Skills
101,323 packaged capabilities and playbooks for agents. Install any of them with one command.
All
This skill automates sending WeChat messages on macOS via AppleScript. Use when the user needs to "send a WeChat message", "message a WeChat friend", "automate WeChat sending", "post in a WeChat group", or mentions "wechat message" or "WeChat automation".
End-to-end pipeline for writing and publishing an article to WeChat Official Account draft box. Use when user says "帮我写一篇新文章", "写文章发公众号", "配图排版发草稿箱", "直接发公众号", "配图发布", "write and publish", or any combination of writing + illustrating + posting to WeChat without manual confirmation steps.
Helps engineering managers build a functional PM–EM partnership and become more product-oriented — produces a 3-pattern PM–EM dynamic model (PM-Led / Engineering-Led / Fake Balanced), a true partnership definition, the "It's Important to Me" card for navigating disagreements, tips for getting engineers closer to customers and usage data (launch vs. landing, session recordings, customer conversations), and an explicit responsibilities exercise. Use when the user says "PM relationship," "PM drives me crazy," "product manager," "roadmap disagreement," "PM overrides me," "PM–EM partnership," "I want to be more product-oriented," or "how do I get engineers closer to customers." Do NOT use for general roadmap prioritization (use `roadmap-planning`) or when urgency is being manufactured by a PM (use `managing-urgency`).
Generate persistent, explorable 3D worlds/environments with Marble by World Labs — from text, a single image, multiple images, video, or a 360 panorama, plus coarse-structure blocking with Chisel. Use this skill when a task needs a navigable 3D scene, VR/immersive backdrop, virtual-production environment, previz set, game-blockout, or web/engine-ready Gaussian splat or mesh, and you must choose an input mode, a Marble model, an export format, and a plan/API route, then review world quality. Covers the Marble web app and the World API, editing/expansion, output formats and what each supports downstream, import into Unity/Unreal/Blender/Houdini/web, pricing/credits, capability limits, and rights/licensing. Not for flat text-to-image, text-to-video clips, or real-time on-the-fly world models (Marble outputs are downloadable, static 3D environments).
Use when deploying Zeabur templates or common services/databases via CLI. Use when user says "deploy template", "install template", or asks to deploy any well-known service, database, or self-hosted app. Common triggers include MongoDB, PostgreSQL, MySQL, Redis, MinIO, SurrealDB, PocketBase, WordPress, Ghost, Halo, n8n, Logto, Umami, Uptime Kuma, Vaultwarden, Prometheus, RSSHub, Miniflux, TTRSS, Memos, AFFiNE, Linkding, Slash, Bytebase, SQL Chat, ApiCat, OpenClaw, and any other open-source service the user wants to deploy to Zeabur.
Expert knowledge of academic writing standards for peer-reviewed papers, including citation integrity, style compliance, clarity, and scientific writing best practices. Use when reviewing or editing academic manuscripts, papers, or research documentation.
Diagnoses misbehaving ADK agents by inspecting sessions, events, tool calls, and the exact request that reached the model. Covers the `adk run` CLI and the `adk web` dev server with its session, trace, and debug HTTP endpoints. Use when an agent returns the wrong answer, ignores a tool or swallows a tool error, hangs, loops, emits raw JSON instead of calling tools, is not discovered by `adk web`, when a sub-agent cannot see the parent conversation, or when you need the LLM request/response, token counts, or logs for a run. Don't use for how ADK is designed internally (use `adk-architecture`), for building a new agent or workflow (use `adk-agent-builder`), for environment or dependency setup failures (use `adk-setup`), or for lint and style nits (use `adk-style`).
Write, review, and maintain architecture decision records with clear context, alternatives, consequences, confirmation links, and lifecycle governance. Use when a consequential technical decision or its enforceable architectural constraint must remain understandable. Do not use for system-wide evolutionary architecture design, general observability, or operating a named test or CI tool.
Manage Cloud Firewall (CFW) Internet Firewall public IP protection switches on Alibaba Cloud. Use this Skill when enabling or disabling firewall protection for public IPs, querying asset protection status, batch protecting ECS/EIP/NAT/SLB resources, toggling all public IPs at once, or configuring auto-protection for new assets. Supports multi-account management via member-uid. 管理阿里云云防火墙(CFW)互联网防火墙的公网IP防护开关。 当用户需要开启云防火墙、管理互联网防火墙开关、查看公网IP是否受保护、 批量开启ECS/EIP/NAT/SLB等资源的防火墙、配置新资产自动防护时使用此Skill。 支持多账号管理(member-uid)。
Master Archon MCP for strategic project management, task tracking, and knowledge base operations. The strategic layer (WHAT/WHEN) that coordinates with Skills (HOW). Use when managing projects, tracking tasks, querying knowledge bases, or implementing the Archon+Skills two-layer architecture.
Lyrics-driven generation of complete songs (including vocals), and also supports strong-rhythm production-grade instrumental music. Triggered only when the user explicitly asks to generate a finished song/audio and provides or agrees to supply style and lyric parameters; ordinary music discussion, lyric writing, and casual singing requests do not trigger it.
Audits the security posture of a CockroachDB cluster (Cloud or self-hosted) across network, authentication, authorization, encryption, audit logging, and backup dimensions. Use when assessing cluster security readiness, preparing for compliance reviews, or investigating security configuration gaps.
Provider-independent production workflow for AI avatar spokesperson videos, including presenter briefs, consent and likeness rights, disclosure and platform labeling, casting, localization, performance direction, brand fit, claims review, lip-sync/audio QA, accessibility, approvals, and delivery packages. Use when creating synthetic presenter videos, corporate explainers, training modules, sales/support avatars, localization spokespeople, product announcements, compliance videos, or social avatar clips.
Facilitate structured idea exploration and product/design specification. Use when a user wants to talk through an idea, refine it via iterative questions, and converge on a clear design/spec (and later an implementation plan), especially after inspecting the current project state.
Extract brand identity from a website URL — voice, colors, typography, imagery, values, and target audience — into a structured brand-profile.json. The profile feeds downstream skills (seo-content-writer, email-composer, frontend-design, pro-deck-builder, cross-platform-audit) for brand-consistent output. Use when the user says brand DNA, brand profile, extract brand, analyze brand, brand voice, brand identity, brand colors, or brand style guide.
Multimodal video understanding: generate descriptions, summaries, and tags for videos, and answer questions based on video content. Trigger when the user wants to understand video content, generate descriptions, or ask questions about a video.
Use Cartesia Sonic and related Cartesia voice APIs for production speech: text-to-speech, realtime WebSocket TTS, voice selection, instant and professional voice cloning, pronunciation/language/emotion controls, voice localization, voice changer, pricing/concurrency planning, privacy/security review, and QA for narration, ads, localization, dubbing, avatars, and interactive voice agents.
Identify at-risk customer accounts by analyzing usage patterns, engagement signals, and support history to generate churn risk scores and intervention recommendations. Use when the user requests churn analysis or provides relevant inputs for this workflow.
Stealth Chromium browser automation that bypasses bot detection. Use this skill whenever the user wants to scrape, screenshot, PDF, or automate a site that blocks normal headless browsers — Cloudflare Turnstile, reCAPTCHA v3, FingerprintJS, ShieldSquare, BrowserScan, "Access Denied" responses, suspicious 403/429s on plain `curl`, or any site where Playwright/Puppeteer get caught. Also use when the user mentions CloakBrowser, stealth Chromium, anti-bot bypass, "headless detection", "real Chrome fingerprint", or wants a one-off "what does this page actually render" check. Prefer this skill over raw `curl`/`fetch` for any site behind a CDN or bot-detection layer.
Enforce repository coding standards for Swift 6.2 strict concurrency, actor isolation, @Observable models, SWIFT_APPROACHABLE_CONCURRENCY, @concurrent functions, and modern Swift APIs across visionOS app code. Use when reviewing, writing, or migrating Swift code in this plugin's scope.
ComfyUI node output types - NodeOutput, UI outputs, PreviewImage, PreviewMask, SavedImages, PreviewAudio, PreviewText, PreviewVideo, Saved3DModels. Use when returning results from nodes, displaying previews, or saving output files.
The judgment layer SSOT for data visualization — load it before drawing any chart, dashboard, or board, not after. It governs "whether it should be drawn this way": what conclusion the chart wants the reader to reach, mean or median, whether aggregation is allowed, stacked or line, whether a dual axis is acceptable, color scheme and number of colors, whether the same entity keeps the same color and position across charts, and how many columns a table should keep. Medium-agnostic: it applies to HTML report pages, React/Vue dashboards, native PPT charts, matplotlib/plotly/ECharts/D3, and charts rendered to PNG alike. Use before drawing charts, or when reviewing existing charts ("what's wrong with these charts", "why are the legends inconsistent"). Not for report page layout/components/interaction (use report-with-html:report-with-html) or the implementation and rendering pitfalls of native PPT charts (use deck-creator:deck-creator).
When the user asks to draw any of the following and deliver it to Feishu — architecture diagram, flowchart, swimlane diagram, sequence diagram, org chart, state machine, funnel, Gantt chart, milestone, matrix quadrant, link architecture, hand-drawn-style architecture, complex business swimlane diagram, ER diagram, mind map — or provides a Feishu document URL and asks to draw inside it, or says "sync to Feishu / write to the Feishu whiteboard", this skill must be used. Beyond the 17 built-in templates above, this skill also provides a freeform-drawing mode: when the user asks for a painting, hand-drawn illustration, mobile app product prototype, map, game screen, Sankey diagram, or reproduction of an arbitrary reference image — anything the templates cannot cover — or explicitly says "freeform drawing / freestyle / no templates", this skill must likewise be used (the output is still an editable Feishu whiteboard). The generation path strictly follows the eight steps Normalize → Select → Plan → Layout → Render → StaticCheck → VQA → Deliver; hardcoding templates, hardcoding coordinates, and inventing business content the user did not mention are forbidden. All styles are read from assets/style-tokens/ (except for native code diagrams and freeform-drawing mode). Before delivery, the quality gate (executed per route) and dual-reviewer visual review must be passed (any score below 9 or any blocker means failure).
Use this skill when the user measures control-plane insertion throughput of DOCA Flow with the doca_flow_perf binary on x86 host or Arm cores—selecting a configs/ JSON policy, weighing the DPDK-versus-DOCA backend, running the smoke iteration then the eval loop, interpreting CPU cycles plus num_pushed / num_failed, or capturing the four-tuple (DOCA release, firmware, JSON policy, worker config) that makes an insertion figure defensible. Also load it for implied asks like "5-tuple hairpin insert figure", "flow-perf result mismatches release notes", "DPDK-vs-DOCA comparison", or "install variance too high". Refuse and route to the proper owner for live-app optimization (doca-flow-tune), wire throughput or latency, or library-internal pipe semantics.
Close out acceptance testing. Finds the open docs/acceptances/ATTENTION.<feature>.md candidate files that start-acceptance left, compares each candidate with docs/acceptances/ATTENTION.md, and for every one not already there asks the user, one at a time, whether it is systematic and settles the exact wording with them before it is moved; then retires the feature file. Runs the project's full test suite, lint and typecheck once first, since the acceptance session skipped them. Use when the user says acceptance is finished, asks to promote or review the acceptance lessons, or invokes /end-acceptance [feature]. Never promotes a lesson in its own wording.
Explore the dataset before designing any model. Place `data_analysis/data_analysis.py` and do not execute it. Run `materialize.py` once, write `data_analysis.md` and JOURNAL § Data understanding. Never design the model, edit `src/<pkg>/`, or modify raw data files. TRIGGER when the user asks to explore, profile, or understand the data; triage sent them here (`status.data_analysis` missing); a data source changed; they want to refresh a recorded EDA; or leakage or research arises on a recorded EDA. STOP when `status.setup.pending` is non-empty (load `setup-ml-project`), when there is no data (send to triage), the request is not raw-data exploration, or EDA is recorded with no refresh and no methodology concern. Do not invent a package root or a root `JOURNAL.md` while setup is pending. HOW TO USE: G-TABULAR via `add-python-package`, infer or ask the targets, load `plot-ml-figure` if installed, then ask the five-option continuation board, including Close.
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.
