Skills
96,981 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Invoked by /tidy-code. Reviews source code for structural quality violations (hidden dependencies, god functions, silent failures, deep nesting) in any language. Produces a read-only findings report with concrete refactoring suggestions. Skips dot-directories (`.git`, `.venv`, IDE/agent scaffolding) and build/vendor output (`node_modules`, `vendor`, `dist`, `build`, `coverage`, `target`). Does not auto-activate on natural-language requests.
The Tong Jincheng perspective: viewing interpersonal relationships, social dynamics and personal growth through the thinking framework of the "deeply affectionate grandmaster" and livestream emotional-content creator. Platform compatibility: GitHub Copilot, Claude, Qwen Code, OpenCode, Cursor. Source material: transcripts of 9 first-hand videos (livestreams/dating vlogs/flirting analysis), about 200,000 characters. Core models: 5. Decision heuristics: 9. Trigger words: 「童锦程」「深情祖师爷」「用童锦程的方式」「从童锦程视角」「景辰怎么看」. Limitations: the material is mainly emotional/interpersonal content, with insufficient data on business/entrepreneurial thinking, so use with caution in purely business decision scenarios. Research date: April 2026.
Detects and removes duplicate transactions across overlapping bank, credit-card, and brokerage statement imports using a stable composite key (account_id, date ±1d, amount_cents, description_normalized). Emits a list of new transactions to commit, a list of suppressed duplicates with their reasons, and a list of suspicious near-duplicates that need human review. Use when ingesting financial statements that may overlap prior drops, merging multiple export sources for the same account, or when user mentions duplicate transactions, deduping a transaction file, or reconciling overlapping statements.
One-off transcription of local audio or video files to text or subtitle files using Transloadit via the official `@transloadit/node` CLI. Use when the user wants speech in local media converted to `.txt`, `.json`, `.srt`, or `.webvtt`; prefer the `speech transcribe` intent with Replicate.
都爆鸭 · general trending topic selection (direct pull without keywords + smart matching with personal IP). First pull down the batch of currently hottest items across the web (without keywords), then smart-match them against the user's personal IP positioning (domain/persona/angle/audience) to pick the hot topics this IP can ride and produce an actionable topic list. Triggers: 选题、爆款选题、追热点、找选题、全网热榜、综合热点、热搜、趋势、选题信号、蹭热点、借势选题.
Design and audit sales-assisted trial conversion for SaaS and developer tools covering activation signals, PQL scoring, usage intent, sales routing, in-product nudges, demo offers, procurement readiness, upgrade prompts, trial extension, nurture, handoff, experiment guardrails, and trust-preserving outreach. Use when free trials or freemium users should convert without spam or dark patterns.
Plan trips with assumptions, not questions. Trip Planner extracts flight constraints from natural conversation, fills in the gaps with smart defaults, and returns structured flight recommendations with direct booking links. Assume first. Show your work. Book the flight. Just say "trip-planner" to get going.
Navigate Taiwan healthcare regulations including NHI system, medical device classification, drug registration, telemedicine rules, and health data protection. Use this skill when the user is building a health tech product for Taiwan, needs to understand NHI, evaluate medical device regulatory pathways, or assess telemedicine compliance — even if they say 'sell a medical device in Taiwan', 'how does NHI work', 'telemedicine regulations', or 'health data privacy in Taiwan'.
Building MCP (Model Context Protocol) servers in TypeScript on Cloudflare Workers with the official @modelcontextprotocol/sdk: registering tools, resources, and prompts, Hono HTTP routing, StreamableHTTPServerTransport setup, authentication (API keys, OAuth, Zero Trust), Cloudflare D1/KV/R2/Vectorize integration, wrangler deployment, and testing. Prevents common failures including export syntax errors, Zod schema validation failures, memory leaks from unclosed transports, CORS misconfiguration, and missing rate limiting. Use when exposing APIs, tools, or data to LLMs as a stateless serverless MCP server on the edge. Not for Python MCP servers (see FastMCP), stateful WebSocket agents (Cloudflare Agents SDK), or local stdio CLI tools.
TypeScript type system, generics, utility types, strict mode, and ecosystem patterns. Use for: typescript, ts, type, generic, utility type, Partial, Pick, Omit, Record, Exclude, Extract, ReturnType, Parameters, keyof, typeof, infer, mapped type, conditional type, template literal type, discriminated union, type guard, type assertion, type narrowing, tsconfig, strict mode, declaration file, zod, valibot, typescript 7, tsgo, native compiler.
Advanced TypeScript type system specialist for complex generics, conditional types, template literals, type inference, performance optimization, and type-level programming. Use for intricate type system challenges, recursive types, brand types, utility type authoring, and type performance issues. Includes comprehensive coverage of 18 advanced type system error patterns.
Use when contributing to TYPO3 Core — Forge issues, Gerrit patches, cherry-picks, CI debugging — or when working on **git.typo3.org**, the t3o site GitLab (`services/t3o-sites/**`: extensions.typo3.org/ter, common/t3olayout, anubis): merge requests, issues, work items, labels, pipelines. Core `main` is v15-in-progress (PHP 8.4/8.5); v14.3 LTS fixes land on branch `14.3` via cherry-pick. Triggers on: forge.typo3.org, review.typo3.org, git.typo3.org, t3o-sites, core patch, Gerrit review, TYPO3 Core contribution, TER website, t3olayout.
Rút hệ thiết kế ĐANG CÓ của một dự án — từ code (token màu, thang kích thước, component) hoặc từ URL site đang chạy thành MỘT trang HTML tham chiếu kiểu Figma UI kit: foundations + component theo nhóm + màn mẫu mobile/desktop, đổi palette/sáng-tối/shape, copy token/HTML/hex. Giá trị lấy bằng cách CHẠY hàm token thật (hoặc đọc style thật bằng Playwright) chứ không gõ tay; có vòng KHÁM PHÁ tự đi theo link + bấm mở phần ẩn, bỏ trùng theo chữ ký, dừng khi bão hoà, và cổng PHỦ bắt mọi pattern tìm được phải có mặt trong kit (không dừng ở phần xương). Mẫu output chuẩn: references/ui-kit-reference.html. KHÁC hallmark (dựng/đổi thiết kế mới): skill này chỉ CHÉP LẠI hệ đang có, không sáng tác. Gọi khi user nói 'ui kit', 'figma ui kit bằng html', 'bóc design system ra', 'trang tham chiếu component', 'design token page', 'component gallery', 'style guide từ code', 'bóc trang này ra html kit', 'làm giàu kit', hoặc /ui-kit-from-code.
Compress the ENTIRE UI/UX of a frontend (React/Next.js/Vue + Tailwind) into ONE self-contained HTML file, running offline, with hardcoded placeholder data, under a size threshold (default 5MB): port each screen to vanilla JS with Tailwind classes copied verbatim from the JSX, compile Tailwind with the project's OWN config, embed fonts + icon subset + logos, with a screen/state picker panel and a phone frame for mobile mode, verified with Playwright. Use when the user says "snapshot UI", "gói toàn bộ giao diện vào 1 file html", "bản demo giao diện offline", "mockup tĩnh có dữ liệu mẫu", "nén UIUX dự án", "gửi giao diện cho người không chạy được app", or invokes /ui-snapshot. DIFFERENT from hallmark/redesign (NEW design) and playwright-verify (only captures/measures a running app) — ui-snapshot REPRODUCES the existing interface, no backend/login needed.
UI/UX design intelligence. 50 styles, 21 palettes, 50 font pairings, 20 charts, 9 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient. Integrations: shadcn/ui MCP for component search and examples
Secure Uniswap v4 hook development assistant. Activates when: - Working with Uniswap v4 hooks, PoolManager, or hook contracts - Keywords: uniswap, v4, hook, hooks, beforeSwap, afterSwap, PoolManager, IHooks, BaseHook - Creating, reviewing, or auditing hook contracts - Implementing custom AMM logic, dynamic fees, or swap modifications - User says "uniswap hooks" or invokes "/uniswap-v4-hooks"
Customize Univer color themes, branded palettes, dark mode, and theme-aware plugin styles across Sheets, Docs, Slides, Bases, Boards, and PDFs. Use when setting an initial or runtime Theme, consuming ThemeService or --univer-* CSS variables, keeping custom UI or canvas code theme-aware, registering a separate Univer Pro chart theme, or migrating theme code from Univer 0.25.0 to current 1.0 source.
Upgrades Helm chart dependencies (PostgreSQL, Vault) in the Chainloop project, including vendorized charts, container images, and CI/CD workflows. Use when the user mentions upgrading Helm charts, Bitnami dependencies, PostgreSQL chart, or Vault chart. CRITICAL - Major version upgrades are FORBIDDEN and must be escalated.
Design and audit usage-metering accuracy for usage-based pricing, credits, quotas, overages, AI/token billing, API calls, seats, storage, event ingestion, idempotency, late events, aggregation, reconciliation, invoices, customer visibility, disputes, corrections, and revenue leakage. Use when billable usage must be trusted by customers and finance.
Design and improve product user onboarding (first-time user experience) to drive activation and early retention. Produces an Onboarding & Activation Pack (aha moment spec, first 30 seconds + first mile plan, onboarding journey map, experiment backlog, measurement plan). Use for Growth teams.
Use before claiming work is complete, fixed, or passing, before committing or creating PRs. Requires inspecting verification evidence that matches each claim before making any success claims, whether freshly run or still-applicable verified evidence whose relevant code, tests, inputs, and environment match. Evidence before assertions, always.
SaaS application template using the Partnership model. AI proposes complete vision based on proven SaaS patterns (auth, dashboard, core features, settings), then you provide context. Perfect for web applications requiring user accounts and core functionality.
Produce rich, finished video content with React Remotion by orchestrating the repository's media-generation skills (nanobanana for images, veo for video clips, lyria for BGM, gemini-tts for narration) and composing them on a data-driven Remotion timeline. Follows an approval-gated workflow: first return a video composition plan for the user to approve, then generate assets, compose, run a multimodal self-review loop, and deliver only when the result meets the quality bar. Use when the user wants to "create a video", "make a promo / explainer / social clip", or combine images, video, music, and voiceover into one polished video.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.
