Jobs, with no local GPU needed.","category":"science-and-research","source":"huggingface/skills","path":"skills/huggingface-llm-trainer","license":"Apache-2.0","licenseNote":"The SKILL.md says its terms are in a LICENSE.txt, but no such file is in the skill folder on 2026-10-01; the repository's LICENSE is Apache-2.0.","worksWith":["Claude Code","Codex","Gemini CLI","Cursor","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"The skill tells the agent to submit jobs through the hf_jobs tool of the Hugging Face MCP server. Without that server, it falls back to the hf jobs command in the terminal.","cliChecked":"2026-10-01","whatItDoes":["Hugging Face LLM trainer is a skill for training models on Hugging Face Jobs, its managed cloud GPU service. It covers supervised fine-tuning, DPO preference training, GRPO reinforcement learning and reward models with the TRL library, and suggests Unsloth for tight GPU memory, models over 13B parameters or vision-language models. It can also convert a trained model to GGUF for Ollama, LM Studio or llama.cpp.","The agent writes a training script in the uv inline-dependency format, adds Trackio for live monitoring, and submits it as a job right away, then reports the job ID, a monitoring link and an estimated time. Bundled scripts estimate cost, inspect datasets before training and pick base models from benchmarks. A hardware table on 2026-10-01 runs from about 0.75 dollars an hour on a T4 to 10 to 20 dollars an hour for models above 13B.","Two failure modes get loud warnings. The training environment is wiped when a job ends, so the model must be pushed to the Hub with a write token passed as a secret, or all results are lost. And the default 30 minute timeout is too short for most runs, so the agent sets one to two hours or more."],"useItWhen":["You want to fine-tune a small open model on your own dataset without owning a GPU.","You want to try DPO or GRPO training and have the job script written for you.","You want a fine-tuned model converted to GGUF to run locally.","You want a cost estimate before starting a training run."],"notFor":["Free accounts: Hugging Face Jobs require a Pro, Team or Enterprise plan, and GPU time is billed per hour.","Training on your own local GPU: the skill is built around Hugging Face's cloud jobs."],"requirements":["A Hugging Face account on a Pro, Team or Enterprise plan","A Hugging Face token with write access","The Hugging Face MCP server, or the hf CLI for the hf jobs fallback"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add huggingface/skills --skill huggingface-llm-trainer"},{"label":"Or, with the hf CLI installed, add it the way Hugging Face's README describes","command":"hf skills add huggingface-llm-trainer"}],"faq":[{"q":"Do I need a paid Hugging Face plan to use the LLM trainer skill?","a":"Yes. The skill states on 2026-10-01 that Hugging Face Jobs require a Pro, Team or Enterprise plan, and GPU hours are charged on top, from about 0.75 dollars an hour on the smallest GPU in its table."},{"q":"What happens to my model when a Hugging Face training job ends?","a":"The job's environment is deleted. The skill makes the agent set push_to_hub and pass your HF_TOKEN as a secret so the trained model is saved to a Hub repo before that happens."}],"sourceUpdated":"2026-07-06","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/heygen-com/hyperframes","owner":"heygen-com","slug":"hyperframes","name":"HyperFrames","skillName":"hyperframes","title":"hyperframes skill: make videos from HTML with an AI agent","summary":"hyperframes is HeyGen's entry-point agent skill for HyperFrames, its open-source framework that renders HTML, CSS and animations into MP4 video; it captures your brief and routes to the right video workflow.","category":"video-and-media","source":"heygen-com/hyperframes","path":"skills/hyperframes","license":"Apache-2.0","worksWith":["Claude Code","Codex","Cursor","Gemini CLI","IBM Bob","OpenCode","GitHub Copilot","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["HyperFrames is the router skill for HeyGen's HyperFrames framework, where a video is an HTML file whose elements declare their timing in data attributes and whose animations can be scrubbed to any frame. The skill is meant to be read first for any request to make, edit, animate or render a video, and it maps the 21 skills the README lists on 2026-10-01.","The agent starts from project state. A specific edit or render on an existing project is done directly; a project with a BRIEF.md resumes its recorded workflow; a fresh request goes through a short intent interview, then one routing step. Workflows include product launch videos from a website, topic explainers, videos from a GitHub pull request, captions on talking-head footage, short motion graphics, beat-synced music videos, slide decks and ports from Remotion.","Domain skills such as hyperframes-core, hyperframes-animation and hyperframes-cli load only when a workflow needs them, and the router installs a workflow skill on demand. The skill also tells the agent to check whether the project's pinned HyperFrames version is behind before rendering, and to upgrade and run a check when it is."],"useItWhen":["You want a 30 to 90 second launch video made from your product's website.","You want a pull request turned into a short changelog video.","You have talking-head footage and want captions or designed overlays added.","You have a Remotion composition and want it ported to HTML."],"notFor":["Recording a plain browser session: the skill itself hands that case to other tools.","Projects that must stay in React video code: Remotion's own skills fit better."],"requirements":["Node.js 22 or newer","FFmpeg"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add heygen-com/hyperframes --skill hyperframes"},{"label":"Or install HyperFrames' core skill set from the current main branch, which HeyGen recommends for agents and scripts","command":"npx hyperframes skills update"},{"label":"Or, in Claude Code, add HeyGen's plugin marketplace","command":"claude plugin marketplace add heygen-com/hyperframes"},{"label":"Then install the plugin","command":"claude plugin install hyperframes@hyperframes"}],"installNote":"HeyGen notes that skills.sh can lag the main branch by hours, which is why it suggests npx hyperframes skills update when you need the newest copy.","sourceUpdated":"2026-09-28","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/pbakaus/impeccable","owner":"pbakaus","slug":"impeccable","name":"Impeccable","skillName":"impeccable","title":"impeccable skill: design, critique and polish frontend UI","summary":"Impeccable is Paul Bakaus's agent skill for frontend design: one skill with 24 commands such as audit, critique and polish, plus a launcher that runs deterministic checks for common AI design mistakes.","category":"frontend-and-design","source":"pbakaus/impeccable","path":".agents/skills/impeccable","license":"Apache-2.0","licenseNote":"Apache-2.0 per the repository and the license key in the Claude Code copy of the SKILL.md. The skill's launcher downloads a separate Impeccable engine binary on first run.","worksWith":["Claude Code","Cursor","Codex CLI","Gemini CLI","GitHub Copilot","Grok Build","Hermes Agent","OpenCode","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"The repository ships a separate copy of the skill per agent folder (.claude, .cursor, .agents and others); this page tracks the .agents copy. Commands are written as /impeccable plus a verb, which assumes an agent with slash commands; elsewhere you ask for the command in words.","cliChecked":"2026-10-01","whatItDoes":["Impeccable is a frontend design skill by Paul Bakaus that grew from Anthropic's frontend-design skill. It covers websites, landing pages, dashboards, app screens, forms and onboarding. You start a project with init, which records durable facts about the product, audience and voice in PRODUCT.md so later commands do not have to ask again. A document command writes DESIGN.md from existing code.","Everything runs through one entry point, for example /impeccable audit blog. The README lists 24 commands on 2026-10-01, including shape to plan before coding, critique for a UX review, audit for accessibility, performance and responsive checks, polish for a final pass, harden for errors and edge cases, and bolder or quieter to change the intensity. A live mode lets you iterate on single elements in the browser.","Before editing, the agent runs a bundled launcher that loads the product and design context, then reads a craft-floor file with the skill's quality bar and bans, such as overused fonts, gray text on colored backgrounds and cards nested in cards. The README counts 61 deterministic detector rules that run without a model or API key, through the CLI or a browser extension."],"useItWhen":["Your agent's UI looks like every other AI-made SaaS page and you want it to look like yours.","You want a design review of an existing screen with concrete fixes.","You want product context written down once and reused by every design task.","You need an accessibility and responsive audit before shipping a page."],"notFor":["Backend or non-UI work: the skill says so in its own description."],"requirements":["Network access on first run, when the launcher downloads the Impeccable engine binary to ~/.impeccable/bin unless it ships beside it"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add pbakaus/impeccable --skill impeccable"},{"label":"Or use the publisher's installer, which also sets up hooks for supported agents","command":"npx impeccable install"},{"label":"Or, in Claude Code, add the publisher's plugin marketplace, then install Impeccable from /plugin","command":"/plugin marketplace add pbakaus/impeccable"}],"installNote":"After installing, run /impeccable init once in your project so the skill can write PRODUCT.md.","sourceUpdated":"2026-10-02","version":"4.4.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:02:16+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/internal-comms","owner":"anthropics","slug":"internal-comms","name":"Internal Comms","skillName":"internal-comms","title":"internal-comms skill: status reports, 3P updates, newsletters","summary":"Internal Comms is Anthropic's open-source agent skill for writing inside a company: the agent drafts 3P updates, company newsletters, FAQs and other internal messages in set formats.","category":"writing","source":"anthropics/skills","path":"skills/internal-comms","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"cliChecked":"2026-10-01","whatItDoes":["Internal Comms is an Anthropic skill for internal company writing. It loads when you ask for a company newsletter, an incident report, a FAQ, a project or status update, a 3P update (progress, plans, problems) or an update for leadership. The agent identifies the type of message, loads the matching guide from the skill's examples folder and follows its format and tone. The newsletter guide, for example, assumes a company of more than 1,000 people.","Internal Comms has four guides. A 3P update (Progress, Plans, Problems) covers one team's week and should take under a minute to read. A company newsletter runs about 20 to 25 short bullets in sections, written in the first person plural. FAQ answers turn the questions many employees are asking into one-sentence questions with short answers. A general guide covers everything else, and the agent first asks about audience, purpose, tone and format.","When tools are connected, the Internal Comms guides tell the agent to gather material from Slack, email, Google Drive and the calendar, favoring posts and documents with many reactions, replies or views. Without that access, the agent asks you what to cover and fits your notes to the format. For a 3P update it confirms the team name and time period first."],"useItWhen":["You owe leadership a weekly team update in a fixed Progress, Plans, Problems format.","You are putting together a company newsletter from what happened in Slack and email this week.","You want an FAQ of the questions people across the company keep asking."],"notFor":["You are writing a long spec, proposal or design doc: use Doc Co-Authoring."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill internal-comms"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/getsentry/iterate-pr","owner":"getsentry","slug":"iterate-pr","name":"Iterate on PR","skillName":"iterate-pr","title":"iterate-pr skill: fix CI failures and review comments on a PR","summary":"iterate-pr is Sentry's agent skill that keeps working on a GitHub pull request until CI passes and high and medium priority review comments are fixed, using bundled scripts and the gh CLI.","category":"code-review-and-git","source":"getsentry/skills","path":"skills/iterate-pr","license":"Apache-2.0","worksWith":["Claude Code","Cursor","Cline","GitHub Copilot"],"agentNote":"The workflow runs in any agent with a shell; one optional step uses Claude Code's Monitor tool to watch CI, with a polling loop as the default.","cliChecked":"2026-10-01","whatItDoes":["iterate-pr is an agent skill from Sentry that drives a GitHub pull request to green. The agent finds the pull request for the current branch, then uses four bundled Python scripts run with uv: one fetches CI check results with failure snippets, one sorts review comments into high, medium, low, bot and resolved, one monitors checks, and one replies to review threads.","The iterate-pr loop fixes high and medium comments directly and asks the user which low-priority ones to handle. For a failed check it reads the full log, states the cause before editing, searches for the same problem elsewhere and fixes the root cause. It reruns the relevant test or linter locally, commits and pushes, then polls every 30 seconds and checks for new feedback once CI passes.","The skill knows which checks to wait for, the review bots Sentry, Warden, Cursor, Bugbot, Seer and CodeQL, and which never to wait for: human approval, draft status, required reviews and Codecov. It stops and reports when only human gates remain, when the branch needs a rebase, or when the same failure survives two attempts."],"useItWhen":["Your pull request has failing CI and you want an agent to fix it and push.","Reviewers and review bots left comments and you want the real issues fixed.","You want an agent to keep watching CI until the checks pass.","You want low-priority nitpicks listed so you can choose which to address."],"notFor":["Merging or approving: the skill never waits for or performs human approval, draft changes or merges."],"requirements":["The GitHub CLI (gh), signed in","uv to run the bundled Python scripts","An open pull request for the current branch"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add getsentry/skills --skill iterate-pr"},{"label":"Or, for Claude Code, add Sentry's plugin marketplace","command":"claude plugin marketplace add getsentry/skills"},{"label":"Then install the sentry-skills plugin, which holds all 28 skills","command":"claude plugin install sentry-skills@sentry-skills"}],"installNote":"Restart Claude Code after installing the plugin. Sentry's README says its skills were written for Sentry employees, so some carry Sentry conventions.","sourceUpdated":"2026-08-24","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/mcp-builder","owner":"anthropics","slug":"mcp-builder","name":"MCP Builder","skillName":"mcp-builder","title":"mcp-builder skill: build and test an MCP server for any API","summary":"MCP Builder is Anthropic's open-source agent skill for building MCP servers in TypeScript or Python, from API research and tool design to testing and a 10-question evaluation.","category":"building-skills-and-mcp","source":"anthropics/skills","path":"skills/mcp-builder","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"agentNote":"Building the server works in any agent, but the evaluation script calls Claude through the Anthropic API, so that step needs an ANTHROPIC_API_KEY whichever agent you use.","cliChecked":"2026-10-01","whatItDoes":["MCP Builder is Anthropic's guide to building MCP (Model Context Protocol) servers, which let AI models use an external API or service through tools. It loads when you build an MCP server in Python with FastMCP or in TypeScript with the official MCP SDK. It recommends TypeScript, with streamable HTTP for remote servers and stdio for local ones.","With MCP Builder, the agent works in four phases. It studies the MCP specification and the service's API, then plans which endpoints to cover, starting with the most common operations. It implements tools with Zod or Pydantic input schemas, consistent name prefixes, pagination, actionable error messages and hints such as read-only or destructive. Then it builds the server and tests it with the MCP Inspector. Reference files cover best practices and each language.","The last MCP Builder phase is an evaluation: 10 realistic, read-only questions, each with one verifiable answer, saved as an XML file. A bundled Python script runs those questions against your server through Claude, over stdio, SSE or HTTP, and reports the results along with Claude's feedback on the tools. It needs the anthropic and mcp Python packages and an ANTHROPIC_API_KEY."],"useItWhen":["You want Claude or another agent to use your product's API and need an MCP server for it.","You are designing MCP tools and want naming, pagination and error handling conventions to follow.","You built an MCP server and want to test whether a model can answer real questions with it."],"notFor":["You want to package know-how or a workflow rather than tools: write a skill with Skill Creator or Writing Skills."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill mcp-builder"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","faq":[{"q":"Do I need an Anthropic API key for mcp-builder?","a":"Only for the evaluation: its Python script runs your 10 questions through Claude and needs the anthropic and mcp packages plus an ANTHROPIC_API_KEY. Researching, designing and building the server works without one."},{"q":"Should I build my MCP server in Python or TypeScript?","a":"The skill supports both, FastMCP for Python and the official MCP SDK for TypeScript, and recommends TypeScript, with streamable HTTP for remote servers and stdio for local ones."}],"sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/trailofbits/mutation-testing","owner":"trailofbits","slug":"mutation-testing","name":"Mutation testing","skillName":"mutation-testing","title":"mutation-testing skill: run mewt or muton and analyze survivors","summary":"mutation-testing is Trail of Bits' agent skill for configuring mutation testing campaigns with its mewt or muton tools, reading surviving mutants, telling equivalent mutants from real test gaps, and hunting bugs they expose.","category":"testing-and-debugging","source":"trailofbits/skills","path":"plugins/mutation-testing/skills/mutation-testing","license":"CC-BY-SA-4.0","licenseNote":"CC-BY-SA-4.0 is a share-alike license. You can copy the skill into a project and change it, even for commercial work, if you credit Trail of Bits, link the license and say what you changed. If you share a changed version, for example in a public repository, it must stay under CC-BY-SA-4.0. It is a content license, not a software license, so ask your legal team before bundling it inside a closed product.","worksWith":["Claude Code","Codex","ChatGPT workspace marketplace"],"cliChecked":"2026-10-01","whatItDoes":["mutation-testing is an agent skill from security firm Trail of Bits for mutation testing: making small deliberate changes to source code and checking whether the test suite notices. It works with Trail of Bits' own tools, mewt and muton, which share one interface, and follows the mewt 4.x commands, while telling the agent to trust the installed tool's help output over its examples.","The mutation-testing skill routes each request to one of three workflows. Configuration sets up a campaign, picks targets, measures test time and trims a campaign that would take too long. Analysis reads each surviving mutant, separates equivalent changes from real testing gaps and recommends specific assertions or inputs in a report. Bug hunting uses survivors to point at untested code and requires a reproduced proof before calling anything a bug.","The skill explains what each mutation type means when it survives and warns that a severity label describes the mutation and not the danger to you: a surviving low-severity mutant in fee logic can outweigh a high-severity one in a log line. Analysis also accepts results from slither-mutate, mull and dextool-mutate, and has notes for Solidity, Move, Cairo and Solana code."],"useItWhen":["Your line coverage is high but you doubt your tests would catch real bugs.","You want mewt or muton set up for a Rust, TypeScript or smart contract project.","Your mutation campaign would take days and you want it scoped down.","You have a list of surviving mutants and want to know which ones matter."],"notFor":["Plain test or line coverage questions with no mutation testing involved: the skill itself says not to load for those."],"requirements":["mewt or muton plus a runnable test suite for running campaigns; analyzing saved results needs neither tool"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add trailofbits/skills --skill mutation-testing"},{"label":"Or, in Claude Code, add Trail of Bits' plugin marketplace","command":"/plugin marketplace add trailofbits/skills"},{"label":"Or, in Codex, add the same marketplace from your terminal","command":"codex plugin marketplace add trailofbits/skills"},{"label":"Then install the mutation-testing plugin in Codex","command":"codex plugin add mutation-testing@trailofbits"}],"installNote":"In Claude Code, open /plugin menu after adding the marketplace and install the mutation-testing plugin. For a ChatGPT workspace, Trail of Bits' README says to import the repository's .claude-plugin/marketplace.json.","faq":[{"q":"Can I copy Trail of Bits' mutation-testing skill into my own repository?","a":"Yes, with conditions. The skills are CC-BY-SA-4.0. You may copy and adapt them, even commercially, if you credit Trail of Bits, link the license and note your changes. Any changed version you share must carry the same CC-BY-SA-4.0 license. Using the skill unchanged in your own work does not put your code under that license."},{"q":"Does the mutation-testing skill work with tools other than mewt and muton?","a":"Partly. Campaign setup is written for mewt and muton only. The analysis workflow also reads results from slither-mutate, mull and dextool-mutate, and saved results can be analyzed without any mutation tool installed."}],"sourceUpdated":"2026-09-16","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/neondatabase/neon-postgres","owner":"neondatabase","slug":"neon-postgres","name":"Neon Postgres","skillName":"neon-postgres","title":"neon-postgres skill: connect, migrate, branch and tune Neon","summary":"neon-postgres is Neon's agent skill for working with Postgres on Neon: connection strings, pooled or direct connections, migrations, branching, built-in diagnostics, autoscaling, scale to zero and search.","category":"databases","source":"neondatabase/agent-skills","path":"skills/neon-postgres","license":"Apache-2.0","worksWith":["Claude Code","Cursor","Kimi Code CLI","Pi","Other Agent Plugins v1 clients"],"agentNote":"The skill works through the Neon CLI by default or the Neon MCP server; the Claude Code, Cursor and Kimi Code plugins install the MCP server, the skills CLI line does not.","cliChecked":"2026-10-01","whatItDoes":["neon-postgres is an agent skill from Neon for Postgres databases hosted on Neon, which Neon now calls Lakebase Postgres. It tells the agent to reuse a DATABASE_URL already present in the prompt, environment or repo, and otherwise to pick or create a project with the Neon CLI or MCP server. A table decides between pooled connections, for web apps and serverless functions, and direct ones, for migrations, dumps, replication and LISTEN.","The neon-postgres skill gives driver choices per platform: node-postgres with Fluid compute on Vercel, node-postgres with Hyperdrive on Cloudflare, and Neon's serverless driver for edge runtimes. Migrations are managed as code through the project's ORM, tested first on a branch of production data and run over a direct connection. Branches are copy-on-write clones, each with its own compute.","For slow queries the skill maps symptoms to read-only neon inspect checks such as outliers, unused-indexes, locks, bloat and cache hit rate, with warnings on how not to misread them, and hands query and index work to a separate postgres-best-practices skill. Other topics include read replicas, logical replication, instant restore, IP allow lists, search, and how autoscaling and scale to zero work."],"useItWhen":["You are connecting an app to Neon and are unsure whether to use the pooled or the direct URL.","You want to test a schema migration on a copy of production data first.","Your Neon database got slow and you want the built-in diagnostics run.","You want to know how scale to zero and cold starts affect your app."],"notFor":["Starting a new backend with auth, files or functions: Neon routes that to its parent neon skill in the same repository."],"requirements":["A Neon account or an existing Neon DATABASE_URL; the Neon CLI or the Neon MCP server for project and branch actions"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add neondatabase/agent-skills --skill neon-postgres"},{"label":"Or, in Claude Code, add Neon's plugin marketplace","command":"/plugin marketplace add neondatabase/agent-skills"},{"label":"Then install the neon-postgres plugin (Neon's skills plus the Neon MCP server)","command":"/plugin install neon-postgres@neon"},{"label":"Or, in Cursor, run this in chat","command":"/add-plugin neon-postgres"}],"installNote":"With the plugin routes, Neon asks you to sign in through OAuth the first time an MCP tool runs.","sourceUpdated":"2026-09-23","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/kepano/obsidian-markdown","owner":"kepano","slug":"obsidian-markdown","name":"Obsidian Markdown","skillName":"obsidian-markdown","title":"obsidian-markdown skill: Obsidian wikilinks, callouts, properties","summary":"obsidian-markdown is Steph Ango's agent skill for writing Obsidian Flavored Markdown: notes with properties, wikilinks, embeds, callouts and block references that render correctly in an Obsidian vault.","category":"documents-and-office","source":"kepano/obsidian-skills","path":"skills/obsidian-markdown","license":"MIT","worksWith":["Claude Code","Codex","OpenCode","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["Obsidian Markdown is a skill in Steph Ango's obsidian-skills repository for agents that write notes into an Obsidian vault. Obsidian extends CommonMark and GitHub Flavored Markdown with its own syntax, and the skill teaches only those extensions, assuming the agent already knows headings, lists and tables.","Its note workflow starts with frontmatter properties such as title, tags and aliases, then the body. Notes inside the vault are linked with wikilinks, including links to a heading or a block ID, while external URLs keep normal Markdown links, because Obsidian updates wikilinks when notes are renamed. Embeds pull in other notes, sections, images at a set width, or a single PDF page.","Callouts get their own section, with the common types such as note, tip, warning and bug, plus foldable callouts that start collapsed or expanded. Reference files list every property type, embed type and callout alias. The last step tells the agent to check that the note renders in Obsidian's reading view."],"useItWhen":["You let an agent write or tidy notes in your Obsidian vault.","You want links between notes as wikilinks that survive renames.","You want meeting notes or summaries with callouts and tags Obsidian understands.","Your agent writes frontmatter that Obsidian does not read as properties.","You want images or a single PDF page embedded in a note at a set size."],"notFor":["Obsidian Bases or JSON Canvas files: the same repository has separate obsidian-bases and json-canvas skills."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add kepano/obsidian-skills --skill obsidian-markdown"},{"label":"Or, in Claude Code, add the publisher's plugin marketplace","command":"/plugin marketplace add kepano/obsidian-skills"},{"label":"Then install the obsidian plugin","command":"/plugin install obsidian@obsidian-skills"}],"sourceUpdated":"2026-03-22","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/pdf","owner":"anthropics","slug":"pdf","name":"PDF","skillName":"pdf","title":"pdf skill: Anthropic's tool to merge, split, fill and read PDFs","summary":"PDF is Anthropic's proprietary agent skill for reading, creating and editing PDF files, from pulling out tables to filling forms and OCR; its license limits it to Claude and Anthropic's own products.","category":"documents-and-office","source":"anthropics/skills","path":"skills/pdf","license":"Proprietary: use only within Anthropic's services, no copying or redistribution","licenseNote":"The LICENSE.txt in the pdf folder (read 2026-10-01) puts use under your agreement with Anthropic and forbids extracting the files from Anthropic's services, keeping copies outside them, copying them, making derivative works and redistributing them. Anthropic's README calls its four document skills source-available, not open source. You may read the files on GitHub, but you may not copy them into Codex, Cursor or any other agent, or publish a changed version.","worksWith":["Claude Code","Claude.ai","Claude API"],"whatItDoes":["The PDF skill teaches Claude to work with PDF files using Python libraries such as pypdf and command-line tools. It covers reading text and tables, merging and splitting files, rotating pages, adding watermarks, creating new PDFs, encrypting and decrypting, extracting images, and running OCR on scanned pages so they become searchable.","Filling in forms has its own guide in the PDF skill and bundled scripts: they check whether a form has fillable fields, read the fields and where they sit on the page, fill them in, and turn pages into images to validate the result.","PDF is one of Anthropic's four document skills, next to Word, PowerPoint and Excel. Unlike most skills in Anthropic's repository it is not open source: the license allows use only within Anthropic's own services and forbids copying, changing or redistributing it. That is why this page lists only Anthropic's install routes."],"useItWhen":["You need the text or tables out of a PDF, such as an invoice or a report.","You want to merge several PDFs into one, or split one into parts.","You have a PDF form to fill in from data you already have.","You have scanned pages and want them searchable."],"notFor":["You work in Codex, Cursor, Gemini CLI or another agent that is not Anthropic's: the pdf license bars copying the skill there.","You want an editable Word document: Anthropic's DOCX skill handles .docx files."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the document-skills plugin, which holds the Word, PDF, PowerPoint and Excel skills","command":"/plugin install document-skills@anthropic-agent-skills"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","faq":[{"q":"Can I use Anthropic's pdf skill in Codex, Cursor or ChatGPT?","a":"No. Its LICENSE.txt (read 2026-10-01) forbids keeping copies outside Anthropic's services, copying and redistribution, so it is meant for Claude.ai, Claude Code through Anthropic's plugin, and the Claude API. The open skills CLI does list it, but installing it into another agent copies the files, which the license forbids."},{"q":"Can the pdf skill make scanned PDFs searchable?","a":"Yes. It runs OCR on scanned pages so their text can be searched and extracted, next to text and table extraction, merging, splitting, rotating, watermarking, encryption and form filling with bundled scripts."}],"sourceUpdated":"2026-02-06","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/openai/pdf","owner":"openai","slug":"pdf","name":"PDF (OpenAI)","skillName":"pdf","title":"pdf skill: OpenAI's tool to create, read and check PDF files","summary":"pdf is OpenAI's open-source agent skill for making and reviewing PDF files: the agent builds them with Python libraries, renders each page to an image and checks the layout before delivering.","category":"documents-and-office","source":"openai/skills","path":"skills/.curated/pdf","license":"Apache-2.0","licenseNote":"The skill folder carries its own Apache-2.0 LICENSE.txt, so you may use, change and share it, including at work. GitHub shows no license for the repository as a whole because each skill has its own.","worksWith":["Codex","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"OpenAI marked openai/skills deprecated on 2026-06-22 and points to its openai/plugins repository for current Codex skills. This copy still installs, but may not get updates.","cliChecked":"2026-10-01","whatItDoes":["The PDF skill from OpenAI's skills catalog tells an agent how to handle PDF files where layout matters. For new documents it uses the reportlab Python library; for reading text and running quick checks it uses pdfplumber or pypdf, while warning that text extraction says nothing reliable about how the page looks.","Its main rule is visual review. After every meaningful change, the agent renders pages to PNG images with Poppler's pdftoppm and inspects them for clipped text, overlapping elements, broken tables, black squares or unreadable characters. It is told not to deliver until the latest render shows no defects, and to check headers, footers and page numbers.","The skill also fixes housekeeping: intermediate files go in tmp/pdfs and final files in output/pdf, with stable descriptive names. If a dependency cannot be installed, the agent is told to say which one is missing and how to install it, rather than carry on without checking the output."],"useItWhen":["You want an agent to produce a report or invoice as a PDF that looks right on every page.","You need to check a generated PDF for clipped text or broken tables before sending it.","You want text pulled out of a PDF for a quick check.","You want an open-licensed PDF skill you can copy and adapt."],"notFor":["Filling PDF forms, splitting or merging files, or OCR on scans: the skill does not cover them. Anthropic's PDF skill does, under a proprietary license."],"requirements":["Python with reportlab, pdfplumber and pypdf","Poppler (pdftoppm) for rendering pages to images"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add openai/skills --skill pdf"},{"label":"Or, inside Codex, use the built-in skill installer","command":"$skill-installer pdf"}],"installNote":"Restart Codex after installing so it picks up the new skill.","faq":[{"q":"Is OpenAI's skills repository still maintained?","a":"Its README has marked it deprecated since 2026-06-22. OpenAI now points to the openai/plugins repository for Codex skills and plugins. Skills in openai/skills still install as of 2026-10-01, but new work happens elsewhere."},{"q":"Can I use OpenAI's PDF skill outside Codex?","a":"Yes. Its folder has an Apache-2.0 license, and the skills CLI listed it for install on 2026-10-01. It only needs Python and Poppler, nothing specific to Codex."}],"sourceUpdated":"2026-02-02","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/microsoft/playwright-cli","owner":"microsoft","slug":"playwright-cli","name":"Playwright CLI","skillName":"playwright-cli","title":"playwright-cli skill: browser automation and web testing","summary":"playwright-cli is Microsoft's agent skill for driving a real browser from the terminal: open pages, click and fill by element reference, take snapshots and screenshots, mock requests and generate Playwright tests.","category":"testing-and-debugging","source":"microsoft/playwright-cli","path":"skills/playwright-cli","license":"Apache-2.0","worksWith":["Claude Code","GitHub Copilot","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"The SKILL.md frontmatter pre-approves only playwright-cli and npx playwright shell commands (allowed-tools). Agents that ignore that key still work but may ask permission for each command.","cliChecked":"2026-10-01","whatItDoes":["Playwright CLI is a skill from Microsoft's Playwright team that teaches a coding agent the playwright-cli command set. Instead of loading a browser tool schema into context, the agent runs short shell commands: open a page, read an accessibility snapshot with element references such as e15, then click, type, fill, drag, upload or select by that reference. Each command prints a fresh snapshot so the agent sees what changed.","The command set covers more than clicking. The agent can save and load login state, read and set cookies and local storage, emulate dark mode, reduced motion or a mobile device, mock network routes, read console messages and requests, record traces and videos, and run Playwright code directly. Named sessions keep separate browsers for separate projects, and a show command opens a dashboard where you can watch or step in.","Reference files bundled with the skill cover writing and generating Playwright tests, request mocking, storage state, tracing and video recording. The README positions the CLI against Playwright MCP: the CLI is meant to use fewer tokens for coding agents, while MCP suits long exploratory loops that need continuous page state."],"useItWhen":["You want your agent to test a sign-up or checkout flow in a real browser and take screenshots.","You need a Playwright test written from steps the agent just performed.","You want to check how a page behaves in dark mode, on a phone, or with an API mocked.","Your agent's browser tool eats too much context and you want a lighter option."],"notFor":["Long exploratory sessions where the agent should keep the page state in view all the time: Microsoft points to Playwright MCP for those."],"requirements":["Node.js 18 or newer","The @playwright/cli package installed globally"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add microsoft/playwright-cli --skill playwright-cli"},{"label":"Install the Playwright CLI itself, which the skill drives","command":"npm install -g @playwright/cli@latest"},{"label":"Or let the CLI install its own skills into your project","command":"playwright-cli install --skills"}],"sourceUpdated":"2026-09-28","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/supabase/supabase-postgres-best-practices","owner":"supabase","slug":"supabase-postgres-best-practices","name":"Postgres Best Practices","skillName":"supabase-postgres-best-practices","title":"supabase-postgres-best-practices skill: indexes, RLS, schemas","summary":"Postgres Best Practices is Supabase's agent skill of Postgres rules for any Postgres database, ranked by impact: indexes, connection pooling, row level security and schema design.","category":"databases","source":"supabase/agent-skills","path":"skills/supabase-postgres-best-practices","license":"MIT","worksWith":["Claude Code","GitHub Copilot","Cursor","Cline","other agents through the skills CLI (Supabase says 18 or more)"],"cliChecked":"2026-10-01","whatItDoes":["Postgres Best Practices is a rule set maintained by Supabase for Postgres running anywhere, not only on Supabase. It loads before an agent writes or changes anything in a Postgres database: tables and column types, migrations, RLS policies and their tests, indexes, triggers, functions, pg_cron and pgmq jobs, pgvector search or data imports. It also loads for slow queries, timeouts, connection exhaustion, locking, bloat, or rows visible to the wrong tenant.","The Postgres Best Practices rules sit in eight categories ranked by impact. Query performance, connection management, and security with row level security are rated critical, and schema design high. Locking, data access patterns, monitoring and advanced features follow. Examples include indexing WHERE, JOIN and foreign key columns, pooling connections, using prepared statements safely with a pooler, cursor pagination instead of OFFSET, and SKIP LOCKED for worker queues.","The Postgres Best Practices security rules cover row level security for multi-tenant data, RLS policies that stay fast, and least privilege for roles. Schema rules cover adding constraints safely in migrations, data types, primary keys and lowercase identifiers. Each rule file explains why it matters and shows incorrect and correct SQL, with EXPLAIN output or metrics where useful and Supabase notes where they apply."],"useItWhen":["You have a query that was fast on test data and is slow in production.","You are adding row level security to a multi-tenant app and want policies that stay fast.","You see your app run out of database connections under load or from serverless functions.","You are writing a migration and want constraints, types and indexes right the first time."],"notFor":["Your task is Supabase-specific, such as Auth, Edge Functions or Storage: use the Supabase skill."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add supabase/agent-skills --skill supabase-postgres-best-practices"},{"label":"Or, for Claude Code, add Supabase's plugin marketplace from your terminal","command":"claude plugin marketplace add supabase/agent-skills"},{"label":"Then install the postgres-best-practices plugin","command":"claude plugin install postgres-best-practices@supabase-agent-skills"}],"faq":[{"q":"Does the Postgres best practices skill only work with Supabase?","a":"No. Supabase maintains it for Postgres running anywhere; Supabase notes appear only in the rules where they apply. It is MIT, per its SKILL.md and the repository's LICENSE file."},{"q":"Which Postgres problems does the skill rank as critical?","a":"Query performance, connection management, and security with row level security. Schema design is rated high, followed by locking, data access patterns, monitoring and advanced features."}],"sourceUpdated":"2026-07-30","version":"1.1.1","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:45:26+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/pptx","owner":"anthropics","slug":"pptx","name":"PPTX","skillName":"pptx","title":"pptx skill: build and edit PowerPoint decks (Claude only)","summary":"PPTX is Anthropic's proprietary agent skill for building PowerPoint decks from scratch or from your template and checking every slide for overflow; its license limits it to Claude and Anthropic's products.","category":"documents-and-office","source":"anthropics/skills","path":"skills/pptx","license":"Proprietary: use only within Anthropic's services, no copying or redistribution","licenseNote":"The LICENSE.txt in the pptx folder (read 2026-10-01) puts use under your agreement with Anthropic and forbids extracting the files from Anthropic's services, keeping copies outside them, copying them, making derivative works and redistributing them. Anthropic's README calls its four document skills source-available, not open source. You may read the files on GitHub, but you may not copy them into Codex, Cursor or any other agent, or publish a changed version.","worksWith":["Claude Code","Claude.ai","Claude API"],"whatItDoes":["The PPTX skill teaches Claude to work with PowerPoint files (.pptx) and templates (.potx). It loads whenever a deck, slides or a presentation file is involved: creating a pitch deck, pulling the text out of a deck, editing slides, combining or splitting files, or working with layouts, speaker notes and comments.","PPTX builds new decks with the pptxgenjs library, following a list of pitfalls that corrupt files or break layouts. Existing decks and templates are unzipped and their slide XML edited, with bundled scripts that make a thumbnail grid of every slide, duplicate slides, remove unused parts and validate the file. Design guidance asks for a palette chosen for the topic, a visual on every slide, fonts that render reliably, and no accent lines under titles. Claude renders every slide as an image and fixes overflow and overlaps.","PPTX is one of Anthropic's four document skills, next to Word, PDF and Excel. Unlike most skills in Anthropic's repository it is not open source: the license allows use only within Anthropic's own services and forbids copying, changing or redistributing it. That is why this page lists only Anthropic's install routes."],"useItWhen":["You need a pitch deck or report presentation as a real PowerPoint file.","You have a company template and want new slides built from its layouts.","You want the text out of an existing deck for a summary or an email.","You need to combine slides from several decks or split one into parts."],"notFor":["You work in Codex, Cursor, Gemini CLI or another agent that is not Anthropic's: the pptx license bars copying the skill there.","You only want colors and fonts applied to a deck: Theme Factory does that and is Apache-2.0."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the document-skills plugin, which holds the Word, PDF, PowerPoint and Excel skills","command":"/plugin install document-skills@anthropic-agent-skills"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","faq":[{"q":"Can I use Anthropic's pptx skill in Codex, Cursor or ChatGPT?","a":"No. Its LICENSE.txt (read 2026-10-01) forbids keeping copies outside Anthropic's services, copying and redistribution, so it is meant for Claude.ai, Claude Code through Anthropic's plugin, and the Claude API. The open skills CLI does list it, but installing it into another agent copies the files, which the license forbids."},{"q":"Can the pptx skill use my company's PowerPoint template?","a":"Yes. It unzips your .pptx or .potx template, builds new slides from its layouts by editing the slide XML with bundled scripts, then renders every slide as an image to find and fix overflow and overlaps."}],"sourceUpdated":"2026-07-17","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/prisma/prisma-upgrade-v7","owner":"prisma","slug":"prisma-upgrade-v7","name":"Prisma upgrade to v7","skillName":"prisma-upgrade-v7","title":"prisma-upgrade-v7 skill: migrate Prisma ORM 6 to 7 with an agent","summary":"prisma-upgrade-v7 is Prisma's agent skill for upgrading a project from Prisma ORM 6 to 7, covering every breaking change: the new generator, driver adapters, prisma.config.ts, ESM and removed features.","category":"databases","source":"prisma/skills","path":"prisma-upgrade-v7","license":"MIT","worksWith":["Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["prisma-upgrade-v7 is an agent skill from Prisma that takes an existing project from Prisma ORM 6 to Prisma ORM 7. It ranks the breaking changes by impact: schema and generator changes, driver adapters and the module system are critical, while config and environment loading, removed features and Accelerate are high. Each area has its own reference file that the agent reads when it reaches that step.","The prisma-upgrade-v7 skill lays the upgrade out in ten steps. The agent updates the packages, chooses ESM or keeps CommonJS with moduleFormat set to cjs, updates tsconfig, switches the generator to prisma-client with an explicit output path, creates prisma.config.ts, installs a driver adapter such as @prisma/adapter-pg, and loads environment variables itself with dotenv. It then fixes imports and client setup and runs prisma generate.","The skill also replaces Prisma.validator with TypeScript's satisfies, moves $use middleware to Client Extensions, and notes that metrics were removed. It stops MongoDB projects early: Prisma 7 has no MongoDB connector, so those projects should stay on version 6 or move to Prisma 8 with the separate prisma-mongodb-upgrade skill."],"useItWhen":["You want to upgrade a project from Prisma 6 to Prisma 7.","After upgrading you get an error saying a driver adapter is required.","Your imports from @prisma/client broke after switching to the prisma-client generator.","You use Prisma middleware or Accelerate and need to know what changes in 7."],"notFor":["MongoDB projects: use prisma-mongodb-upgrade from the same repository.","New projects: Prisma's README says new setups now default to Prisma ORM 8, set up with its prisma-orm-setup skill."],"requirements":["Node.js 20.19.0 or newer and TypeScript 5.4.0 or newer","A SQL database such as PostgreSQL, MySQL or SQLite"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add prisma/skills --skill prisma-upgrade-v7"}],"sourceUpdated":"2026-09-29","version":"7.6.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:45:26+02:00"},{"url":"https://cookiesforai.app/skills/coreyhaines31/programmatic-seo","owner":"coreyhaines31","slug":"programmatic-seo","name":"Programmatic SEO","skillName":"programmatic-seo","title":"programmatic-seo skill: build SEO pages at scale from templates","summary":"programmatic-seo is Corey Haines's agent skill for building SEO pages at scale from templates and data, such as location, comparison or integration pages, without tripping search engine spam policies.","category":"marketing-and-seo","source":"coreyhaines31/marketingskills","path":"skills/programmatic-seo","license":"MIT","worksWith":["Claude Code","Codex","Cursor","Windsurf","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["Programmatic SEO is one of Corey Haines's marketing skills, for sites that want many similar pages aimed at different keywords: city pages, X versus Y comparisons, integration pages, glossaries or directories. It reads your product marketing context first, then asks what keyword pattern, data and goals you have before proposing anything.","The skill groups page types into 12 playbooks, including templates, curation, conversions, comparisons, examples, locations, personas, integrations, glossary, translations, directory and profiles, each with an example search pattern and a reference file with more detail. It helps you choose a playbook that matches the data you actually own, since proprietary data is one of its core principles.","Implementation covers keyword pattern research, data requirements, template design, internal linking and indexation. Its rules push for unique value on every page, clean URLs and real search intent, and warn against thin pages that search engines treat as spam. The output is a strategy document and a page template, plus checklists for before and after launch."],"useItWhen":["You want to create a page for every city, integration or competitor you serve.","You have a dataset and want to know which page pattern it can support.","You want a page template and internal linking plan before generating hundreds of pages.","You worry your template pages look thin to Google."],"notFor":["Fixing an existing site's SEO problems: use seo-audit from the same set."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add coreyhaines31/marketingskills --skill programmatic-seo"},{"label":"Or, in Claude Code, add the publisher's plugin marketplace","command":"/plugin marketplace add coreyhaines31/marketingskills"},{"label":"Then install all the marketing skills","command":"/plugin install marketing-skills"}],"sourceUpdated":"2026-05-14","version":"2.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/vercel-react-best-practices","owner":"vercel-labs","slug":"vercel-react-best-practices","name":"React Best Practices","skillName":"vercel-react-best-practices","title":"vercel-react-best-practices skill: faster React and Next.js","summary":"React Best Practices is Vercel Engineering's agent skill of 70 performance rules for React and Next.js, ranked by impact, which the agent applies when it writes or reviews code.","category":"coding-and-frameworks","source":"vercel-labs/agent-skills","path":"skills/react-best-practices","license":"MIT","licenseNote":"The SKILL.md frontmatter says license: MIT. The vercel-labs/agent-skills repository has no LICENSE file and GitHub reports no license; its README ends with \"License: MIT\" (checked 2026-10-01). MIT lets you use, change and share the skill if you keep the copyright notice.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["React Best Practices is a performance guide for React and Next.js maintained by Vercel Engineering. It loads when an agent writes, reviews or refactors React components or Next.js pages, implements data fetching, or works on bundle size and load times. It holds 70 rules in eight categories, ordered by impact so the biggest problems come first.","The two critical categories in React Best Practices are request waterfalls and bundle size: start independent work in parallel with Promise.all(), move await into the branch that needs it, stream content with Suspense, import directly instead of through barrel files, and load heavy components with next/dynamic. The server rules include authenticating server actions like API routes, deduplicating work per request with React.cache(), and passing less data to client components.","Lower React Best Practices categories cover client data fetching with SWR, re-render optimization, rendering performance, JavaScript micro-optimizations and a few advanced patterns, such as deriving state during render instead of in effects, using a ternary instead of && for conditional rendering, and Set or Map for repeated lookups. Each rule has its own file in the rules folder with an explanation, an incorrect code example and a correct one, and AGENTS.md compiles every rule into one document."],"useItWhen":["You have slow Next.js pages and suspect data fetching waterfalls.","You have a JavaScript bundle that keeps growing and want to know what to split or defer.","You want an agent to review a React pull request for performance problems.","You are writing new React components and want them fast from the start."],"notFor":["Your problem is component API design and boolean props: use React Composition Patterns.","You build a React Native or Expo app: use the React Native skill."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill vercel-react-best-practices"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-04-14","version":"1.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/vercel-composition-patterns","owner":"vercel-labs","slug":"vercel-composition-patterns","name":"React Composition Patterns","skillName":"vercel-composition-patterns","title":"vercel-composition-patterns skill: replace React boolean props","summary":"React Composition Patterns is Vercel's agent skill for React component APIs: compound components, lifted state and explicit variants instead of a growing list of boolean props.","category":"coding-and-frameworks","source":"vercel-labs/agent-skills","path":"skills/composition-patterns","license":"MIT","licenseNote":"The SKILL.md frontmatter says license: MIT. The vercel-labs/agent-skills repository has no LICENSE file and GitHub reports no license; its README ends with \"License: MIT\" (checked 2026-10-01). MIT lets you use, change and share the skill if you keep the copyright notice.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["React Composition Patterns is Vercel's guide to building React components that stay flexible as a codebase grows. It loads when an agent refactors components that have piled up boolean props, builds a reusable component library, designs a component API, or works with compound components, render props or context providers. The guide says these patterns make a codebase easier for both people and AI agents to work with.","React Composition Patterns holds eight rules in four groups, ordered by impact. Component architecture comes first: replace boolean props with composition, and structure complex components as compound components that share context. The state rules keep the provider as the only place that knows how state is managed, define a generic context interface with state, actions and meta, and lift state into providers so sibling components can reach it.","Two implementation rules in React Composition Patterns prefer explicit variant components over boolean modes, and children over renderX props. A React 19 rule, skipped on React 18 or earlier, treats ref as a regular prop instead of wrapping components in forwardRef, and reads context with use() instead of useContext(). Each rule file shows incorrect and correct code, and AGENTS.md compiles every rule into one document."],"useItWhen":["You have a component with so many boolean props that every new case adds another flag.","You are building a component library and want APIs other developers can compose.","You are moving a codebase to React 19 and want forwardRef and useContext replaced.","You want an agent to review your component architecture before it grows further."],"notFor":["Your problem is speed, such as request waterfalls or bundle size: use React Best Practices."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill vercel-composition-patterns"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-01-28","version":"1.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/vercel-react-native-skills","owner":"vercel-labs","slug":"vercel-react-native-skills","name":"React Native","skillName":"vercel-react-native-skills","title":"vercel-react-native-skills: faster React Native apps","summary":"React Native is Vercel's agent skill of React Native and Expo rules, ranked by impact: list performance, animations, native navigation and UI patterns for faster mobile apps.","category":"coding-and-frameworks","source":"vercel-labs/agent-skills","path":"skills/react-native-skills","license":"MIT","licenseNote":"The SKILL.md frontmatter says license: MIT. The vercel-labs/agent-skills repository has no LICENSE file and GitHub reports no license; its README ends with \"License: MIT\" (checked 2026-10-01). MIT lets you use, change and share the skill if you keep the copyright notice.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["The React Native skill is Vercel's set of best practices for React Native and Expo apps. It loads when an agent builds React Native components, tunes list and scroll performance, adds animations, or works with native modules and platform APIs. The rules are grouped into eight categories and ordered by impact, from list performance down to configuration.","In the React Native skill, list performance comes first: virtualize large lists with FlashList, memoize list items, keep callbacks stable, avoid inline style objects and use item types for mixed lists. The animation rules animate only transform and opacity, use useDerivedValue for computed animations and Gesture.Tap instead of Pressable for animated presses. Navigation uses native stack and native tabs over JavaScript navigators.","The React Native skill's UI rules prefer expo-image for images, Pressable over TouchableOpacity, native context menus and modals, and onLayout over measure(). Further rules cover state subscriptions and the React Compiler, wrapping text in Text components, and keeping native dependencies in the app package of a monorepo. Each rule file shows incorrect and correct code, and AGENTS.md compiles every rule into one document."],"useItWhen":["You have a long list in your React Native app that stutters when you scroll.","You are adding animations with Reanimated and want them to run smoothly.","You are choosing between native and JavaScript navigators, menus or modals.","You keep your React Native app in a monorepo and native modules are not autolinking."],"notFor":["You build for the web with React or Next.js: use React Best Practices."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill vercel-react-native-skills"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-01-27","version":"1.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/vercel-react-view-transitions","owner":"vercel-labs","slug":"vercel-react-view-transitions","name":"React View Transitions","skillName":"vercel-react-view-transitions","title":"vercel-react-view-transitions skill: animate route changes","summary":"React View Transitions is Vercel's agent skill for animating route changes, shared elements and list reorders with React's ViewTransition component, in Next.js and other React apps.","category":"coding-and-frameworks","source":"vercel-labs/agent-skills","path":"skills/react-view-transitions","license":"MIT","licenseNote":"The SKILL.md frontmatter says license: MIT. The vercel-labs/agent-skills repository has no LICENSE file and GitHub reports no license; its README ends with \"License: MIT\" (checked 2026-10-01). MIT lets you use, change and share the skill if you keep the copyright notice.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["React View Transitions is Vercel's guide to animating between UI states with React's View Transition API, which uses the browser's own document.startViewTransition. It loads when you want page transitions, animated route changes, shared element morphs, enter and exit animations, list reorders or forward and back navigation animations, including in Next.js, without a third-party animation library.","React View Transitions sets an order: shared elements first, then Suspense reveals, list identity, state changes and route changes, and each animation has to communicate a spatial relationship. Directional slides are kept for list to detail navigation and ordered sequences, not tabs. It explains the common traps: enter and exit only fire when ViewTransition comes before any DOM node, only startTransition, useDeferredValue and Suspense trigger it, and router.back() carries no transition types.","React View Transitions also covers availability: the Next.js App Router already bundles the React canary build, other apps install react@canary, and unsupported browsers skip the animations. Five reference files hold a step-by-step implementation workflow that starts with an audit, CSS recipes including reduced motion styles, patterns, troubleshooting, and Next.js details such as transitionTypes on next/link."],"useItWhen":["You want a thumbnail to morph into the full image when someone opens a detail page.","You want route changes in Next.js to slide forward and back instead of cutting.","You added view transitions that do nothing and need to find out why.","You want list items to animate when they reorder, appear or disappear."],"notFor":["Your app cannot use React canary: outside the Next.js App Router, ViewTransition needs react@canary."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill vercel-react-view-transitions"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-08-28","version":"1.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/receiving-code-review","owner":"obra","slug":"receiving-code-review","name":"Receiving Code Review","skillName":"receiving-code-review","title":"receiving-code-review skill: check review feedback before acting","summary":"Receiving Code Review is an agent skill from Superpowers: your agent checks review feedback against the code before acting, pushes back with reasons when it is wrong, and skips the flattery.","category":"code-review-and-git","source":"obra/superpowers","path":"skills/receiving-code-review","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Receiving Code Review is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when the agent receives code review feedback, before it implements any suggestion, and matters most when the feedback is unclear or looks technically wrong. The agent reads all of it, restates each point or asks about it, checks it against the actual codebase, and only then responds and implements, one item at a time with a test for each.","Receiving Code Review bans performative replies such as \"You're absolutely right\" or \"Great point\", and any thanks: the agent states the fix or simply makes it. If some items are unclear, it asks about them before implementing any, since items may be related. Feedback from you is trusted once understood. Suggestions from outside reviewers are checked first: whether they break something, whether the current code has a reason, and whether the reviewer has the full context.","Under Receiving Code Review, when a reviewer asks for something to be implemented \"properly\", the agent searches the codebase for real usage and suggests removing the code if nothing calls it (YAGNI). It pushes back with technical reasoning when a suggestion is wrong for this stack, and stops to discuss with you when feedback conflicts with your earlier decisions. Fixes go in a set order: blocking issues, then simple fixes, then complex ones. On GitHub it replies inside the review comment thread."],"useItWhen":["Your agent agrees with every review comment and changes code without checking it.","You are passing an outside reviewer's comments on a pull request to your agent.","You have review items that are unclear and want the agent to ask before it implements the rest."],"notFor":["You want a review of your own changes: use Requesting Code Review."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill receiving-code-review"}],"sourceUpdated":"2026-07-24","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/hashicorp/refactor-module","owner":"hashicorp","slug":"refactor-module","name":"Refactor module","skillName":"refactor-module","title":"refactor-module skill: split Terraform code into modules","summary":"refactor-module is HashiCorp's agent skill for turning a monolithic Terraform configuration into reusable modules, with typed interfaces, documentation, tests and a state migration that recreates no resources.","category":"infrastructure-as-code","source":"hashicorp/agent-skills","path":"plugins/terraform/skills/refactor-module","license":"MPL-2.0","licenseNote":"MPL-2.0 is a file-level copyleft license: you can use and copy the skill in any project, including closed source ones, but if you distribute changed versions of its files, those files stay under MPL-2.0 and must be shared as source.","worksWith":["GitHub Copilot","Claude Code","Codex","Cursor","OpenCode","IBM Bob"],"cliChecked":"2026-10-01","whatItDoes":["refactor-module is an agent skill from HashiCorp that walks an agent through breaking a large Terraform configuration into modules. It starts with analysis: grouping resources by function, finding repeated patterns, mapping dependencies and judging how hard the state migration will be. It then designs each module's interface as typed variables with validation and a set of outputs, and decides what belongs inside a module and what stays outside.","State migration is the risky step, and refactor-module handles it explicitly. The agent lists current addresses with terraform state list or terraform show -json rather than parsing the raw state file, and never echoes state contents, since they hold secrets in plain text. It writes moved blocks for Terraform 1.1 and newer, or terraform state mv commands for older versions, then checks that a plan shows no changes before anything is applied.","The skill also covers module documentation, testing through the terraform-test skill, semantic versioning for modules, and three refactoring patterns: grouping, layering and composition. It warns against over-generic modules and against modules that reference each other directly. Its success criteria include one responsibility per module and no resource recreation after the move."],"useItWhen":["Your Terraform lives in one large main.tf with copied blocks for each subnet or environment.","You want to turn existing resources into a module without destroying and recreating them.","You need moved blocks or terraform state mv commands written for a refactor.","You want a reusable internal module with documented inputs and outputs."],"notFor":["Reviewing the output of terraform plan before an apply: HashiCorp's 20 skills (read 2026-10-01) include none made only for that job.","Writing a module from scratch with no existing state: terraform-style-guide is enough."],"requirements":["Terraform CLI and read access to the current state (terraform state list); moved blocks need Terraform 1.1 or newer"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add hashicorp/agent-skills --skill refactor-module"},{"label":"Or, for Claude Code, add HashiCorp's plugin marketplace","command":"claude plugin marketplace add hashicorp/agent-skills"},{"label":"Then install the terraform plugin, which holds all 16 Terraform skills","command":"claude plugin install terraform@hashicorp"}],"installNote":"In Codex, HashiCorp's README says to add the repository's .agents/plugins/marketplace.json as a repository marketplace, then install the terraform plugin.","faq":[{"q":"Does HashiCorp have an agent skill that reviews terraform plan output?","a":"No. On 2026-10-01 hashicorp/agent-skills held 20 skills, 16 for Terraform and 4 for Packer, and none is dedicated to reading or reviewing terraform plan output. The closest is terraform-test, which can run tests in plan mode against assertions you write."},{"q":"Can I copy this HashiCorp skill into my own repository?","a":"Yes. The skill is MPL-2.0. You can copy and use it in any project, open or closed source. If you change its files and distribute them, the changed files must stay MPL-2.0 and be available as source; your own files are not affected."}],"sourceUpdated":"2026-08-26","version":"0.0.1","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/remotion-dev/remotion-best-practices","owner":"remotion-dev","slug":"remotion-best-practices","name":"Remotion best practices","skillName":"remotion-best-practices","title":"remotion-best-practices skill: videos in React with an AI agent","summary":"remotion-best-practices is Remotion's own agent skill and the entry point to its other video skills: it routes an agent to the right guide for creating, previewing, captioning and rendering Remotion videos in React.","category":"video-and-media","source":"remotion-dev/skills","path":"skills/remotion-best-practices","license":"No license file in the skills repo","licenseNote":"GitHub reports no license for remotion-dev/skills on 2026-10-01, and the README says it is generated from the packages/skills folder of the Remotion monorepo. Remotion itself uses the Remotion License (github.com/remotion-dev/remotion/blob/main/LICENSE.md, read 2026-10-01): free for individuals, for-profit companies with up to 3 employees and non-profits, while larger for-profit companies need a paid company license.","worksWith":["Claude Code","Codex","Kimi Code","Cursor","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["Remotion best practices is the router skill in Remotion's skills repository. Remotion is a framework for making videos with React code. The skill itself is short: it tells the agent which bundled guide to load for the task, with folders for creating a project, writing Remotion markup, maps, captions, multimedia, interactivity, Studio, rendering, SaaS apps, documentation lookup and upgrades.","Its workflow puts a live preview first. When you ask for a new video, the agent sets up a project if needed, starts Remotion Studio and opens the preview as soon as the project runs, so you can watch progress and steer. It renders an MP4 only when you ask for it in plain words, such as export or give me the MP4, and points out that Studio has its own render interface.","The skill also protects your edits. If the agent finds a change in the code it did not make, it is told to assume the change was intentional or to ask, not to overwrite it. Separate notes cover how to open Studio in Cursor, in agents with an in-app browser and in agents without one."],"useItWhen":["You want an agent to make a promo or explainer video in code and show you a live preview.","You want animated captions or subtitles added to a Remotion video.","You are building an app that renders videos with Remotion and want the right architecture.","You need to upgrade Remotion and its skills to the current version."],"notFor":["Larger for-profit companies that have not bought a Remotion company license: the free license covers only individuals, companies of up to 3 employees and non-profits."],"requirements":["A Remotion project, or permission for the agent to create one"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add remotion-dev/skills --skill remotion-best-practices"},{"label":"Or start a new Remotion project, which offers to add the skills","command":"bun create video"}],"faq":[{"q":"Is Remotion free for companies?","a":"Only small ones. The Remotion License, read on 2026-10-01, lets individuals, for-profit organizations with up to 3 employees and non-profits use Remotion for free, including commercially. Other for-profit companies need a company license from remotion.pro."},{"q":"What license do the Remotion agent skills have?","a":"The remotion-dev/skills repository has no license file, and GitHub reports none on 2026-10-01. The skills are generated from Remotion's main repository, whose LICENSE.md is the Remotion License, so treat them as covered by Remotion's terms unless Remotion says otherwise."}],"sourceUpdated":"2026-10-01","version":"4.0.532","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/requesting-code-review","owner":"obra","slug":"requesting-code-review","name":"Requesting Code Review","skillName":"requesting-code-review","title":"requesting-code-review skill: AI code review before you merge","summary":"Requesting Code Review is an agent skill from Superpowers: your agent sends its changes to a separate reviewer subagent that checks them against the plan and grades each issue by severity.","category":"code-review-and-git","source":"obra/superpowers","path":"skills/requesting-code-review","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have. This skill needs a subagent tool: Claude Code's Agent tool, or in Codex the multi-agent feature switched on in config.toml (multi_agent = true).","cliChecked":"2026-10-01","whatItDoes":["Requesting Code Review is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads after a task, after a major feature, or before a merge, and it is required after every task when you work with Subagent-Driven Development. The agent does not review its own diff. It takes the git range for the work and dispatches a reviewer subagent with context written for the review, never the session's history.","The code-reviewer.md template bundled with Requesting Code Review fills in what was built, the plan or requirements, and the base and head commits. The reviewer works read-only, never spawns reviewers of its own, and checks plan alignment, code quality, architecture, testing and production readiness. It lists the behaviors it set aside as outside the plan, and treats the spec as a vision document: where the spec is silent, what a reasonable user would expect counts as a requirement.","The Requesting Code Review reviewer returns strengths, issues graded Critical, Important or Minor with a file and line for each, recommendations, and a clear ready-to-merge verdict. The agent fixes Critical issues at once, fixes Important ones before moving on, and notes Minor ones for later. When the reviewer is wrong, the agent pushes back with technical reasoning and shows the code or tests that prove it works."],"useItWhen":["Your agent finished a feature and you want a second opinion before you merge.","You want review findings sorted by severity, with a file and line for each.","You want the changes checked against your plan, not only for style.","Your agent is stuck and a fresh look at the code could help."],"notFor":["You are on the receiving end of review comments: use Receiving Code Review."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill requesting-code-review"}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/getsentry/security-review","owner":"getsentry","slug":"security-review","name":"Security review","skillName":"security-review","title":"security-review skill: find exploitable vulnerabilities in code","summary":"security-review is Sentry's agent skill for reviewing code for exploitable vulnerabilities, such as injection, XSS, broken access control and weak crypto, reporting only findings it has confirmed with high confidence.","category":"security","source":"getsentry/skills","path":"skills/security-review","license":"CC-BY-SA-4.0","licenseNote":"The repository is Apache-2.0, but the security-review folder has its own LICENSE: its reference material is adapted from the OWASP Cheat Sheet Series and stays under CC-BY-SA-4.0. You may copy and adapt it, even commercially, if you credit the source and share any changed version under the same license.","worksWith":["Claude Code","Cursor","Cline","GitHub Copilot"],"agentNote":"The SKILL.md lists Claude Code tool names in allowed-tools (Read, Grep, Glob, Bash, Task). Other agents ignore that line and use their own file and shell tools.","cliChecked":"2026-10-01","whatItDoes":["security-review is an agent skill from Sentry that reviews code for security holes an attacker could actually use. Its main rule is to report only high-confidence findings: a vulnerable pattern plus input that an attacker controls. Medium-confidence issues are marked as needing verification, and theoretical or best-practice issues are not reported at all. The agent reports only on the code it was given but researches the whole codebase first.","The security-review skill spends much of its length on avoiding false positives. Values from settings, environment variables, config files and constants count as server-controlled, not attacker-controlled. Auto-escaping in Django, React and Vue templates and parameterized ORM queries are treated as safe unless bypassed, for example through dangerouslySetInnerHTML, v-html, mark_safe or raw SQL built with string interpolation.","The skill loads topic guides by context: authorization and injection for API routes, XSS and CSRF for templates, SSRF for outgoing requests, and others for crypto, file uploads, deserialization and supply chain. It has language guides for Python, JavaScript, Go, Rust and Java, and infrastructure guides for Docker, Kubernetes, Terraform, CI and cloud IAM. Findings carry a severity from critical to low."],"useItWhen":["You want a security review of a pull request without a flood of false alarms.","You are about to ship a new API endpoint and want access control and injection checked.","You want a second look at code that handles file uploads, tokens or outgoing requests.","You need an OWASP-style review for Python, JavaScript, Go, Rust or Java code."],"notFor":["GitHub Actions workflow files: use gha-security-review from the same repository."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add getsentry/skills --skill security-review"},{"label":"Or, for Claude Code, add Sentry's plugin marketplace","command":"claude plugin marketplace add getsentry/skills"},{"label":"Then install the sentry-skills plugin, which holds all 28 skills","command":"claude plugin install sentry-skills@sentry-skills"}],"installNote":"Restart Claude Code after installing the plugin. Sentry's README says its skills were written for Sentry employees, so some carry Sentry conventions.","faq":[{"q":"Can I copy Sentry's security-review skill into my own project?","a":"Yes, with conditions. Its reference material is CC-BY-SA-4.0 because it is adapted from the OWASP Cheat Sheet Series. You may copy and change it, even in a commercial project, if you credit the source, link the license, say what you changed, and release any changed version under CC-BY-SA-4.0."},{"q":"Does security-review report every possible vulnerability?","a":"No, by design. It reports only high-confidence findings where attacker-controlled input reaches a vulnerable pattern, marks medium-confidence ones as needing verification, and drops theoretical issues. That keeps reports short but can miss issues it could not confirm."}],"sourceUpdated":"2026-09-28","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/trailofbits/semgrep-rule-creator","owner":"trailofbits","slug":"semgrep-rule-creator","name":"Semgrep rule creator","skillName":"semgrep-rule-creator","title":"semgrep-rule-creator skill: write and test custom Semgrep rules","summary":"semgrep-rule-creator is Trail of Bits' agent skill for writing custom Semgrep rules that detect vulnerabilities and bug patterns, test-first, with taint mode preferred and every test required to pass.","category":"security","source":"trailofbits/skills","path":"plugins/semgrep-rule-creator/skills/semgrep-rule-creator","license":"CC-BY-SA-4.0","licenseNote":"CC-BY-SA-4.0 is a share-alike license. You can copy the skill into a project and change it, even for commercial work, if you credit Trail of Bits, link the license and say what you changed. If you share a changed version, for example in a public repository, it must stay under CC-BY-SA-4.0. It is a content license, not a software license, so ask your legal team before bundling it inside a closed product.","worksWith":["Claude Code","Codex","ChatGPT workspace marketplace"],"cliChecked":"2026-10-01","whatItDoes":["semgrep-rule-creator is an agent skill from security firm Trail of Bits for writing new Semgrep rules, the YAML pattern files that Semgrep runs to find bugs and vulnerabilities in source code. It covers rules for a specific bug pattern, a class of security vulnerability, data flow from user input to a dangerous function, and coding standards a team wants enforced.","The semgrep-rule-creator workflow is strict and test-first. The agent writes a test file with both vulnerable cases and safe cases before the rule, looks at Semgrep's syntax tree dump of the code, writes the rule, then iterates until every test passes with semgrep --test. Only then may it simplify the patterns. Each rule lives in its own folder with exactly one YAML file and one test file.","The skill prefers taint mode for injection-style bugs, because a plain pattern such as eval matches safe constant input too, while taint mode alerts only when untrusted data reaches the sink. It lists anti-patterns to avoid: rules so broad they match every call, rules so narrow they miss variations, generic-language rules, and tests with no safe cases."],"useItWhen":["You found a bug pattern in your code and want a Semgrep rule that catches every instance.","You want a taint rule that tracks user input to a dangerous function.","You want to enforce a team coding standard in CI with Semgrep.","Your existing custom rule has too many false positives and needs tests."],"notFor":["Scanning a codebase with Semgrep rulesets that already exist: Trail of Bits' static-analysis plugin covers that.","Porting a rule to another language: use semgrep-rule-variant-creator from the same repository."],"requirements":["The Semgrep CLI installed locally"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add trailofbits/skills --skill semgrep-rule-creator"},{"label":"Or, in Claude Code, add Trail of Bits' plugin marketplace","command":"/plugin marketplace add trailofbits/skills"},{"label":"Or, in Codex, add the same marketplace from your terminal","command":"codex plugin marketplace add trailofbits/skills"},{"label":"Then install the semgrep-rule-creator plugin in Codex","command":"codex plugin add semgrep-rule-creator@trailofbits"}],"installNote":"In Claude Code, open /plugin menu after adding the marketplace and install the semgrep-rule-creator plugin. For a ChatGPT workspace, Trail of Bits' README says to import the repository's .claude-plugin/marketplace.json.","faq":[{"q":"Can I copy Trail of Bits' semgrep-rule-creator skill into my own repository?","a":"Yes, with conditions. The skills are CC-BY-SA-4.0. You may copy and adapt them, even commercially, if you credit Trail of Bits, link the license and note your changes. Any changed version you share must carry the same CC-BY-SA-4.0 license. Using the skill unchanged in your own work does not put your code under that license."},{"q":"Does semgrep-rule-creator run Semgrep scans on my whole codebase?","a":"No. It writes and tests new rules, one rule and one test file at a time. For scanning a codebase with existing rulesets, Trail of Bits points to its static-analysis plugin, which has separate semgrep, codeql and sarif-parsing skills."}],"sourceUpdated":"2026-09-16","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/coreyhaines31/seo-audit","owner":"coreyhaines31","slug":"seo-audit","name":"SEO audit","skillName":"seo-audit","title":"seo-audit skill: find technical and on-page SEO issues","summary":"seo-audit is Corey Haines's agent skill for diagnosing why a site is not ranking: the agent checks crawling, indexing, speed, on-page elements and content quality, then writes a prioritized fix list.","category":"marketing-and-seo","source":"coreyhaines31/marketingskills","path":"skills/seo-audit","license":"MIT","worksWith":["Claude Code","Codex","Cursor","Windsurf","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["SEO audit is one of Corey Haines's marketing skills and loads when you say your site is not ranking, traffic dropped, or you want a technical or on-page SEO check. Like the other skills in the set, it first reads a product marketing context file if one exists, so it does not ask again about your product and audience. It then asks about site type, goals, known issues and whether Search Console data is available.","The audit runs in a fixed order: crawlability and indexation first, then site speed and Core Web Vitals, mobile, HTTPS and URL structure, then titles, meta descriptions, headings, images, internal links and keyword targeting, and finally content quality and E-E-A-T signals. Sections cover international sites with hreflang, and common issues for SaaS, e-commerce, blogs and local businesses.","The skill warns about a trap that produces false findings: fetching a page with curl or a web fetch tool often cannot see schema markup injected by JavaScript, so the agent should render the page in a browser or use Google's Rich Results Test before reporting missing schema. The report lists each issue with impact, evidence, fix and priority, followed by an action plan."],"useItWhen":["Your organic traffic dropped and you want to know why.","You want a technical SEO check of a site before or after a migration.","You want titles, meta descriptions and headings reviewed page by page.","You want findings ranked so you fix what blocks indexing first."],"notFor":["Showing up in AI answers from ChatGPT, Perplexity or Google AI Overviews: the same set's ai-seo skill covers that.","Building many pages from data: use programmatic-seo from the same set."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add coreyhaines31/marketingskills --skill seo-audit"},{"label":"Or, in Claude Code, add the publisher's plugin marketplace","command":"/plugin marketplace add coreyhaines31/marketingskills"},{"label":"Then install all the marketing skills","command":"/plugin install marketing-skills"}],"sourceUpdated":"2026-08-19","version":"2.0.1","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/shadcn-ui/shadcn","owner":"shadcn-ui","slug":"shadcn","name":"shadcn","skillName":"shadcn","title":"shadcn skill: add, compose and fix shadcn/ui components","summary":"shadcn is the shadcn/ui project's own agent skill: it reads your components.json, then guides the agent to find, install, compose and style shadcn/ui components with the shadcn CLI and the project's rules.","category":"frontend-and-design","source":"shadcn-ui/ui","path":"skills/shadcn","license":"MIT","worksWith":["Claude Code","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"The SKILL.md injects live project context by running npx shadcn@latest info --json through a command-substitution line that Claude Code executes when the skill loads. In agents without that feature the agent has to run the command itself, which the skill also tells it to do when context is stale.","cliChecked":"2026-10-01","whatItDoes":["shadcn is the agent skill published in the shadcn/ui repository and recommended in its documentation. It loads in any project with a components.json file and gives the agent your setup: framework, Tailwind version, path aliases, base library (Radix, Base UI or React Aria), icon library and the components already installed. The agent then uses the shadcn CLI to search registries, read component docs and add components instead of writing custom markup.","The skill enforces rules with incorrect and correct code pairs in separate files. Forms use FieldGroup and Field, option sets use ToggleGroup, dialogs always get a title, colors come from semantic tokens instead of raw Tailwind values, spacing uses gap instead of space-y, and icons inside buttons take a data-icon attribute instead of size classes. Chat interfaces are built from the library's message and bubble primitives.","It also covers maintenance. To update a component you have edited, the agent previews upstream changes with dry-run and diff and merges them, and never overwrites without your approval. Preset codes are applied through the CLI, and when you ask for a block without naming a registry, the agent asks which one instead of guessing."],"useItWhen":["You want your agent to build a settings page or dashboard from shadcn/ui parts instead of custom divs.","Your agent writes shadcn/ui code that ignores your aliases, icon library or Radix versus Base UI setup.","You want to pull a block from a community registry and have its imports fixed.","You need to update edited components from upstream without losing your changes."],"notFor":["Projects that do not use shadcn/ui: the skill only activates around a components.json file."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add shadcn-ui/ui --skill shadcn"},{"label":"Or use the command from the shadcn/ui docs, which installs the repository's skills","command":"npx skills add shadcn/ui"}],"sourceUpdated":"2026-09-04","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/skill-creator","owner":"anthropics","slug":"skill-creator","name":"Skill Creator","skillName":"skill-creator","title":"skill-creator skill: create, test and improve agent skills","summary":"Skill Creator is Anthropic's open-source agent skill for making skills: the agent drafts a SKILL.md with you, tests it with and without the skill, and tunes its description so it triggers.","category":"building-skills-and-mcp","source":"anthropics/skills","path":"skills/skill-creator","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill folder (read 2026-10-01); the SKILL.md itself has no license key and the anthropics/skills repository has no root license file. In practice you may use, change and share it in any agent if you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"agentNote":"Parallel test runs, baselines and blind comparison need subagents, which Claude Code and Cowork have; in Claude.ai tests run one at a time and benchmarking is skipped. Description tuning calls claude -p, so it needs the Claude Code command-line tool.","cliChecked":"2026-10-01","whatItDoes":["Skill Creator is Anthropic's skill for writing new skills and improving existing ones. It loads when you want to create a skill from scratch, edit one, test it with evals, benchmark it or improve how reliably its description triggers. The agent works out where you are in the process: it can capture a workflow from the current conversation, interview you about edge cases and formats, then write the SKILL.md.","For testing with Skill Creator, the agent writes two or three realistic prompts and runs each one with the skill and without it (or against the old version), using sub-agents in parallel. It grades the outputs against checks, aggregates pass rate, time and tokens into a benchmark, and opens a bundled review page where you read each output and leave feedback. It then revises the skill from your feedback and runs the tests again.","A separate Skill Creator loop tunes the description: 20 queries that should and should not trigger the skill, split into training and held-out sets, each run three times, for up to five rounds of rewrites. That loop needs the claude command-line tool, so it runs in Claude Code and is skipped in Claude.ai, where tests also run one at a time. A script packages the result as a .skill file."],"useItWhen":["You keep giving your agent the same instructions and want to turn them into a skill.","You wrote a skill and want to know whether it actually improves the results.","Your skill does not load when it should, and you want its description tuned.","You want to package a skill as a file you can install or share."],"notFor":["You want a strict test-first method that closes loopholes in discipline skills: Writing Skills from Superpowers."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill skill-creator"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","faq":[{"q":"Does skill-creator work outside Claude Code?","a":"Partly. Drafting a SKILL.md and packaging it work in any agent with Python and a filesystem. Parallel test runs need subagents, and the description tuning loop needs the claude command-line tool (claude -p), which only Claude Code provides."},{"q":"How is skill-creator different from Superpowers' writing-skills?","a":"Skill Creator measures a skill: runs with and without it, graded benchmarks, a review page and description tuning. Writing Skills applies test-driven development: pressure scenarios with subagents before and after the skill exists, aimed at closing loopholes."}],"sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/slack-gif-creator","owner":"anthropics","slug":"slack-gif-creator","name":"Slack GIF Creator","skillName":"slack-gif-creator","title":"slack-gif-creator skill: animated emoji and GIFs for Slack","summary":"Slack GIF Creator is Anthropic's open-source agent skill for animated GIFs sized for Slack: the agent draws the frames in Python, adds motion and checks the file meets Slack's limits.","category":"video-and-media","source":"anthropics/skills","path":"skills/slack-gif-creator","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"cliChecked":"2026-10-01","whatItDoes":["Slack GIF Creator is an Anthropic skill for making animated GIFs to use in Slack, as custom emoji or in messages. It loads on requests such as \"make me a GIF of X doing Y for Slack\". It sets the targets: 128 by 128 pixels for emoji and 480 by 480 for message GIFs, 10 to 30 frames per second, 48 to 128 colors, and under three seconds for emoji.","With Slack GIF Creator, the agent draws each frame with Python's Pillow library, using shapes such as circles, polygons and lines rather than emoji fonts or stock graphics. It can also animate an image you upload or use it as a reference. The skill describes animation patterns (shake, pulse, bounce, spin, fade, slide, zoom and particle bursts) and asks for thicker lines, gradients and layered shapes so the result does not look like placeholder art.","Slack GIF Creator's bundled Python modules assemble frames into an optimized GIF, provide easing curves for smoother motion, offer helpers for backgrounds, circles, stars and text, and validate whether a file is ready for Slack. They need pillow, imageio and numpy. File size is reduced only when you ask, through fewer frames, fewer colors, smaller dimensions or dropped duplicate frames."],"useItWhen":["You want a custom animated emoji for your team's Slack.","You want a short reaction GIF of an idea, such as a shape that bounces or a star that bursts.","You have an image and want it animated for Slack.","Your GIF is too large for Slack and you need a smaller file."],"notFor":["You want a still poster or image: use Canvas Design."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill slack-gif-creator"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/stripe/stripe-best-practices","owner":"stripe","slug":"stripe-best-practices","name":"Stripe best practices","skillName":"stripe-best-practices","title":"stripe-best-practices skill: build Stripe payments correctly","summary":"stripe-best-practices is Stripe's own agent skill for integration decisions: it steers the agent to the right Stripe API for payments, subscriptions, Connect, tax and security, using current API and SDK versions.","category":"coding-and-frameworks","source":"stripe/ai","path":"skills/stripe-best-practices","license":"MIT","worksWith":["Claude Code","Codex","Cursor","Grok Build","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"Most details live in reference pages on docs.stripe.com that the agent is told to read before answering, so it needs web access to follow the skill fully.","cliChecked":"2026-10-01","whatItDoes":["Stripe best practices is the main skill in Stripe's AI repository and loads whenever an agent plans, builds, changes, tests or reviews a Stripe integration. On 2026-10-01 it names Stripe API version 2026-08-26.dahlia as the latest and lists current SDK versions for Ruby, Python, Node.js, PHP, Java, Go and .NET, telling the agent to use them unless you say otherwise.","A routing table maps each job to an API: Checkout Sessions for one-time payments, Checkout Sessions with the Payment Element for embedded forms, Setup Intents for saving cards, Accounts v2 for Connect platforms, Billing APIs for subscriptions, Metronome for new usage-based billing, Stripe Tax with the Registrations API for sales tax and VAT, and v2 Financial Accounts for embedded banking. The agent reads the matching reference page before writing code.","Its critical rules target common mistakes. The agent must not pass payment_method_types except for Terminal, must treat webhooks as required and fulfil orders in a webhook handler instead of the success page, must confirm an active tax registration before turning on automatic tax, and must use a StripeClient instance instead of the deprecated global API key. It recommends restricted API keys over secret keys and separate sandboxes for development and CI."],"useItWhen":["You are adding Stripe Checkout or subscriptions to an app and want the current recommended API.","You are building a marketplace on Stripe Connect.","You want to collect sales tax or VAT with Stripe Tax without silently collecting nothing.","You want an agent to review an existing Stripe integration for deprecated patterns."],"notFor":["Other payment providers: every rule here is specific to Stripe's APIs."],"requirements":["A Stripe account, or a test sandbox created with the Stripe CLI (stripe sandbox create)"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add stripe/ai --skill stripe-best-practices"},{"label":"Or, in Claude Code, install Stripe's plugin from the official marketplace","command":"claude plugin install stripe@claude-plugins-official"},{"label":"Or, in Codex, install Stripe's plugin from the OpenAI-curated marketplace","command":"codex plugin add stripe@openai-curated"}],"installNote":"Stripe recommends its plugins where available because they update automatically. Copies installed with the skills CLI need npx skills update to stay current.","sourceUpdated":"2026-10-02","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:02:16+02:00"},{"url":"https://cookiesforai.app/skills/obra/subagent-driven-development","owner":"obra","slug":"subagent-driven-development","name":"Subagent-Driven Development","skillName":"subagent-driven-development","title":"subagent-driven-development skill: a fresh subagent per task","summary":"Subagent-Driven Development is an agent skill from Superpowers: your agent carries out a plan by giving each task to a fresh subagent, reviewing every task, then reviewing the whole branch.","category":"planning-and-workflow","source":"obra/superpowers","path":"skills/subagent-driven-development","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have. This skill needs a subagent tool: Claude Code's Agent tool, or in Codex the multi-agent feature switched on in config.toml (multi_agent = true).","cliChecked":"2026-10-01","whatItDoes":["Subagent-Driven Development is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when the agent executes an implementation plan whose tasks are mostly independent. The main agent acts as a controller: a fresh implementer subagent takes each task, a task reviewer checks spec compliance and code quality after it, and a reviewer on the most capable available model checks the whole branch at the end.","Subagent-Driven Development's bundled scripts set up a git-ignored workspace per plan, cut each task's brief out of the plan, and package the diff for the reviewer as a file. Progress goes into a ledger that survives context compaction, so a controller that loses its place does not redo finished tasks. Review findings go through fix rounds, at most five per task: the first three resume the same implementer, the last two use a fresh one on a more capable model.","In Subagent-Driven Development, each role gets the least capable model that can handle it, and every dispatch names its model. The agent runs all tasks without checking in, records its own rulings on plan conflicts, and stops only for destructive or security-sensitive actions, a merge, push or publish, or a plan too broken to follow. It works in a worktree from Using Git Worktrees, takes its final reviewer from Requesting Code Review, and ends with Finishing a Development Branch."],"useItWhen":["You have a plan with many independent tasks and want each one reviewed before the next starts.","Your agent loses track of long plans after its context is compacted.","You want to keep cost down by running simple tasks on cheaper models.","You want the agent to work through the whole plan without asking whether to continue."],"notFor":["Your agent has no subagent tool, or you want one context to do every task: use Executing Plans."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill subagent-driven-development"}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/supabase/supabase","owner":"supabase","slug":"supabase","name":"Supabase","skillName":"supabase","title":"supabase skill: auth, RLS, Edge Functions and Postgres","summary":"Supabase is Supabase's own agent skill for any Supabase task, with a security checklist for auth and row level security and a rule to check the current docs first.","category":"databases","source":"supabase/agent-skills","path":"skills/supabase","license":"MIT","licenseNote":"The SKILL.md has no license key; MIT comes from the LICENSE file at the root of supabase/agent-skills (checked 2026-10-01).","worksWith":["Claude Code","GitHub Copilot","Cursor","Cline","other agents through the skills CLI (Supabase says 18 or more)"],"cliChecked":"2026-10-01","whatItDoes":["The Supabase skill is Supabase's official skill for AI coding agents. It loads for any task that touches Supabase: the database, Auth, Edge Functions, Realtime, Storage, Cron and Queues, the supabase-js and SSR libraries in frameworks such as Next.js and SvelteKit, the Supabase CLI and MCP server, migrations, and debugging errors from the logs.","The Supabase skill's first rule is that Supabase changes often, so the agent reads the Supabase changelog and current docs instead of relying on what it remembers. It also tells the agent to test every fix with a real query, and to change approach after two or three failed attempts instead of repeating the same command.","The Supabase skill carries a checklist of security traps that fail silently: access rules built on metadata users can edit themselves, secret keys in browser code, views that skip row level security, updates that need a SELECT policy to work, and the deprecated auth.role() check that lets anonymous users through."],"useItWhen":["You are building sign-in, sessions or row level security on Supabase.","A table you created with SQL is unexpectedly not reachable from your app.","You are writing migrations or Edge Functions and want them checked against current Supabase docs.","You want an agent to review a Supabase project for security gaps."],"notFor":["You work on plain Postgres outside Supabase: use Postgres Best Practices."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add supabase/agent-skills --skill supabase"},{"label":"Or, for Claude Code, add Supabase's plugin marketplace from your terminal","command":"claude plugin marketplace add supabase/agent-skills"},{"label":"Then install the supabase plugin","command":"claude plugin install supabase@supabase-agent-skills"}],"sourceUpdated":"2026-09-28","version":"0.1.2","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:45:26+02:00"},{"url":"https://cookiesforai.app/skills/trailofbits/supply-chain-risk-auditor","owner":"trailofbits","slug":"supply-chain-risk-auditor","name":"Supply chain risk auditor","skillName":"supply-chain-risk-auditor","title":"supply-chain-risk-auditor skill: audit npm, PyPI and Go packages","summary":"supply-chain-risk-auditor is Trail of Bits' agent skill for auditing a project's npm, PyPI and Go dependencies for known advisories, abandoned upstreams, publisher concentration and install scripts, measured by bundled scripts.","category":"security","source":"trailofbits/skills","path":"plugins/supply-chain-risk-auditor/skills/supply-chain-risk-auditor","license":"CC-BY-SA-4.0","licenseNote":"CC-BY-SA-4.0 is a share-alike license. You can copy the skill into a project and change it, even for commercial work, if you credit Trail of Bits, link the license and say what you changed. If you share a changed version, for example in a public repository, it must stay under CC-BY-SA-4.0. It is a content license, not a software license, so ask your legal team before bundling it inside a closed product.","worksWith":["Claude Code","Codex","ChatGPT workspace marketplace"],"cliChecked":"2026-10-01","whatItDoes":["supply-chain-risk-auditor is an agent skill from security firm Trail of Bits that produces a risk report for a project's direct dependencies on npm, PyPI and Go, plus an advisory sweep of everything the lockfile resolves. It checks version-matched security advisories, archived or abandoned upstream repositories, how few people can publish a package on npm, and scripts that run at install time.","The measuring in supply-chain-risk-auditor is done by two bundled Python scripts, not by the agent. Trail of Bits explains why: hand-collected figures were wrong before, such as GitHub contributor counts suggesting five maintainers where npm showed one. The scripts mark each criterion as clean, flagged or unassessable with a reason, and refuse to print a report when they measured nothing.","The agent adds only judgment, kept apart from the measured data: what to fix first, whether an advisory fix is a patch or a major upgrade away, replacement candidates for abandoned packages, and whether npm ci with install scripts disabled would work. It reads lockfiles from npm (package-lock), uv and Go modules; it does not read poetry.lock, pnpm-lock.yaml or yarn.lock."],"useItWhen":["You want to know which of your dependencies have known vulnerabilities at the versions you use.","You suspect some of your packages are abandoned or archived upstream.","You are about to start a security engagement and need a dependency risk baseline.","You want to know which npm packages run scripts when they install."],"notFor":["License compliance checks or scanning your own source code for secrets: the skill does neither.","Projects outside npm, PyPI and Go, such as Rust crates or Maven."],"requirements":["uv to run the bundled scripts","The GitHub CLI signed in; without it GitHub allows 60 requests an hour and repository checks come back unassessable","A package.json, pyproject.toml, requirements file or go.mod"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add trailofbits/skills --skill supply-chain-risk-auditor"},{"label":"Or, in Claude Code, add Trail of Bits' plugin marketplace","command":"/plugin marketplace add trailofbits/skills"},{"label":"Or, in Codex, add the same marketplace from your terminal","command":"codex plugin marketplace add trailofbits/skills"},{"label":"Then install the supply-chain-risk-auditor plugin in Codex","command":"codex plugin add supply-chain-risk-auditor@trailofbits"}],"installNote":"In Claude Code, open /plugin menu after adding the marketplace and install the supply-chain-risk-auditor plugin. For a ChatGPT workspace, Trail of Bits' README says to import the repository's .claude-plugin/marketplace.json.","faq":[{"q":"Can I copy Trail of Bits' supply-chain-risk-auditor skill into my own repository?","a":"Yes, with conditions. The skills are CC-BY-SA-4.0. You may copy and adapt them, even commercially, if you credit Trail of Bits, link the license and note your changes. Any changed version you share must carry the same CC-BY-SA-4.0 license. Using the skill unchanged in your own work does not put your code under that license."},{"q":"Does supply-chain-risk-auditor support pnpm or Yarn lockfiles?","a":"Not on 2026-10-01. It reads package-lock.json, npm-shrinkwrap.json, uv.lock and Go 1.17+ go.mod files. When yarn.lock, pnpm-lock.yaml or poetry.lock are present, the report notes this, and it uses pinned versions or, without them, the latest release."}],"sourceUpdated":"2026-09-16","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/obra/systematic-debugging","owner":"obra","slug":"systematic-debugging","name":"Systematic Debugging","skillName":"systematic-debugging","title":"systematic-debugging skill: find the root cause before fixing","summary":"Systematic Debugging is an agent skill from Superpowers that makes your agent find the root cause of a bug or failing test before it proposes any fix, instead of guessing.","category":"testing-and-debugging","source":"obra/superpowers","path":"skills/systematic-debugging","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Systematic Debugging is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads for any bug, test failure, build failure or unexpected behavior, before the agent proposes a fix. Its central rule: no fixes without a root cause investigation first. The rule holds when the issue looks simple, when time is short, and after earlier fixes have failed.","Under Systematic Debugging, the agent works in four phases. First it reads the error messages in full, reproduces the problem, checks recent changes, and in systems with several parts logs what enters and leaves each one to see where it breaks. Then it compares the broken code with similar working code. Then it tests one hypothesis at a time with the smallest change. Finally it writes a failing test with Test-Driven Development, makes one fix, and confirms it with Verification Before Completion.","Under Systematic Debugging, if three fixes have failed, the agent stops and questions the architecture with you instead of trying a fourth. The folder holds supporting guides on tracing a bug backward through the call stack, adding validation at several layers, and replacing fixed timeouts with waiting for a condition, plus a script that finds which test leaves unwanted files or state behind."],"useItWhen":["Your agent keeps trying quick fixes that do not solve the bug.","You have a failing test and want to know why before anything changes.","You have a build or CI step that breaks somewhere across several layers and cannot see where.","Your tests are flaky and seem to depend on timing."],"notFor":["You have several unrelated failures at once: use Dispatching Parallel Agents."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill systematic-debugging"}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/mattpocock/tdd","owner":"mattpocock","slug":"tdd","name":"TDD","skillName":"tdd","title":"tdd skill: test-driven development with an AI coding agent","summary":"tdd is Matt Pocock's agent skill for test-driven development: the agent agrees which public interfaces to test with you, then works one failing test and one minimal fix at a time.","category":"testing-and-debugging","source":"mattpocock/skills","path":"skills/engineering/tdd","license":"MIT","worksWith":["Claude Code","Codex","Any agent that reads SKILL.md files, through the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["TDD is a skill in Matt Pocock's skills collection that shapes how an agent writes tests while building a feature or fixing a bug. It loads when you ask for test-first work, mention red, green, refactor, or want integration tests. Its core idea is that tests check behavior through public interfaces, so code can be rewritten without breaking them.","Before writing any test, the agent lists the seams it plans to test, meaning the public boundaries where behavior can be observed, and asks you to confirm them. No test goes in at an unconfirmed seam. The loop then runs in vertical slices: one test, one minimal implementation, repeat. Refactoring is left out of the loop and belongs to a separate review step.","The skill names three anti-patterns: tests coupled to implementation, tautological tests that recompute the expected value the same way the code does, and horizontal slicing where all tests are written before any code. Two short guides cover what a good test looks like and when to mock, which the skill limits to system boundaries such as external APIs, time and randomness."],"useItWhen":["You want your agent to write tests first instead of after the code.","Your agent's tests break on every refactor even when behavior stays the same.","You want to decide which parts of a module get tested before the agent starts.","Your agent mocks your own modules and you want it to stop."],"notFor":["A strict rule that forbids keeping code written before its test: obra's Test-Driven Development skill in Superpowers enforces that."],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add mattpocock/skills --skill tdd"},{"label":"Or, in Claude Code, install the whole collection from the official marketplace","command":"/plugin install mattpocock-skills"}],"installNote":"The author says to pick one route: installing both the plugin and the skills CLI copy leaves every skill twice.","sourceUpdated":"2026-09-17","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/hashicorp/terraform-policy","owner":"hashicorp","slug":"terraform-policy","name":"Terraform policy","skillName":"terraform-policy","title":"terraform-policy skill: write, test and convert Sentinel policies","summary":"terraform-policy is HashiCorp's agent skill for writing Terraform Policy files (.policy.hcl), testing them with .policytest.hcl, and converting Sentinel policies to the new format.","category":"infrastructure-as-code","source":"hashicorp/agent-skills","path":"plugins/terraform/skills/terraform-policy","license":"MPL-2.0","licenseNote":"MPL-2.0 is a file-level copyleft license: you can use and copy the skill in any project, including closed source ones, but if you distribute changed versions of its files, those files stay under MPL-2.0 and must be shared as source.","worksWith":["GitHub Copilot","Claude Code","Codex","Cursor","OpenCode","IBM Bob"],"cliChecked":"2026-10-01","whatItDoes":["terraform-policy is an agent skill from HashiCorp for Terraform Policy, the HCL-based policy language checked by the tfpolicy CLI. It handles four jobs: writing a new .policy.hcl policy from a requirement such as blocking unencrypted EC2 volumes, converting a .sentinel policy, writing or fixing a .policytest.hcl test, and migrating a whole Sentinel policy library.","The terraform-policy skill is a router. Its SKILL.md is short and sends the agent to one of two reference guides: one for authoring and converting policies, one for writing tests. Before giving any instructions, the agent checks the installed tfpolicy version, because the skill keeps guidance for the 0.2.x and 0.3.x lines and their rules differ, for example on whether mock resources in tests must declare attributes.","The skill also lists features that only exist in newer versions, such as the core::alltrue and core::anytrue functions and schema checks that run before any test in 0.3.x. If the version is unknown, the agent is told to ask or to give guidance that keeps the two paths apart. A README and example policies sit in the same folder."],"useItWhen":["You are moving from Sentinel to Terraform Policy and want policies converted.","You need a new policy, such as requiring encryption on every storage volume.","A .policytest.hcl test fails and you want help finding out why.","You want tests written for an existing .policy.hcl file."],"notFor":["Reviewing the output of terraform plan before an apply: HashiCorp's 20 skills (read 2026-10-01) include none made only for that job.","Testing Terraform modules (.tftest.hcl): use terraform-test from the same repository."],"requirements":["The tfpolicy CLI, version 0.2.x or 0.3.x"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add hashicorp/agent-skills --skill terraform-policy"},{"label":"Or, for Claude Code, add HashiCorp's plugin marketplace","command":"claude plugin marketplace add hashicorp/agent-skills"},{"label":"Then install the terraform plugin, which holds all 16 Terraform skills","command":"claude plugin install terraform@hashicorp"}],"installNote":"In Codex, HashiCorp's README says to add the repository's .agents/plugins/marketplace.json as a repository marketplace, then install the terraform plugin.","faq":[{"q":"Does HashiCorp have an agent skill that reviews terraform plan output?","a":"No. On 2026-10-01 hashicorp/agent-skills held 20 skills, 16 for Terraform and 4 for Packer, and none is dedicated to reading or reviewing terraform plan output. The closest is terraform-test, which can run tests in plan mode against assertions you write."},{"q":"Can I copy this HashiCorp skill into my own repository?","a":"Yes. The skill is MPL-2.0. You can copy and use it in any project, open or closed source. If you change its files and distribute them, the changed files must stay MPL-2.0 and be available as source; your own files are not affected."}],"sourceUpdated":"2026-09-28","version":"0.1.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/antonbabenko/terraform-skill","owner":"antonbabenko","slug":"terraform-skill","name":"Terraform skill","skillName":"terraform-skill","title":"terraform-skill: write, test and review Terraform and OpenTofu","summary":"terraform-skill is Anton Babenko's agent skill for writing, testing and reviewing Terraform and OpenTofu, which names the risk before it writes code and never recommends a production apply without a reviewed plan.","category":"infrastructure-as-code","source":"antonbabenko/terraform-skill","path":"skills/terraform-skill","license":"Apache-2.0","licenseNote":"Apache-2.0 lets you use, change and share the skill in any project, including closed source ones, if you keep its license and notices. GitHub shows NOASSERTION because it could not match the LICENSE file, which opens with a copyright line and two lines naming websites before the full Apache License 2.0 text; the SKILL.md frontmatter also says Apache-2.0.","worksWith":["Claude Code","Cursor","GitHub Copilot","Gemini CLI","OpenCode","Codex","Kiro","Antigravity","Autohand Code"],"cliChecked":"2026-10-02","whatItDoes":["terraform-skill is an agent skill by Anton Babenko that makes an agent diagnose a Terraform or OpenTofu task before it writes code. The agent first records the runtime and version, the providers, the state backend, where the code runs and how critical the environment is. It then names the failure category, such as identity churn after a refactor, secrets leaking into state, an oversized blast radius, CI drift or state corruption, and loads only the reference file for that category.","Every answer under terraform-skill ends with the same five parts: assumptions and version floor, the risk addressed, the fix and its tradeoffs, the exact commands that validate it (fmt -check, validate, plan -out, a policy check), and how to roll back a change to state. Two rules matter most for reviews: no production apply is recommended without a reviewed plan artifact and an approval, and no destroy runs before terraform plan -destroy has listed every resource it would delete.","terraform-skill also bundles eight reference files the agent reads only when a task needs them: native terraform test versus Terratest, why computed values need command = apply in tests, module layout and variable contracts, S3 native state locking from Terraform 1.10, write_only arguments from 1.11 to keep secrets out of state, and CI templates for GitHub Actions, GitLab CI and Atlantis. Examples default to AWS; the README says Azure and GCP are covered too."],"useItWhen":["You want your agent to review a Terraform or OpenTofu change and refuse a production apply without a saved, reviewed plan.","You are about to run a targeted destroy and want every resource it would delete listed first.","You are choosing between native terraform test and Terratest, or adding mock providers so CI needs no cloud credentials.","You want a CI pipeline that applies exactly the plan someone reviewed instead of planning again."],"notFor":["A ready-made report on a saved plan file: terraform-skill makes a reviewed plan a gate and shows how to export a plan with terraform show -json, but reading the plan is left to the agent and to you.","Writing Terraform Policy or moving Sentinel policies: HashiCorp's terraform-policy skill is built for that."],"requirements":["Terraform 1.0 or newer, or OpenTofu 1.6 or newer; native tests need 1.6 and mock providers 1.7","Optional: HashiCorp's Terraform MCP server for registry lookups, and the terraform-ls language server"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add antonbabenko/terraform-skill --skill terraform-skill"},{"label":"Or, in Claude Code, add Anton Babenko's plugin marketplace","command":"/plugin marketplace add antonbabenko/agent-plugins"},{"label":"Then install the terraform-skill plugin","command":"/plugin install terraform-skill@antonbabenko"},{"label":"Or, in Gemini CLI, install it as an extension","command":"gemini extensions install https://github.com/antonbabenko/terraform-skill"}],"installNote":"For Cursor, Codex, OpenCode, Kiro and other agents, the README gives a git clone into each agent's skills folder.","faq":[{"q":"Does terraform-skill review terraform plan output?","a":"Partly. It does not read a saved plan file for you, but it forbids recommending a production apply without a reviewed plan artifact and an approval, and requires terraform plan -destroy with every deleted resource listed before any destroy. In CI it applies the reviewed plan from the plan stage instead of planning again."},{"q":"Is terraform-skill from HashiCorp?","a":"No. It is Anton Babenko's own project. HashiCorp publishes its skills in hashicorp/agent-skills, which held 16 Terraform skills on 2026-10-01; none of them is built mainly for plan review either."}],"sourceUpdated":"2026-06-03","version":"1.17.1","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/hashicorp/terraform-style-guide","owner":"hashicorp","slug":"terraform-style-guide","name":"Terraform style guide","skillName":"terraform-style-guide","title":"terraform-style-guide skill: write and review Terraform HCL","summary":"terraform-style-guide is HashiCorp's agent skill for writing and reviewing Terraform HCL to the official style conventions: file layout, formatting, naming, typed variables, version pinning and a review checklist.","category":"infrastructure-as-code","source":"hashicorp/agent-skills","path":"plugins/terraform/skills/terraform-style-guide","license":"MPL-2.0","licenseNote":"MPL-2.0 is a file-level copyleft license: you can use and copy the skill in any project, including closed source ones, but if you distribute changed versions of its files, those files stay under MPL-2.0 and must be shared as source.","worksWith":["GitHub Copilot","Claude Code","Codex","Cursor","OpenCode","IBM Bob"],"cliChecked":"2026-10-01","whatItDoes":["terraform-style-guide is an agent skill from HashiCorp that loads when an agent writes, reviews or generates Terraform configuration. It sets the order of work: provider and version constraints first, then data sources, resources in dependency order, outputs, and variables for anything configurable. Files follow a fixed split: terraform.tf, providers.tf, main.tf, variables.tf, outputs.tf and locals.tf, with variables and outputs kept alphabetical.","The formatting and naming rules in terraform-style-guide follow HashiCorp's published style guide. Code uses two-space indentation and aligned equals signs, meta-arguments come before arguments, and lifecycle blocks go last. Names are lowercase with underscores, singular, and do not repeat the resource type. Every variable needs a type and description, every output a description, and secrets are marked sensitive. for_each is preferred over count, except for conditional creation.","The skill ends with a review checklist: terraform fmt and terraform validate pass, versions are pinned, and no credentials are hardcoded. A separate SECURITY.md in the skill folder covers encryption, keeping secrets out of state and ephemeral resources. It also lists what never to commit, such as state files and plan files, and what always to commit, such as the dependency lock file."],"useItWhen":["You want an agent to write a new Terraform configuration that a HashiCorp reviewer would accept.","You are reviewing a Terraform pull request and want a consistent checklist.","Your Terraform files are one long main.tf and you want them split the standard way.","You keep seeing untyped variables or outputs without descriptions in your modules."],"notFor":["Reviewing the output of terraform plan before an apply: HashiCorp's 20 skills (read 2026-10-01) include none made only for that job.","Splitting a large configuration into reusable modules: use refactor-module from the same repository."],"requirements":["Terraform CLI for terraform fmt and terraform validate; tflint, checkov or tfsec are optional extras the skill mentions"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add hashicorp/agent-skills --skill terraform-style-guide"},{"label":"Or, for Claude Code, add HashiCorp's plugin marketplace","command":"claude plugin marketplace add hashicorp/agent-skills"},{"label":"Then install the terraform plugin, which holds all 16 Terraform skills","command":"claude plugin install terraform@hashicorp"}],"installNote":"In Codex, HashiCorp's README says to add the repository's .agents/plugins/marketplace.json as a repository marketplace, then install the terraform plugin.","faq":[{"q":"Does HashiCorp have an agent skill that reviews terraform plan output?","a":"No. On 2026-10-01 hashicorp/agent-skills held 20 skills, 16 for Terraform and 4 for Packer, and none is dedicated to reading or reviewing terraform plan output. The closest is terraform-test, which can run tests in plan mode against assertions you write."},{"q":"Can I copy this HashiCorp skill into my own repository?","a":"Yes. The skill is MPL-2.0. You can copy and use it in any project, open or closed source. If you change its files and distribute them, the changed files must stay MPL-2.0 and be available as source; your own files are not affected."}],"sourceUpdated":"2026-08-10","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/hashicorp/terraform-test","owner":"hashicorp","slug":"terraform-test","name":"Terraform test","skillName":"terraform-test","title":"terraform-test skill: write and run Terraform tests with an agent","summary":"terraform-test is HashiCorp's agent skill for writing and running Terraform's built-in tests: .tftest.hcl files with run blocks, assertions, expected failures, mock providers and CI pipelines.","category":"infrastructure-as-code","source":"hashicorp/agent-skills","path":"plugins/terraform/skills/terraform-test","license":"MPL-2.0","licenseNote":"MPL-2.0 is a file-level copyleft license: you can use and copy the skill in any project, including closed source ones, but if you distribute changed versions of its files, those files stay under MPL-2.0 and must be shared as source.","worksWith":["GitHub Copilot","Claude Code","Codex","Cursor","OpenCode","IBM Bob"],"cliChecked":"2026-10-01","whatItDoes":["terraform-test is an agent skill from HashiCorp that teaches an agent the terraform test framework. It covers test files ending in .tftest.hcl or .tftest.json, run blocks with their own variables and module sources, assert blocks with clear error messages, and expect_failures for checking that validation rules reject bad input. Tests can run in plan mode, which creates nothing, or apply mode, which builds real resources and destroys them afterwards in reverse order.","The skill in terraform-test gives ready patterns for common checks: outputs, conditional resources, resource counts, tags, data source lookups, sequential runs that pass values between each other, parallel runs, shared state keys and multiple aliased providers. It recommends naming plan-mode files *_unit_test.tftest.hcl and apply-mode files *_integration_test.tftest.hcl so CI can run them separately, and defaulting to plan mode unless real behavior must be tested.","Three reference files in the terraform-test folder load only when needed. One covers mock providers, which need Terraform 1.7.0 or newer and let tests run without cloud credentials. One gives GitHub Actions and GitLab CI pipelines that run unit tests on every pull request and integration tests on merge. The last is a complete example suite for a VPC module."],"useItWhen":["You have a Terraform module with no tests and want unit tests that run in plan mode.","You want to test a module without cloud credentials by mocking the provider.","You need to check that a variable's validation rule rejects bad values.","You want Terraform tests running in GitHub Actions or GitLab CI."],"notFor":["Reviewing the output of terraform plan before an apply: HashiCorp's 20 skills (read 2026-10-01) include none made only for that job.","Testing Terraform Policy files (.policytest.hcl): use terraform-policy from the same repository."],"requirements":["Terraform CLI with the test command; mock providers need Terraform 1.7.0 or newer","Cloud credentials only for apply-mode integration tests"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add hashicorp/agent-skills --skill terraform-test"},{"label":"Or, for Claude Code, add HashiCorp's plugin marketplace","command":"claude plugin marketplace add hashicorp/agent-skills"},{"label":"Then install the terraform plugin, which holds all 16 Terraform skills","command":"claude plugin install terraform@hashicorp"}],"installNote":"In Codex, HashiCorp's README says to add the repository's .agents/plugins/marketplace.json as a repository marketplace, then install the terraform plugin.","faq":[{"q":"Does HashiCorp have an agent skill that reviews terraform plan output?","a":"No. On 2026-10-01 hashicorp/agent-skills held 20 skills, 16 for Terraform and 4 for Packer, and none is dedicated to reading or reviewing terraform plan output. The closest is terraform-test, which can run tests in plan mode against assertions you write."},{"q":"Can I copy this HashiCorp skill into my own repository?","a":"Yes. The skill is MPL-2.0. You can copy and use it in any project, open or closed source. If you change its files and distribute them, the changed files must stay MPL-2.0 and be available as source; your own files are not affected."}],"sourceUpdated":"2026-08-10","version":"0.0.2","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/test-driven-development","owner":"obra","slug":"test-driven-development","name":"Test-Driven Development","skillName":"test-driven-development","title":"test-driven-development skill: AI agents write tests first","summary":"Test-Driven Development is an agent skill from Superpowers: your agent writes a failing test first, watches it fail, then writes only the code needed to make it pass.","category":"testing-and-debugging","source":"obra/superpowers","path":"skills/test-driven-development","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Test-Driven Development is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when the agent implements any feature or bug fix, before it writes implementation code. The rule is strict: \"No production code without a failing test first.\" If code was written before its test, the agent deletes it and starts over, and does not keep it as a reference.","Under Test-Driven Development, each change follows red, green, refactor. The agent writes one small test for one behavior, runs it and confirms it fails for the expected reason, and writes the simplest code that passes. It then runs the project's whole test suite, not just the new file, and reports any failure by name. Cleanup comes only after the tests pass. Exceptions such as throwaway prototypes, generated code or configuration files need your permission.","A guide bundled with Test-Driven Development, writing-good-tests.md, covers the tests themselves: name the break each test should catch, assert on real behavior instead of mocks, keep test-only code out of production classes, and learn what a dependency does before replacing it with a mock. The skill also lists the excuses agents give for skipping tests, such as \"too simple to test\" or \"I'll test after\", with an answer to each."],"useItWhen":["Your agent writes code first and adds tests afterwards, if at all.","You are fixing a bug and want a test that proves it is gone and stays gone.","Your agent's tests pass but only check mocks, not real behavior."],"notFor":["You have a bug and do not know its cause yet: start with Systematic Debugging, which hands the fix to this skill."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill test-driven-development"}],"faq":[{"q":"What does the TDD skill do if the agent already wrote the code?","a":"It tells the agent to delete that code and start over from a failing test, not keep it as a reference. Exceptions such as throwaway prototypes, generated code or configuration files need your permission."},{"q":"Can I install only the TDD skill without the rest of Superpowers?","a":"Yes: npx skills add obra/superpowers --skill test-driven-development installs just this folder. It still refers to sibling skills such as Systematic Debugging, and without the full plugin no session-start hook reminds the agent to use it."}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/theme-factory","owner":"anthropics","slug":"theme-factory","name":"Theme Factory","skillName":"theme-factory","title":"theme-factory skill: color and font themes for slides and docs","summary":"Theme Factory is Anthropic's open-source agent skill with 10 ready color and font themes, plus custom ones on request, that an agent applies to slides, documents and web pages.","category":"frontend-and-design","source":"anthropics/skills","path":"skills/theme-factory","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"cliChecked":"2026-10-01","whatItDoes":["Theme Factory is an Anthropic skill for giving an artifact a consistent look. It works on slide decks, documents, reports and HTML landing pages. It ships ten themes, each a file with a color palette in hex codes and a pair of fonts for headings and body text: Ocean Depths, Sunset Boulevard, Forest Canopy, Modern Minimalist, Golden Hour, Arctic Frost, Desert Rose, Tech Innovation, Botanical Garden and Midnight Galaxy.","With Theme Factory, the agent first shows you a bundled PDF that previews all ten, then waits for you to choose. It reads the chosen theme's file and applies its colors and fonts across the whole artifact, keeping text readable against its background. When none of the ten fits, the agent creates a new theme from your description, names it in the same style and shows it for your review before applying it."],"useItWhen":["You have a slide deck or report that needs a consistent color palette and font pairing.","You want to pick a look from visual previews instead of describing colors yourself.","You want a custom theme built from a short description, such as the mood or the audience."],"notFor":["You want a new layout or design direction, not only colors and fonts: use Frontend Design or Canvas Design."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill theme-factory"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/nextlevelbuilder/ui-ux-pro-max","owner":"nextlevelbuilder","slug":"ui-ux-pro-max","name":"UI UX Pro Max","skillName":"ui-ux-pro-max","title":"ui-ux-pro-max skill: design systems and UX rules for agents","summary":"ui-ux-pro-max is an agent skill from Next Level Builder that gives an agent a searchable local library of UI styles, color palettes, font pairings and UX rules, and generates a design system per project.","category":"frontend-and-design","source":"nextlevelbuilder/ui-ux-pro-max-skill","path":".claude/skills/ui-ux-pro-max","license":"MIT","worksWith":["Claude Code","Cursor","Windsurf","Codex CLI","Gemini CLI","GitHub Copilot","Kiro","OpenCode","Any agent that reads SKILL.md files, through the skills CLI"],"agentNote":"The search commands in the SKILL.md use Claude Code's ${CLAUDE_PLUGIN_ROOT} variable in the script path. In other agents, or in a skills CLI copy, the agent has to point Python at search.py inside the installed skill folder.","cliChecked":"2026-10-01","whatItDoes":["UI UX Pro Max is a design skill whose core is a set of local data files and a Python search script, not just written rules. On 2026-10-01 its SKILL.md counts 79 searchable styles, 192 product palettes with reasoning profiles, 74 font pairings, 119 UX guidelines, 25 chart types and 22 technology stacks, from React and SwiftUI to Flutter and WPF.","For a new project or page, the agent runs the script with a design-system flag and a description such as a beauty spa booking site. The output names a page pattern, a style, colors, a font pairing, key effects, anti-patterns to avoid and a pre-delivery checklist. With a persist flag it saves this as design-system files in the project, so later sessions read the same decisions instead of regenerating them.","The skill ranks its rules in ten priority groups, starting with accessibility and touch targets, then performance, style choice, responsive layout, typography and color, animation, forms, navigation and charts. The agent queries one domain at a time for details, so the full rule text stays out of context until it is needed."],"useItWhen":["You are starting a new app or landing page and want a coherent style, palette and font pairing chosen for its industry.","You want design decisions saved in the repo so every session follows them.","You want an accessibility and UX checklist run before a UI ships.","You build for a stack other than React, such as SwiftUI, Flutter or Jetpack Compose."],"notFor":["Backend, API or infrastructure work: the skill says to skip it unless the task changes how something looks or behaves."],"requirements":["Python 3 for the search script (standard library only, per the README)"],"install":[{"label":"Install the skill with the skills CLI","command":"npx skills add nextlevelbuilder/ui-ux-pro-max-skill --skill ui-ux-pro-max"},{"label":"Or, in Claude Code, add the publisher's plugin marketplace","command":"/plugin marketplace add nextlevelbuilder/ui-ux-pro-max-skill"},{"label":"Then install the plugin","command":"/plugin install ui-ux-pro-max@ui-ux-pro-max-skill"},{"label":"Or install the publisher's CLI, then run uipro init with your agent, for example --ai cursor","command":"npm install -g ui-ux-pro-max-cli"}],"sourceUpdated":"2026-09-27","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/using-git-worktrees","owner":"obra","slug":"using-git-worktrees","name":"Using Git Worktrees","skillName":"using-git-worktrees","title":"using-git-worktrees skill: an isolated worktree for agent work","summary":"Using Git Worktrees is an agent skill from Superpowers: before feature work, your agent makes sure it is in an isolated git worktree, sets up the project and checks that the tests pass.","category":"code-review-and-git","source":"obra/superpowers","path":"skills/using-git-worktrees","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Using Git Worktrees is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads before feature work that needs isolation from the current workspace, and before executing an implementation plan. It first checks whether the agent is already in a linked worktree, with a guard for git submodules, and if so it creates nothing new. In a normal checkout it asks for your consent, unless your instructions already state a preference.","To create the workspace under Using Git Worktrees, the agent prefers a native tool from its harness, such as an EnterWorktree tool or a /worktree command, and falls back to git worktree add only when none exists. For the fallback it uses a directory named in your instructions, then an existing .worktrees/ or worktrees/ folder, then .worktrees/ by default. It checks that the folder is git-ignored, and adds it to .gitignore and commits if not.","Using Git Worktrees then runs the project setup it detects (npm install, cargo build, pip or poetry install, go mod download) and runs the tests to confirm a clean baseline. If the tests fail, it reports the failures and asks whether to continue. If a sandbox blocks worktree creation, it tells you and works in the current directory instead."],"useItWhen":["You want your agent to build a feature without touching the branch you are working on.","You want each piece of work to start from a clean checkout with passing tests.","You are about to have an agent carry out an implementation plan."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill using-git-worktrees"}],"sourceUpdated":"2026-07-24","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/using-superpowers","owner":"obra","slug":"using-superpowers","name":"Using Superpowers","skillName":"using-superpowers","title":"using-superpowers skill: make your agent check and use its skills","summary":"Using Superpowers is the agent skill that makes your agent check for a relevant Superpowers skill before it responds, announce it, and follow it whenever one applies.","category":"planning-and-workflow","source":"obra/superpowers","path":"skills/using-superpowers","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Using Superpowers is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads at the start of any conversation and sets how the agent finds and uses skills. The agent checks for relevant skills before any response or action, including clarifying questions and exploring the codebase. If there is even a 1% chance a skill applies, the agent invokes it, announces which skill it is using, and follows it.","In Using Superpowers, process skills come first and set the approach, then implementation skills carry it out: a request to build something starts with Brainstorming, and a bug starts with Systematic Debugging. Before entering plan mode, the agent brainstorms first. A table of red flags lists the thoughts that mean the agent is about to skip a skill, such as \"this is just a simple question\" or \"let me explore the codebase first\".","Reference files in Using Superpowers map the actions skills ask for to the tools of each agent, with notes for Claude Code, Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Your own instructions, in files such as CLAUDE.md, AGENTS.md or GEMINI.md or in direct requests, take precedence over skills, and skills override the agent's default behavior. A subagent dispatched for a specific task is told to ignore this skill."],"useItWhen":["You installed Superpowers and want your agent to use its skills without being reminded.","Your agent jumps into code or questions before checking whether a skill fits the task.","You use Superpowers in more than one agent tool and want the same behavior in each."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill using-superpowers"}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/vercel-cli-with-tokens","owner":"vercel-labs","slug":"vercel-cli-with-tokens","name":"Vercel CLI with Tokens","skillName":"vercel-cli-with-tokens","title":"vercel-cli-with-tokens skill: deploy with a Vercel access token","summary":"Vercel CLI with Tokens is Vercel's agent skill for running the Vercel CLI with an access token instead of vercel login: deploys, environment variables, logs and domains.","category":"cloud-and-deployment","source":"vercel-labs/agent-skills","path":"skills/vercel-cli-with-tokens","license":"MIT, stated only in the repository README (no license file)","licenseNote":"The SKILL.md has no license key, there is no LICENSE file in the skill folder or the repository, and GitHub reports no license (checked 2026-10-01). The only grant is the line \"License: MIT\" at the end of the repository README, which covers the whole collection.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"cliChecked":"2026-10-01","whatItDoes":["Vercel CLI with Tokens is Vercel's skill for deploying and managing Vercel projects when the CLI signs in with an access token rather than an interactive vercel login. It loads for tasks such as deploying to Vercel, setting up a project or adding environment variables. The agent looks for the token in the environment and in .env files before asking you, and points you to vercel.com/account/tokens if none is found.","Under Vercel CLI with Tokens, the agent exports the token as VERCEL_TOKEN and never passes it as a --token flag, because command-line arguments show up in shell history and process listings. With VERCEL_PROJECT_ID and VERCEL_ORG_ID set together, it deploys without linking. Otherwise it links the project, by git repository when possible. It deploys a preview unless you ask for production, and asks before any git push.","Vercel CLI with Tokens also covers setting environment variables per environment, listing and inspecting deployments, reading build and runtime logs, and adding domains. For projects managed by Stripe Projects, it lists the upgrade and downgrade commands and asks before any plan change, because upgrades bill a real card. Troubleshooting steps cover a missing or expired token, the wrong team and failed builds."],"useItWhen":["You want the agent to work with a Vercel access token instead of an interactive login.","You have a Vercel token in a .env file and want the agent to deploy with it without exposing it.","You want the agent to add, list or pull environment variables for a Vercel project.","You have a failed Vercel deployment and want the agent to read the build logs."],"notFor":["You can sign in interactively and just want a preview link: use Deploy to Vercel."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill vercel-cli-with-tokens"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-04-17","version":"1.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/vercel-optimize","owner":"vercel-labs","slug":"vercel-optimize","name":"Vercel Optimize","skillName":"vercel-optimize","title":"vercel-optimize skill: lower your Vercel bill, speed up routes","summary":"Vercel Optimize is Vercel's agent skill that audits a deployed project from its production metrics and returns ranked, verified fixes to cut cost and speed up slow routes.","category":"cloud-and-deployment","source":"vercel-labs/agent-skills","path":"skills/vercel-optimize","license":"MIT, stated only in the repository README (no license file)","licenseNote":"The SKILL.md has no license key, there is no LICENSE file in the skill folder or the repository, and GitHub reports no license (checked 2026-10-01). The only grant is the line \"License: MIT\" at the end of the repository README, which covers the whole collection.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"agentNote":"With three or more investigations the SKILL.md gives each to its own sub-agent when the host supports them; hosts without sub-agents run them one after another.","cliChecked":"2026-10-01","whatItDoes":["Vercel Optimize runs a cost and performance audit on a project already deployed to Vercel. It loads for requests about a lower Vercel bill, slow or expensive routes, caching, Core Web Vitals, Fluid compute, Fast Data Transfer, Bot Management, Build Minutes or Function Invocations. Next.js, SvelteKit and Nuxt are supported and Astro is limited; for other frameworks the agent asks before running a limited audit.","The Vercel Optimize audit starts from metrics, not from reading code. Bundled Node.js scripts collect 14 days of production metrics, usage and project config through the Vercel CLI, scan the codebase and merge the results. A gate script picks which candidates deserve investigation, six by default, and the agent reads only the files tied to each candidate. With three or more investigations it can give each one to its own sub-agent.","A Vercel Optimize verification script checks every recommendation against real files, line numbers and a bundled library of version-matched Vercel and framework docs, and strips citations that do not fit. The report states cost impact in words, never in dollar amounts. The skill needs Vercel CLI v53 or later, a signed-in CLI in a linked project, Node.js 20 or later, and Observability Plus for route-level recommendations."],"useItWhen":["You saw your Vercel bill go up and want to know which routes or settings drive it.","You have routes in a Next.js, SvelteKit or Nuxt app that are slow or rarely served from cache.","You have poor Core Web Vitals and want fixes tied to real traffic.","You want each recommendation to cite the exact files and docs behind it."],"notFor":["Your project is not deployed on Vercel yet: the audit starts from production metrics.","You want React performance rules applied while you code: use React Best Practices."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill vercel-optimize"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-06-10","version":"1.2.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/obra/verification-before-completion","owner":"obra","slug":"verification-before-completion","name":"Verification Before Completion","skillName":"verification-before-completion","title":"verification-before-completion skill: test before claiming done","summary":"Verification Before Completion is an agent skill from Superpowers: your agent runs the command that proves a claim and reads the output before it says work is done, fixed or passing.","category":"testing-and-debugging","source":"obra/superpowers","path":"skills/verification-before-completion","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Verification Before Completion is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when the agent is about to say work is complete, fixed or passing, and before it commits or opens a pull request. Its rule: \"No completion claims without fresh verification evidence.\" If the agent has not run the check in the current message, it cannot claim the check passes.","Before any claim, Verification Before Completion has the agent name the command that proves it, run it in full, read the whole output and the exit code, and only then state the result with the evidence. A table spells out what each claim needs: passing tests need test output with zero failures, a build needs exit code 0, a fixed bug needs the original symptom tested again, and a subagent's reported success needs the change confirmed in the version control diff.","The red flags in Verification Before Completion include words such as \"should\", \"probably\" and \"seems to\", and expressions of satisfaction like \"Done\" before any check has run. The rule covers paraphrases and anything that implies success, not only exact phrases. A regression test counts only after a red-green cycle: it passes, fails when the fix is reverted, and passes again when the fix is restored."],"useItWhen":["Your agent says the tests pass when it has not run them.","You delegate work to subagents and need their reports checked against the actual changes.","You want every commit or pull request from your agent backed by real test and build output."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill verification-before-completion"}],"sourceUpdated":"2026-07-24","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/bradautomates/watch","owner":"bradautomates","slug":"watch","name":"Watch","skillName":"watch","title":"watch skill: let your AI agent watch YouTube and other videos","summary":"Watch is an open-source agent skill that turns a video link or file into frames and a transcript, or hands it to Google's Gemini, so your agent can answer questions about it.","category":"video-and-media","source":"bradautomates/claude-video","path":"skills/watch","license":"MIT","worksWith":["Claude Code (terminal, VS Code and Claude Desktop Code sessions)","Codex (app, CLI and IDE extension)","Cursor","GitHub Copilot","other local agents through the skills CLI"],"agentNote":"The SKILL.md lists Claude Code's allowed tools (Bash, Read, AskUserQuestion) but says they are examples: any host with a shell, an image viewer and a way to ask you questions works. The /watch command is Claude Code's slash command form; in Codex you ask for watch by name.","cliChecked":"2026-10-01","whatItDoes":["Watch is an agent skill that lets your agent answer questions about a video. Give it a video link or a file on your computer plus a question, and the agent answers from what is in the video. Links can come from YouTube or any other site yt-dlp can download, and the frames and transcript it works from carry timestamps.","Watch has two engines. The local engine downloads the video with yt-dlp, pulls frames with ffmpeg and takes the transcript from the video's own captions, falling back to WhisperX on your computer or a cloud Whisper service (Groq or OpenAI). With a free Google Gemini API key, the Gemini engine lets Google's video model watch the whole video, sound included: YouTube links are sent to Google, and other videos are uploaded to Google and deleted after the answer.","On first use of Watch, a Python setup script checks for the tools it needs and asks which engine you want. The local engine needs yt-dlp and ffmpeg installed. Settings and API keys stay in ~/.config/watch/.env on your machine, and the optional local WhisperX transcriber installs into its own environment under ~/.cache/watch."],"useItWhen":["You want the key points of a long YouTube talk or tutorial without watching all of it.","You need to know what happens at a certain moment in a screen recording or bug report video.","You want your agent to learn from a product demo or course video and answer questions about it."],"notFor":["You use Claude.ai chat, Cowork or a browser-only app: the publisher says Watch does not work there.","Your videos must not leave your machine: use the local engine, since the Gemini engine sends YouTube links to Google and uploads other videos."],"install":[{"label":"Inside Claude Code, add the claude-video marketplace","command":"/plugin marketplace add bradautomates/claude-video"},{"label":"Then install the watch plugin","command":"/plugin install watch@claude-video"},{"label":"Or, for other agents, run this in your computer's terminal (needs Node.js)","command":"npx skills add bradautomates/claude-video -g --skill watch"}],"installNote":"The publisher says it does not work in the Claude chat app or in Cowork.","sourceUpdated":"2026-09-25","version":"0.3.2","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/web-artifacts-builder","owner":"anthropics","slug":"web-artifacts-builder","name":"Web Artifacts Builder","skillName":"web-artifacts-builder","title":"web-artifacts-builder skill: React apps as Claude.ai artifacts","summary":"Web Artifacts Builder is Anthropic's agent skill for complex Claude.ai artifacts: the agent sets up a React, Tailwind and shadcn/ui project, then bundles it into one HTML file.","category":"frontend-and-design","source":"anthropics/skills","path":"skills/web-artifacts-builder","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"agentNote":"Web Artifacts Builder is written for Claude.ai artifacts. Its scripts only need Node 18 or later and a shell, so another agent can produce the same single HTML file, but there is no artifact view outside Claude.","cliChecked":"2026-10-01","whatItDoes":["Web Artifacts Builder is an Anthropic skill for Claude.ai HTML artifacts that need state management, routing or shadcn/ui components. It is meant for complex, multi-component artifacts, not simple single-file HTML or JSX ones. The stack is React 18 with TypeScript, Vite, Tailwind CSS, shadcn/ui and Parcel for bundling. Its design rule steers away from heavily centered layouts, purple gradients, uniform rounded corners and the Inter font.","With Web Artifacts Builder, Claude runs a bundled setup script that creates a configured project with path aliases, more than 40 shadcn/ui components and their Radix UI dependencies. It needs Node 18 or later and picks a Vite version to match. Claude edits the code, then runs a second script that builds with Parcel and inlines all JavaScript, CSS and assets into a single bundle.html to share as an artifact. Testing is optional and comes after you see the result, when you ask or when issues arise."],"useItWhen":["You want a multi-view tool or dashboard in Claude.ai that needs state and routing.","You want shadcn/ui components in an artifact.","You need a React app delivered as one self-contained HTML file."],"notFor":["You want a simple single-file HTML or JSX artifact: the skill says it is not for those."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill web-artifacts-builder"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/web-design-guidelines","owner":"vercel-labs","slug":"web-design-guidelines","name":"Web Design Guidelines","skillName":"web-design-guidelines","title":"web-design-guidelines skill: review UI code for accessibility","summary":"Web Design Guidelines is Vercel's agent skill for UI review: the agent checks your files against Vercel's Web Interface Guidelines and lists each issue by file and line.","category":"frontend-and-design","source":"vercel-labs/agent-skills","path":"skills/web-design-guidelines","license":"MIT, stated only in the repository README (no license file)","licenseNote":"The SKILL.md has no license key, there is no LICENSE file in the skill folder or the repository, and GitHub reports no license (checked 2026-10-01). The only grant is the line \"License: MIT\" at the end of the repository README, which covers the whole collection.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"agentNote":"The SKILL.md tells the agent to use WebFetch, Claude Code's fetch tool, and declares an argument hint for Claude Code's slash command. Any agent that can fetch a URL can follow it.","cliChecked":"2026-10-01","whatItDoes":["Web Design Guidelines is Vercel's skill for reviewing UI code against its Web Interface Guidelines, a rule set Vercel keeps in a separate repository. It loads when you ask an agent to review your UI, check accessibility, audit design, review UX or check a site against best practices. You pass it a file or a file pattern; with none, the agent asks which files to review.","Before each Web Design Guidelines review the agent fetches the latest guidelines from GitHub, so the rules are not stored in the skill and the agent needs web access. The current rules include accessibility, focus states, forms, animation, typography, images, performance, navigation and state, touch, safe areas, dark mode, localization, hydration safety, hover states and copy, plus anti-patterns to flag. Findings come back grouped by file in a short file:line format."],"useItWhen":["You finished a page or component and want an accessibility and UX check before you ship it.","You want UI issues listed with file and line numbers you can click through.","You are reviewing a pull request that changes forms, focus states or animations."],"notFor":["You want docs or other prose reviewed: use Writing Guidelines.","Your agent has no web access: the rules are fetched from GitHub on every run."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill web-design-guidelines"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-01-16","version":"1.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/webapp-testing","owner":"anthropics","slug":"webapp-testing","name":"Webapp Testing","skillName":"webapp-testing","title":"webapp-testing skill: test local web apps with Playwright","summary":"Webapp Testing is Anthropic's open-source agent skill for testing local web apps: the agent writes Python Playwright scripts that click through your app, take screenshots and read console logs.","category":"testing-and-debugging","source":"anthropics/skills","path":"skills/webapp-testing","license":"Apache-2.0","licenseNote":"Apache-2.0 comes from the LICENSE.txt in the skill's own folder (read 2026-10-01); the anthropics/skills repository has no root license file. In practice you may use, change and share the skill in any agent, commercially too, as long as you keep the license text and notices and mark your changes.","worksWith":["Claude Code","Claude.ai","Claude API"],"cliChecked":"2026-10-01","whatItDoes":["Webapp Testing is an Anthropic skill for interacting with and testing web applications running on your own machine. It loads when you want to verify that a frontend works, debug UI behavior, capture browser screenshots or read browser logs. The agent writes native Python Playwright scripts that launch Chromium in headless mode.","A decision tree in Webapp Testing picks the approach. For a static HTML file, the agent reads the file to find selectors. For a dynamic app it uses reconnaissance then action: open the page, wait until network activity settles, take a screenshot or inspect the DOM, then act on the selectors it found. The skill flags one common mistake: inspecting a dynamic page before it has finished loading.","A helper script bundled with Webapp Testing starts one or more servers, such as a backend and a frontend, waits until each port answers, runs your test script and then cleans up. The agent runs it with --help and uses it as a black box instead of reading its source. Example scripts show how to find buttons, links and inputs, automate a local HTML file and capture console logs."],"useItWhen":["You changed a frontend and want an agent to click through it and confirm it still works.","You have a UI bug that only shows up in the browser and want screenshots and console logs to debug it.","Your app needs a backend and a frontend running at the same time before it can be tested."],"notFor":["You want UI code reviewed for accessibility and UX rules without running the app: use Web Design Guidelines."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the example-skills plugin, which holds this skill and Anthropic's other example skills","command":"/plugin install example-skills@anthropic-agent-skills"},{"label":"Or install only this skill, for any agent, with the skills CLI","command":"npx skills add anthropics/skills --skill webapp-testing"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","sourceUpdated":"2026-04-20","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/cloudflare/workers-best-practices","owner":"cloudflare","slug":"workers-best-practices","name":"Workers best practices","skillName":"workers-best-practices","title":"workers-best-practices skill: write and review Cloudflare Workers","summary":"workers-best-practices is Cloudflare's agent skill for writing, reviewing and configuring production Cloudflare Workers, built around a table of anti-patterns to flag and current docs over memorized APIs.","category":"cloud-and-deployment","source":"cloudflare/skills","path":"skills/workers-best-practices","license":"Apache-2.0","worksWith":["Claude Code","Codex","GitHub Copilot in VS Code","Cursor","OpenCode","Pi"],"cliChecked":"2026-10-01","whatItDoes":["workers-best-practices is an agent skill from Cloudflare that loads when an agent writes, reviews or configures a Cloudflare Worker. Its first instruction is that the agent's built-in knowledge of Workers APIs may be out of date, so it should first look at what the project actually uses (package versions, the generated types and the compatibility settings in the Wrangler config) and fetch Cloudflare's documentation before stating facts about APIs, limits or runtime behavior.","The core of workers-best-practices is a list of 15 anti-patterns with the fix for each. Examples: buffering an unbounded response body instead of streaming it, Math.random for security tokens, background work not tied to ctx.waitUntil, request state kept in module-level variables, hand-written Env types instead of wrangler types, comparing secrets with plain string equality, and calling the REST API when a binding exists.","The skill also asks for a current compatibility date on new Workers and for Workers Logs and Traces to be switched on in production, noting that traces need their own setting. Three reference files cover configuration and observability, runtime patterns, and platform API checks. Durable Objects and Wrangler commands are handed to their own skills."],"useItWhen":["You want a code review of a Worker before it goes to production.","Your Worker runs out of memory or loses background work after the response.","You are starting a new Worker and want bindings, types and observability set up right.","An agent keeps writing Workers code from an outdated memory of the API."],"notFor":["Running Wrangler commands, deploying or managing KV, R2 and D1: use the wrangler skill from the same repository."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add cloudflare/skills --skill workers-best-practices"},{"label":"Or, in Claude Code, add Cloudflare's plugin marketplace","command":"/plugin marketplace add cloudflare/skills"},{"label":"Then install the cloudflare plugin (all 16 skills plus Cloudflare's MCP server)","command":"/plugin install cloudflare@cloudflare"},{"label":"Or, in Codex, add the same marketplace from your terminal","command":"codex plugin marketplace add cloudflare/skills"},{"label":"Then install the cloudflare plugin and start a new Codex session","command":"codex plugin add cloudflare@cloudflare"}],"installNote":"In VS Code with GitHub Copilot, Cloudflare's README says to enable chat.plugins.enabled, run Chat: Install Plugin From Source, and enter https://github.com/cloudflare/skills. In Cursor, install it from the Cursor Marketplace.","sourceUpdated":"2026-10-01","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/cloudflare/wrangler","owner":"cloudflare","slug":"wrangler","name":"Wrangler","skillName":"wrangler","title":"wrangler skill: deploy Cloudflare Workers with an AI agent","summary":"wrangler is Cloudflare's agent skill for running and troubleshooting the Wrangler CLI: configuring Worker projects, local development, Previews, secrets, deployments, rollbacks and Cloudflare resources such as KV, R2 and D1.","category":"cloud-and-deployment","source":"cloudflare/skills","path":"skills/wrangler","license":"Apache-2.0","worksWith":["Claude Code","Codex","GitHub Copilot in VS Code","Cursor","OpenCode","Pi"],"agentNote":"The skill prefers the docs tool of Cloudflare's MCP server when the agent has it, which the Claude Code and Codex plugins install; without it the agent fetches the linked documentation pages directly.","cliChecked":"2026-10-01","whatItDoes":["wrangler is an agent skill from Cloudflare for work done through the Wrangler CLI. Before writing any command it has the agent inspect the project: package manager, the locally installed Wrangler version, scripts, framework and the config file actually used for build and deploy. It tells the agent not to rely on memorized flags and not to upgrade Wrangler silently, and to look up the exact command or config field in Cloudflare's docs.","The wrangler skill carries guardrails for changes that touch real resources. The agent confirms the target account, Worker and environment first, checks which bindings are local and which are remote, reconciles dashboard edits that a deploy would overwrite, and keeps secret values out of commands and logs. It treats wrangler secret put as a deployment, because it ships a new version at once, and warns that rollbacks do not roll back data.","A section on Workers Previews covers branch and pull request environments, which need project-local Wrangler 4.135.0 or later, and reminds the agent that Preview URLs are public unless protected. Validation means regenerating types, running tests and a dry-run deploy, then reporting what changed and what stayed unverified."],"useItWhen":["You want to deploy a Worker or a framework app to Cloudflare from the terminal.","You need a KV, R2, D1 or Queues binding added to your Wrangler config.","You want a preview environment for each pull request.","A wrangler command fails with an authentication or permissions error."],"notFor":["Projects that use cloudflare.config.ts or the newer cf CLI: the skill itself says to follow Cloudflare's cf CLI docs instead."],"requirements":["A Cloudflare account and Wrangler installed in the project; Workers Previews need Wrangler 4.135.0 or later"],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add cloudflare/skills --skill wrangler"},{"label":"Or, in Claude Code, add Cloudflare's plugin marketplace","command":"/plugin marketplace add cloudflare/skills"},{"label":"Then install the cloudflare plugin (all 16 skills plus Cloudflare's MCP server)","command":"/plugin install cloudflare@cloudflare"},{"label":"Or, in Codex, add the same marketplace from your terminal","command":"codex plugin marketplace add cloudflare/skills"},{"label":"Then install the cloudflare plugin and start a new Codex session","command":"codex plugin add cloudflare@cloudflare"}],"installNote":"In VS Code with GitHub Copilot, Cloudflare's README says to enable chat.plugins.enabled, run Chat: Install Plugin From Source, and enter https://github.com/cloudflare/skills. In Cursor, install it from the Cursor Marketplace.","sourceUpdated":"2026-10-01","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/vercel-labs/writing-guidelines","owner":"vercel-labs","slug":"writing-guidelines","name":"Writing Guidelines","skillName":"writing-guidelines","title":"writing-guidelines skill: review docs for style, tone, AI tells","summary":"Writing Guidelines is Vercel's agent skill for docs review: the agent checks your pages against Vercel's Writing Guidelines and lists each style or tone issue by file and line.","category":"writing","source":"vercel-labs/agent-skills","path":"skills/writing-guidelines","license":"MIT, stated only in the repository README (no license file)","licenseNote":"The SKILL.md has no license key, there is no LICENSE file in the skill folder or the repository, and GitHub reports no license (checked 2026-10-01). The only grant is the line \"License: MIT\" at the end of the repository README, which covers the whole collection.","worksWith":["any AI coding agent that reads Agent Skills, installed with the skills CLI"],"agentNote":"The SKILL.md tells the agent to use WebFetch, Claude Code's fetch tool, and declares an argument hint for Claude Code's slash command. Any agent that can fetch a URL can follow it.","cliChecked":"2026-10-01","whatItDoes":["Writing Guidelines is Vercel's skill for reviewing documentation and other prose against its Writing Guidelines, a rule set kept in a separate Vercel repository. It loads when you ask an agent to review your docs, check writing style, audit prose, review voice and tone, or check a page against the writing handbook. You pass a file or a file pattern; with none, the agent asks which files to review.","Before each Writing Guidelines review the agent fetches the latest guidelines from GitHub, so the rules are not stored in the skill and the agent needs web access. The current rules include content type, voice and tone, banned words, concision, headings, structure, lists, code, placeholders, units, punctuation and links, plus signs of AI-generated text such as recap transitions and choppy sentences. Findings come back grouped by file in a short file:line format."],"useItWhen":["You wrote a docs page and want its voice, tone and structure checked before you publish.","You want AI-drafted docs cleared of recap transitions, choppy sentences and other tells.","You want style issues listed with file and line numbers so you can fix them one by one."],"notFor":["You want UI code reviewed: use Web Design Guidelines.","You are drafting a new doc from scratch: Doc Co-Authoring guides the writing itself."],"install":[{"label":"Install this skill with the skills CLI","command":"npx skills add vercel-labs/agent-skills --skill writing-guidelines"},{"label":"Or install every Vercel skill at once","command":"npx skills add vercel-labs/agent-skills"}],"sourceUpdated":"2026-05-26","version":"1.0.0","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/obra/writing-plans","owner":"obra","slug":"writing-plans","name":"Writing Plans","skillName":"writing-plans","title":"writing-plans skill: turn a spec into an implementation plan","summary":"Writing Plans is an agent skill from Superpowers: your agent turns a spec into a step-by-step implementation plan with exact files, signatures and tests, before touching code.","category":"planning-and-workflow","source":"obra/superpowers","path":"skills/writing-plans","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.","cliChecked":"2026-10-01","whatItDoes":["Writing Plans is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads once a multi-step task has a spec or a set of requirements and no code has been written yet. The agent writes the plan for an engineer who has never seen the codebase or the spec, recording the decisions that engineer could not make alone: the files to touch, the names and signatures to use, the values taken from the spec, and the tests that prove each task.","Writing Plans saves plans to docs/superpowers/plans/ unless you prefer another place. Each starts with a header naming the goal, architecture, tech stack and spec, a Global Constraints section with values copied from the spec, and a Review Focus list of failure modes no test covers yet. Every task lists its exact files and the interfaces it consumes and produces, split into small checkbox steps: write the failing test, run it, implement, run it again, commit.","Before Writing Plans hands over, the agent checks the plan against the spec for requirements with no task, steps that decide nothing, names that differ between tasks, and a plan far longer than its spec. It then asks you to review the plan and choose how to run it: Subagent-Driven Development, where a fresh subagent does each task, or native execution with Executing Plans. An isolated worktree, if used, comes from Using Git Worktrees."],"useItWhen":["You have a spec or design and want it broken into small, testable tasks before any code is written.","You plan to hand the work to subagents or another session and need a plan that stands on its own.","Your agent's plans say things like \"handle edge cases\" and you want exact files and tests instead."],"notFor":["You do not have a spec or design yet: start with Brainstorming."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill writing-plans"}],"sourceUpdated":"2026-09-25","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T03:01:13+02:00"},{"url":"https://cookiesforai.app/skills/obra/writing-skills","owner":"obra","slug":"writing-skills","name":"Writing Skills","skillName":"writing-skills","title":"writing-skills: create and test agent skills test-first","summary":"Writing Skills is an agent skill from Superpowers: your agent writes new skills test-first, running pressure scenarios with subagents before and after the skill exists.","category":"building-skills-and-mcp","source":"obra/superpowers","path":"skills/writing-skills","license":"MIT","worksWith":["Claude Code","Codex (app and CLI)","Cursor","Gemini CLI","GitHub Copilot CLI","OpenCode","Antigravity","Devin CLI","Factory Droid","Grok Build CLI","Kimi Code","Pi","Qwen Code","Hermes Agent","Muse"],"agentNote":"Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have. This skill needs a subagent tool: Claude Code's Agent tool, or in Codex the multi-agent feature switched on in config.toml (multi_agent = true).","cliChecked":"2026-10-01","whatItDoes":["Writing Skills is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when creating a skill, editing one, or checking that a skill works before deployment. It treats skill writing as test-driven development applied to documentation, and requires an understanding of the Test-Driven Development skill first. Its rule mirrors that one: no skill, and no edit to a skill, without a failing test first.","The test in Writing Skills is a pressure scenario run with a subagent. The agent first runs it without the skill and records what the subagent does and the excuses it gives, then writes the smallest skill that addresses those failures, runs the scenario again to confirm the subagent complies, and closes new loopholes. For wording, it runs micro-tests: five or more samples per variant against a control with no guidance, reading every flagged match by hand.","Writing Skills also covers structure: a name of letters, numbers and hyphens, and a description that starts with \"Use when\" and names only the triggering conditions, since agents may follow a description that summarizes the workflow instead of reading the skill. It sets word targets for skills that load often. Bundled files include Anthropic's skill authoring best practices, a guide to testing skills with subagents, and a script that renders a skill's flowcharts to SVG."],"useItWhen":["You are writing your own skill and want proof that it changes how the agent behaves.","Your skill exists but the agent skips it or finds a way around its rules.","You want a skill description that makes the agent load the skill at the right time.","You are editing an existing skill and want to check it still works."],"notFor":["You want graded benchmarks, a review page and description tuning: use Skill Creator from Anthropic."],"install":[{"label":"In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace","command":"/plugin install superpowers@claude-plugins-official"},{"label":"Or install only this skill with the skills CLI","command":"npx skills add obra/superpowers --skill writing-skills"}],"sourceUpdated":"2026-09-19","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"},{"url":"https://cookiesforai.app/skills/anthropics/xlsx","owner":"anthropics","slug":"xlsx","name":"XLSX","skillName":"xlsx","title":"xlsx skill: Excel files with working formulas (Claude only)","summary":"XLSX is Anthropic's proprietary agent skill for creating, editing and cleaning spreadsheets with live formulas checked for errors; its license limits it to Claude and Anthropic's own products.","category":"documents-and-office","source":"anthropics/skills","path":"skills/xlsx","license":"Proprietary: use only within Anthropic's services, no copying or redistribution","licenseNote":"The LICENSE.txt in the xlsx folder (read 2026-10-01) puts use under your agreement with Anthropic and forbids extracting the files from Anthropic's services, keeping copies outside them, copying them, making derivative works and redistributing them. Anthropic's README calls its four document skills source-available, not open source. You may read the files on GitHub, but you may not copy them into Codex, Cursor or any other agent, or publish a changed version.","worksWith":["Claude Code","Claude.ai","Claude API"],"whatItDoes":["The XLSX skill teaches Claude to work with spreadsheet files: .xlsx, .xlsm, .xltx, .csv and .tsv. It loads when a spreadsheet is the main input or output, such as adding columns or formulas, building a chart, cleaning messy data or converting between table formats. It does not load when the result is a Word document, an HTML report, a script or a Google Sheets integration.","With the XLSX skill, Claude writes formulas instead of pasting computed results, so the sheet updates when inputs change, and it documents every assumption and hardcoded number. A bundled script recalculates the workbook with LibreOffice and names the cells with errors, and Claude fixes them all before delivering. It keeps to functions LibreOffice can check, avoiding XLOOKUP and FILTER. When editing your file, it matches your conventions and leaves existing formulas alone. Financial models get standard colors and number formats.","XLSX is one of Anthropic's four document skills, next to Word, PDF and PowerPoint. Unlike most skills in Anthropic's repository it is not open source: the license allows use only within Anthropic's own services and forbids copying, changing or redistributing it. That is why this page lists only Anthropic's install routes."],"useItWhen":["You need a financial model or budget that recalculates when you change the inputs.","You have a messy CSV export with bad headers or broken rows and want a clean spreadsheet.","You want to add formulas, columns or a chart to an existing Excel file without breaking what is there.","You need to convert data between CSV, TSV and Excel formats."],"notFor":["You work in Codex, Cursor, Gemini CLI or another agent that is not Anthropic's: the xlsx license bars copying the skill there.","You want a Word report or a Google Sheets integration: the skill says it does not load for those; Anthropic's DOCX skill covers Word."],"install":[{"label":"In Claude Code, add Anthropic's skills marketplace","command":"/plugin marketplace add anthropics/skills"},{"label":"Then install the document-skills plugin, which holds the Word, PDF, PowerPoint and Excel skills","command":"/plugin install document-skills@anthropic-agent-skills"}],"installNote":"In Claude.ai, Anthropic says its example skills are already available on paid plans: switch them on under Customize > Skills.","faq":[{"q":"Can I use Anthropic's xlsx skill in Codex, Cursor or ChatGPT?","a":"No. Its LICENSE.txt (read 2026-10-01) forbids keeping copies outside Anthropic's services, copying and redistribution, so it is meant for Claude.ai, Claude Code through Anthropic's plugin, and the Claude API. The open skills CLI does list it, but installing it into another agent copies the files, which the license forbids."},{"q":"Why does the xlsx skill avoid XLOOKUP?","a":"It recalculates every workbook with LibreOffice to find formula errors, and in files it writes LibreOffice cannot evaluate XLOOKUP, XMATCH, SORT, FILTER, UNIQUE or SEQUENCE. So it uses INDEX and MATCH for lookups and sorts or filters in Python before writing cells."}],"sourceUpdated":"2026-07-17","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-02T01:58:30+02:00"}]}