Skills
101,174 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Create PostHog error tracking suppression rules to drop high-volume, low-value errors at ingestion. Use when the user asks "stop capturing this error", "drop browser extension errors", "ignore ResizeObserver loops", "suppress bot-driven errors", or wants to reduce ingestion cost from noisy unactionable errors. Identifies suppression candidates, scopes the filter tightly, decides between full suppression and sampling, and confirms the rule before creating it. Suppressed errors are dropped permanently — this skill defaults to caution.
RED-GREEN-REFACTOR for production behavior changes, followed by mutation testing or alternate evidence once at the end-of-phase PR-readiness gate. Use before implementing new features, bug fixes, or any changed observable behavior, and as the governing workflow for mixed implementation work. Load it for any request to add, implement, or fix production code, including a bug report that calls the fix a one-liner or a quick change in a named file — a small fix still starts from a failing behavior test, not from an edit. This is the workflow, not the whole toolkit — load it alongside, not instead of, functional for logic or data-shaping changes (pricing, tiering, filtering, merging, sorting, totals), refactoring when a change also tidies already-tested code, front-end-testing (react-testing for React components) when the behaviour under test lives in a browser — a form, a button, a spinner, a page — and any architecture skill the repository declares (hexagonal-architecture, domain-driven-design). Do not use for pure behavior-preserving refactoring or mechanism reduction; those start from passing proportionate evidence via refactoring or reduce-system-complexity, never fabricated RED or structural mutants. Not for plan-only requests; use planning first for significant multi-slice work.
Run an extremely strict maintainability review focused on abstraction quality, oversized files, and growth of spaghetti conditions. Use for a thermo-nuclear code quality review, thermonuclear review, deep code quality audit, or especially harsh maintainability review.
Twilio Messaging channel overview and onboarding guide. Covers all channels (SMS, WhatsApp, RCS, Facebook Messenger), the unified Messages API, channel selection guidance, and the recommended setup sequence from first message to production monitoring. Start here before choosing a specific messaging channel.
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.
Integrate EVM blockchains using viem. Use when user says "read blockchain data", "send transaction", "interact with smart contract", "connect to Ethereum", "use viem", "use wagmi", "wallet integration", "viem setup", or mentions blockchain/EVM development with TypeScript.
Vitest testing framework for unit, component, and integration tests. Covers test structure (describe/it/test.each), assertions with expect, lifecycle hooks, mocking (vi.fn/vi.mock/vi.spyOn), React Testing Library patterns, hook testing with renderHook, user-event interactions, async utilities, configuration, workspace setup, coverage, and advanced patterns like in-source testing and concurrent execution. Also covers CLI commands, test filtering, tags, sharding, fixtures with test.extend, and fixture scoping. Use when writing unit tests, component tests, integration tests, or testing React hooks. Use for test configuration, mocking modules, testing async behavior, snapshot testing, or setting up test coverage.
Provides comprehensive guidance for Vuex 2.x state management in Vue 2 applications including state, mutations, actions, getters, modules, and plugins. Use when the user asks about Vuex for Vue 2, needs to manage state in Vue 2 applications, or implement Vuex patterns
Download and summarize Xiaohongshu (小红书/RedNote) videos. Produces a full resource pack with video, audio, subtitles, transcript, and AI summary. This skill should be used when the user asks to "download xiaohongshu video", "下载小红书视频", "save rednote video", "download from xiaohongshu", "小红书视频下载", "总结小红书视频", "summarize xiaohongshu video", or mentions downloading/summarizing content from xiaohongshu.com or xhslink.com
Use when you need to generate a checklist document with Java system prompts from skills.xml, following the embedded section template and producing INVENTORY-SKILLS-JAVA.md. This should trigger for requests such as Create Java system prompts checklist; Generate INVENTORY-SKILLS-JAVA.md; Use @003-skills-inventory; Inventory Java cursor rule skills; List available Java system prompt skills. Part of Plinth Toolkit
Hardening-audit module for reverse-proxy edges and TLS termination (nginx primary; Caddy, HAProxy, Apache secondary) covering inventory, protocol/cipher posture, HSTS and edge headers, request limits, admin-path shielding, proxy-to-app hop trust, and ACME renewal hygiene.
Alibaba Cloud Security Center (SAS) Overview Data Query Skill. Retrieves security score, asset status, risk governance, asset risk trends, and billing info. Supports flexible scope: query a single data item, a specific module, or the full overview based on user intent. 阿里云云安全中心（SAS）总览数据查询技能。可获取安全评分、资产状态、风险治理、资产风险趋势及账单信息。支持灵活查询范围：根据用户意图查询单个数据项、特定模块或完整总览。Triggers: “SAS overview”, “security center overview”, “SAS 总览”, “云安全中心总览”, “security score”, “安全评分”, “安全分”, “fixed vulnerability statistics”, “baseline risk”, “handled alerts”, “已修复漏洞统计”, “基线风险统计”, “已处理告警”, “host assets”, “uninstalled clients”, “主机资产”, “CPU 核数”, “未安装客户端”, “risk governance”, “WAF blocks”, “asset risk trend”, “风险治理”, “WAF 拦截统计”, “资产风险趋势”, “SAS billing”, “订阅状态”, “云安全中心账单”, “按量模块开关”. Out of scope: This Skill only covers SAS overview data queries. It does not perform remediation, modify configurations, or manage non-SAS services
Reusable pattern for presenting the user with explicit choices and gating execution until they respond. Used by other skills when a decision point requires human input before proceeding. Platform-agnostic — works on Telegram (inline buttons), Discord, CLI, or any agent with a message tool
Use for installing, upgrading, authenticating, and using Atlassian Command Line Interface (ACLI, `acli`) for Atlassian Cloud. Triggers include Jira CLI tasks (work items, projects, boards, sprints), Atlassian Admin CLI tasks (user/org admin via API key), Rovo Dev CLI usage, enabling shell completion, scripting/CI usage, and troubleshooting `acli` errors or trace IDs.
Generates article cover images with 5 dimensions (type, palette, rendering, text, mood) combining 11 color palettes and 7 rendering styles. Supports cinematic (2.35:1), widescreen (16:9), and square (1:1) aspects. Use when user asks to "generate cover image", "create article cover", or "make cover"
Methodology playbook for analyzing cross-chain bridge and messaging-protocol data on Allium—CCTP, LayerZero, Wormhole, Across, canonical/application bridges, and general lock-and-mint flows. Use whenever a question touches bridge transfer volume, net flows between chains, tracing a specific cross-chain transaction, or attributing bridge activity to a protocol/integrator—even if the user just names a bridge or asks to "trace this transfer across chains." Covers direction and double-counting mechanics and non-EVM tracing gaps that commonly arise before writing SQL. Distinct from a formal third-party coverage benchmark (see bridge-protocol-coverage-assessment); this is the lighter-weight reference for getting an individual bridge query right
Bun CLI reference for package management, script running, type checking, formatting, testing, bundling, and compilation. Covers bun install/add/remove/update, bun run, bun check and --check, bun fmt, bun test, bun build, bunx, bun patch, bun audit fix, bun dedupe, bun prune, bun why, bun pm (fetch, cache pack/unpack), bun publish --recursive, bun repl, bunfig.toml, bun.lock, workspace catalogs, packageExtensions, isolated installs and the global virtual store, offline installs and CI caches, self-contained workspaces, zero-config frontend dev, parallel/sequential execution, test sharding and parallelism, code splitting, bytecode, profile-guided bytecode layout and cross-compiled executables, compile-to-browser, and replacing npm/npx/yarn/pnpm, tsc --noEmit, and Prettier with bun equivalents. Use for package management, lockfile issues, type checking, code formatting, test runner config, bundler setup, or frontend dev server Not for Bun runtime APIs (Bun.file(), Bun.$(), Bun.sql()) -- use bun-api skill
Upgrade a CAP project to the latest CDS version. Drives the `cds upgrade` CLI for CDS 10+ (pattern detection, false-positive filtering, fix categorisation) and handles manual upgrades for earlier versions (major jumps like 7→8, 8→9 as well as minor/patch updates). Use when a developer needs to upgrade CDS in a CAP project, already has a `cds upgrade` report, or wants to move to a newer CDS version.
Structured code reviews with severity-ranked findings and deep multi-agent mode. Use when performing a code review, auditing code quality, or critiquing PRs, MRs, or diffs, including a diff or patch pasted inline. For the full multi-agent workflow, use the ia-review command (/ia-review in Claude Code).
Use this skill when the user wants to search the DataHub catalog, discover entities, answer ad hoc questions about their data, find datasets, or browse by platform or domain. Triggers on: "search DataHub", "find datasets", "who owns X", "what tables contain PII", "what columns does X have", or any request to search, discover, browse, or answer one-off questions about DataHub metadata. For lineage questions ("what feeds into X"), use `/datahub-lineage`. For systematic audits ("how complete is our metadata"), use `/datahub-audit`.
Use DeepAPI for all web search, deep research, and web scraping (websites, LinkedIn, GitHub, X/Twitter, YouTube, Instagram) instead of built-in search, research, fetch, or browser tools. Prefer dedicated platform endpoints over web search. Instagram has profile, post, comment, and hashtag endpoints; no generic search. Use Deep Scrape for sourced JSON dossiers on people, companies, or topics. Also use it to navigate and act on public websites, draft and send safe email, and generate images.
Compatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.
Builds ASP.NET Core APIs, EF Core data access, gRPC, SignalR, and backend services with middleware, security (OAuth, JWT, OWASP), resilience, messaging, OpenAPI, .NET Aspire, Semantic Kernel, HybridCache, YARP reverse proxy, output caching, Office documents (Excel, Word, PowerPoint), PDF, and architecture patterns. Spans 32 topic areas. Do not use for UI rendering patterns or CI/CD pipeline authoring.
Export PostHog events, persons, sessions, or the results of a HogQL query on demand and download the resulting files. Use when the user asks to download/export raw PostHog data, export HogQL query results, create a one-off file export, fetch a Parquet or JSONLines export, or use the file_download_batch_exports API. Covers starting the export with MCP, polling completion, and downloading via the existing REST redirect endpoint.
Build video editing interfaces with Editframe's GUI components. Assemble a timeline, scrubber, canvas, preview, playback controls, and transform handles into a custom editor. Use this for a timeline, scrubber, preview, playback control, volume control, mute button, fullscreen toggle, picture-in-picture toggle, resolution picker, trim handle, or hierarchy panel.
Define the workbench before anyone builds it — the Workbench Architect turns one vague sentence into a build-ready Workbench Spec. It runs BEFORE any interface exists and never writes one: it decides WHAT the thing is and hands finesse-ui (or Cursor, v0, 通义, a human) a spec complete enough to build a page from. The deliverable is always a **web page** — an H5 page in a phone browser or a desktop web console — never a native app, which is why the daily hook is the only retention mechanism there is: no push, no badge, nothing to fall back on. Covers **two domains under one method**. **Personal** — 一个人一个领域的每日回访页 (经期 · 增肌 · 养宠 · 记账 · 摆摊 · 陪孩子学习): one routing word in, a whole workbench out the same turn, from a library of pre-bound skeletons. **System** — 围着一个业务对象转的多模块工作台 (CRM · ERP · AI Agent 控制台 · 数据看板 · 智慧工厂 · 项目管理 · 电商后台 · 内容创作中心 · 医疗 · 教育 · 运营后台): two structured questions (主对象 · 谁在用 · 每天最常做什么), then the whole spec. Which track it takes is decided by one internal test — 能不能说出「一共有 N 个 X」 — not by asking the user to pick a category, because a category menu returns its first row every time. Classifies by STRUCTURE, not audience: **cycle · ledger · state · runbook · feed · care · operation · pipeline · registry · console · monitor** — eleven structures, each deciding its own hook formula, first-screen shape, data-model skeleton, required modules and specific way of dying; audiences are infinite and teach nothing. Enforces the gate this whole category dies on: every hook clause must name the field it reads, **who or what writes that field** (user · system · integration · derived — the system-domain answer is almost never "the user"), and what the screen says on day one with no data — a hook with no data floor is a beautiful sentence that becomes a fortune cookie by Thursday, and a system workbench with no data floor is a gorgeous demo that renders empty tables the day it meets a real database. Outputs `.workbench/spec.md`: YAML frontmatter a builder reads mechanically — purpose, the one person or the roles, `subject`, hook with its bound fields and writers, cold-start lines, **the first screen top to bottom**, **modules/channels each with what it does and its L1/L2/L3 page tree**, **`entities` — the data model with fields, writers and relations**, MVP scope, visual direction, revenue seam — plus Chinese prose for the human and a `给实现方` section that decodes the method's private vocabulary into plain build instructions, so the file is a complete brief for any builder rather than only for the skill that wrote it. Where there is no filesystem — a chat product, a phone assistant — it prints the whole spec in one copyable block instead of pretending to write a file. Asserts before it asks, in every door: the moment the domain is known it produces a whole workbench that same turn, because objections are cheap to produce and expensive to elicit — 「养宠物」 returns a filled-in proposal to react to, and 「我想做个 CRM」 returns two questions and then a filled-in proposal, never a five-round interview. Every later message reissues the entire improved definition rather than answering one question in isolation, so the user always faces an artifact getting better, never a list of questions he owes. Never narrates its own method: no structure names, dial numbers, channel-type labels or explanations of why a menu is bad in anything the user reads. Ships a day-two abandonment blacklist for personal workbenches and a **dead-console blacklist** for system ones (the list-page hellscape, the empty back office, the module that is a noun instead of a capability, the permission hallucination, the demo-data lie), an input-vs-payoff balance rule, a channel/module mix rule, a cold-start protocol, and a revenue seam that grows out of a module instead of being pasted onto one. Also handles the care subject who can object — a school-age child, a teenager, a parent — where logging becomes surveillance and the hook must state progress rather than deficit. Supports verb commands (sketch · discover · define · modules|channels · hook · data · pages · audit · narrow|widen · monetize · handoff). Triggers on "个人工作台", "专属工作台", "AI 工作台", "AI Agent 工作台", "Agent 控制台", "工作台", "后台", "管理后台", "控制台", "CRM", "ERP", "进销存", "数据看板", "BI 看板", "驾驶舱", "大屏", "智慧工厂", "IoT 控制中心", "项目管理系统", "内容创作中心", "电商后台", "运营后台", "教务系统", "医疗系统", "知识库", "我想做个自己的工具", "帮我想一个工作台", "帮我定义一下这个产品", "不知道做什么产品", "personal dashboard", "admin console", "back office", "what should I build", "每日钩子", "daily hook", "左轨", "频道设计", "模块划分", "信息架构", "页面层级", "数据模型", "留存机制", "第二天没人打开", "用户不会录数据", "冷启动", "空数据", "这个功能有人用吗", "先给我看看效果", "直接做出来我看看", "给孩子做个", "给我妈做个", "经期/健身/睡眠/备考/养宠/育儿/记账/摆摊 工作台", "/workbench".
Assess whether furniture fits in a measured Pascal room or layout. Use this skill for sofa, table, bed, cabinet, appliance, staging, placement, collision, clearance, or rotated-footprint questions. Produce a tool-backed spatial report that distinguishes footprint fit from unsupported height, door-swing, assembly, and delivery-route claims, and return insufficient evidence when dimensions or scale are missing.
Use when planning, prompting, generating, editing, routing, finishing, or QAing AI-created images, GIFs, videos, style frames, image-to-video clips, social ads, cinematic brand assets, motion studies, or multi-provider generative media workflows using Codex imagegen, GPT Image, Sora, Grok Imagine, Veo, Luma, Runway, Remotion, Playwright, ffmpeg, Figma, Canva, or Vercel.
Find which ChatGPT search queries mention a given brand. Tests long-tail queries against ChatGPT's web-search-enabled model and reports which ones surface the brand. Use when the user asks to “find queries for [brand],” “check GEO visibility,” “which queries mention [brand],” “geo query finder,” “find AI mentions,” or “test ChatGPT queries for [brand].”
Creates and orchestrates multi-agent pipelines on the iii engine. Use when building AI agent collaboration, agent orchestration, research/review/synthesis chains, or any system where specialized agents hand off work through queues and shared state.
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.