Skills
98,067 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Deep learning framework (PyTorch Lightning / lightning package). Organize PyTorch code into LightningModules, configure Trainers for multi-GPU/TPU, implement data pipelines, callbacks, logging (W&B, TensorBoard, MLflow), distributed training (DDP, FSDP, DeepSpeed), for scalable neural network training.
Performs exact symbolic mathematics with SymPy for algebra, calculus, equation solving, symbolic linear algebra, physics, and lambdify or LaTeX code generation. Use when a task needs symbolic results, explicit assumptions, or exact arithmetic; use NumPy or SciPy for purely numerical workloads.
Searches current documentation, release notes, API references and public repository issues for libraries, frameworks, SDKs, APIs and CLI tools with `firecrawl developer`. Use before answering when the user names a version, is upgrading, pastes an error from a library or tool, or asks how to use, configure or migrate a specific library or API. Model knowledge stops at a training cutoff and misses new releases, renamed APIs, changed defaults and recent fixes, so when a version, option, error or API is unfamiliar or may have changed, look it up instead of guessing. Also checks whether a bug was fixed and in which release. Not for general programming concepts, code that uses no specific library, or refactoring code the user supplied.
Draft, revise, polish, compress, or adapt CCF manuscript prose for presentations. Use for writing, paper polishing, research documents for completing papers, full first drafts, experiments left blank, rewriting, paper compression, abstracts, sections, and slides. Draft content before fitting pages; preserve evidence and format. Assessment without rewriting belongs to ccf-paper-reviewer; rebuttals belong to ccf-rebuttal-writer.
Creates, analyzes, and visualizes complex networks and graphs in Python with NetworkX. Use when working with network/graph data structures, computing graph algorithms (shortest paths, centrality, clustering), detecting communities, generating synthetic networks (random, scale-free, small-world), reading/writing graph file formats, or drawing network topologies. Common applications include social, biological, transportation, and citation networks.
This skill should be used for time series machine learning tasks including classification, regression, clustering, forecasting, anomaly detection, segmentation, and similarity search. Use when working with temporal data, sequential patterns, or time-indexed observations requiring specialized algorithms beyond standard ML approaches. Particularly suited for univariate and multivariate time series analysis with scikit-learn compatible APIs.
Chat with any real person or fictional character in their own voice by automatically finding their speech online, extracting a clean reference sample, and generating audio replies. Also supports generating a matching voice from an uploaded image. Use when the user says "我想跟xxx聊天", "你来扮演xxx跟我说话", "让xxx给我讲讲这篇文章", "我想跟图片中的人说话", or similar.
Use this skill for generative video editing, text-to-video, image-referenced video generation, first-frame-to-video, first-and-last-frame transitions, and video extensions using Gemini Omni 1.1 Flash (gemini-omni-1.1-flash) via the official google-genai SDK. Includes workflows for pre-processing/optimizing high-resolution or long source videos with ffmpeg, stripping audio for full sound regeneration, and handling turn-by-turn video editing and parallel execution.
Pure text-to-image generation for all creative scenarios (no reference photo). Covers logo design, poster design, illustration, meme, game assets, social media content, 3D rendering, education, fashion, food, pet, wedding, holiday marketing, and artistic styles. Use when generating images from text descriptions without a reference photo (e.g. design a logo, create a poster, generate game art, make a meme, 3D render).
GPT Researcher is an autonomous deep research agent that conducts web and local research, producing detailed reports with citations. Use this skill when helping developers understand, extend, debug, or integrate with GPT Researcher - including adding features, understanding the architecture, working with the API, customizing research workflows, adding new retrievers, integrating MCP data sources, or troubleshooting research pipelines.
Design content structure, hierarchy, labelling, and the navigation model. Use when organising what exists. For the UI that exposes it use `navigation-patterns` (interaction-design); for user-generated grouping evidence use `card-sort-analysis` (design-research).
Corporate due-diligence workflow — resolve a brand or website to its registered legal entity, map group structure and beneficial ownership, profile officers and directors, enumerate the digital estate, and screen litigation, insolvency, procurement, sanctions, PEP and adverse media. Use when asked to check out, vet or research a company, verify a supplier or counterparty before signing or paying, or assess whether a business is real. Applies to vendor and third-party risk, KYC and KYB onboarding, M&A and investor diligence, procurement integrity, and shell-company assessment. Reference at useosint.com/skills/x-ray-a-company.
Create AI avatar and talking head videos via inference.sh CLI. Recommended: P-Video-Avatar (fastest, cheapest, built-in TTS). Also: OmniHuman, Fabric, PixVerse. Audio: Inworld TTS-2 (100+ languages, emotion steering for characters), ElevenLabs, Kokoro. Capabilities: audio-driven avatars, text-to-avatar, lipsync videos, talking head generation, virtual presenters, UGC content. Use for: AI presenters, explainer videos, virtual influencers, dubbing, marketing videos, UGC ads, gaming avatars, NPC dialogue. Triggers: ai avatar, talking head, lipsync, avatar video, virtual presenter, ai spokesperson, audio driven video, heygen alternative, synthesia alternative, talking avatar, lip sync, video avatar, ai presenter, digital human, ugc, ugc video, ugc ad, avatar ugc
Use only when the user explicitly requests a review or audit of backend code under `api/`. Supports pending-change, file-focused, and pasted-diff reviews. Do not use for implementation-only requests, diagnosis without review intent, frontend code, or backend code outside `api/`.
Create LinkedIn post graphics. Choose between an HTML/CSS structured graphic and an AI-generated infographic based on the post content. Use this skill whenever the user says “design a graphic,” “create a visual,” “make an image,” “graphic for my post,” “LinkedIn image,” or wants any visual content to pair with a LinkedIn post. Also use it when the user finishes writing a post and wants a matching graphic.
Author Prisma Next migrations — choose db update vs migration plan, edit the framework-rendered migration.ts (replace placeholder sentinels with dataTransform closures), recover from MIGRATION.HASH_MISMATCH or PN-MIG-2001 unfilled placeholder. Use for prisma migrate dev, prisma migrate deploy, prisma db push, db update, db update --dry-run, migration plan, migrate, migration new, migration show, db verify, db sign, data migration, this.dataTransform, dataTransform, placeholder, generated migration.ts, edit migration.ts, MIGRATION.HASH_MISMATCH, schema drift.
Required first preflight before every CCFA skill, including research, review, retrieval, experiments, visuals, and maintenance. Keep reasoning and communication direct, remove empty defensive framing, and preserve evidence and uncertainty. Also use for 去防御性 and 论文人性化. Apply prose edits only within the authorized task; specialist skills retain ownership.
Use when writing or reviewing n8n expressions (`{{...}}` syntax), `$json` / `$node` references, Luxon date code, or expression errors. Triggers on `{{}}`, `$json`, `$node`, `$input`, `DateTime`, `Luxon`, "expression error", "evaluating", "format date", "transform field", or any node-parameter assignment.
Hugging Face Transformers for loading Hub models, running pipeline inference, text generation, and Trainer fine-tuning on NLP, vision, audio, and multimodal tasks. Applies when working with AutoModel, pipelines, tokenizers, generation configs, or TrainingArguments within Transformers
Use when converting Java source files to idiomatic Kotlin, when user mentions "java to kotlin", "j2k", "convert java", "migrate java to kotlin", or when working with .java files that need to become .kt files. Handles framework-aware conversion for Spring, Lombok, Hibernate, Jackson, Micronaut, Quarkus, Dagger/Hilt, RxJava, JUnit, Guice, Retrofit, and Mockito.
Generates or edits images with AI models through the OpenRouter Image API (Gemini, Seedream, Recraft, GPT-Image, Riverflow). Use for photos, illustrations, artwork, concept art, visual assets, logos, and image editing or compositing from reference images. For flowcharts, circuits, pathways, and other technical diagrams, use the scientific-schematics skill instead.
Manages Zotero reference libraries using the pyzotero Python client: retrieves, creates, updates, and deletes items, collections, tags, and attachments via the Zotero Web API v3 or local API. Applies when working with Zotero libraries programmatically, managing bibliographic references, exporting citations, searching library contents, uploading PDF attachments, or building research automation workflows that integrate with Zotero.
Help users craft product positioning and messaging. Use when someone is launching a product, differentiating from competitors, writing marketing copy, struggling to explain what their product does, or working on value propositions and taglines.
Develop and optimize rough CCF research ideas into problems, insights, mechanisms, and evidence plans. Use for 优化idea, 具象化idea, 找方向, 方向探索, and rescue routes when development is the requested deliverable. Judging whether an idea is worthwhile, novel, or coherent belongs to ccf-idea-reviewer even without scores; manuscript edits belong to ccf-paper-writer.
INVOKE THIS SKILL when routing a LangGraph agent with a decision model (TypeSafe Jev, SemIf) instead of an LLM, or when auditing an existing agent for LLM calls that only produce a routing decision. Covers langchain-typesafe Noul/Choice/Score, reading answers correctly, threshold design, and LangSmith Gateway wiring.
Turn a newsletter edition into a 12-20 second promo clip for LinkedIn, Instagram or X that drives people to read and subscribe. Hook question, three proof frames from the edition, then the newsletter name and where to read it. Use when someone says "promo for my newsletter", "trailer for this edition", "get people to my Substack", "teaser video" or "promote this issue".
Use to extend an existing agent project with memory, app integration, VPC, multi-agent, migration, model, browser, code interpreter, payments, or resource removal. Triggers: "add memory", "remember across sessions", "call agent from app", "invoke agent from code", "agent auth", "streaming", "VPC", "VPC connectivity", "can't reach from VPC", "multi-agent", "A2A", "A2A auth", "orchestrator not delegating", "specialist not called", "migrate Bedrock Agent", "migration issue", "change model", "browser tool", "code interpreter", "delete agent", "tear down", "agentcore remove", "cross-account memory", "add payments capability to my agent", "wire payments plugin", "integrate x402 payments with the agent I'm building", "add MPP payments", "Machine Payments Protocol". External APIs via Gateway: use agents-connect. New project: use agents-get-started. CLI/dev-server errors: use agents-debug. Runtime x402/MPP payments: use agents-pay. Migration-specific Strands vs LangGraph routes here
Build a topical authority cluster from a seed keyword or topic. Use when the user asks about topic clusters, content silos, pillar pages, topical authority, content architecture, or how to structure content around a theme. For finding the right keywords, see find-keywords. For writing individual pieces, see brief.
Review manuscript claims, evidence, and writing without rewriting. Use for 文章审核, 审稿, 稿件有什么硬伤, 结论站得住吗, 投稿成熟度, writing review, and version comparison. Return a structured scientific or writing review; scores need not be explicitly requested. Concept-only judgment belongs to ccf-idea-reviewer even with a full PDF; requested prose edits belong to ccf-paper-writer.
APM - install, onboard, instrument, enable, set up, configure, traces, services, dependencies, performance analysis, Data Streams Monitoring (DSM), queue lag, pipeline latency. Use for any request involving Datadog APM setup, instrumentation (SSI, ddtrace, agent install), or analysis
Captures learnings, errors, corrections, and feature requests to enable continuous improvement. Use when: (1) User corrects Claude ('No, that's wrong...', 'Actually...'), (2) User requests a capability that doesn't exist, (3) Claude realizes its knowledge is outdated or incorrect, (4) A better approach is discovered for a recurring task, (5) Receiving a Handoff block from self-healing (a recurring verified heal with Recurrence-Count at least 3) to distill into a memory file or new skill. For ACTIVE runtime failures where the agent needs to apply and verify a fix mid-task, use `self-healing` instead (it files HEAL- entries with proof; self-improvement promotes accumulated patterns). Also review learnings before major tasks. For CI-only/headless learning capture, use self-improvement-ci.
Beisen applicant follow-up. Use to view an applicant's current progress, query the application list under a job, filter resumes pending action, view applicants in interview, follow up with applicants awaiting offers, verify applicants pending onboarding, and view candidate application details. Use when the user asks about candidates, applicants, applications, resume screening status, interview-stage lists, offer/onboarding status, or a specific applicant's details.
Use when building anything multi-step, anything that looks repeatable, anything the user mentions reusing, or any workflow with more than ~10 nodes. Triggers on “reuse”, “I do this in another workflow”, “extract”, “modular”, “shared logic”, “subworkflow”, multi-step builds, or any task that mentions logic the user has built before.
Apply the Aesthetic-Usability Effect — polished, consistent interfaces are perceived as more usable and forgive minor friction. Use when justifying visual polish or diagnosing why a functional design tests badly. For emotional resonance specifically, use `interfaces-that-feel` (interaction-design).
Grassroots-first campaign design for anyone being outspent — startups vs. incumbents, NGOs vs. corporate comms, movements vs. state-backed machines, solo brands vs. big-budget competitors. Ideates awareness, launch, fundraising, mobilization, community-build, counter-narrative, referral, founder-story, and coalition campaigns. Triggers on "campaign plan", "marketing strategy", "ad budget", "should I advertise", "paid vs organic", "launch plan", "grassroots", "low budget marketing", "NGO campaign", "outspent", "competitor has bigger budget", "how do I compete without money". Also trigger on any spend asymmetry, collapsing organic reach, rising CPAs, or a trust/credibility problem — even without the word "campaign". Nudge activation when the user debates buying ads, boosting posts, or hiring influencers; they are likely about to burn money on a channel that will not persuade.
Give the same motion brief to three AI models, keep each first try, render all three the same way and stack them into one comparison video with model names and logos. Makes a 16:9 side-by-side for a newsletter and a 1080 x 1920 stacked reel for Instagram and TikTok. Use when someone says "compare three models", "same prompt three AIs", "model showdown", "which AI is the best motion designer", "side by side comparison video" or wants a fair head-to-head of AI outputs.
Scales pandas, NumPy, and custom Python research workflows beyond memory or across clusters with Dask. Covers DataFrames, Arrays, Bags, Futures, chunking, schedulers, and distributed diagnostics. Use for partitioned file processing, scientific array computation, or parallel tasks whose memory and dependency structure require Dask.
Edit a vigiles .spec.ts to change a compiled instruction file (CLAUDE.md / AGENTS.md) — add, modify, or remove a rule, section, command, or file pointer (including a request to add a "key file" — the keyFiles field is deprecated). Use whenever you need to change a CLAUDE.md/AGENTS.md that carries a vigiles hash (edit the spec, never the artifact), including adding a new enforce()/check()/guidance() rule.
Design n8n AI agents the right way. Use when building or editing any @n8n/n8n-nodes-langchain.* AI node — an AI Agent, LLM chain, Text Classifier, or Information Extractor — and whenever the user mentions AI agents, LLM with tools, tool calling, $fromAI, system prompts, agent memory, sessionId, structured/JSON output, output parser, RAG, vector store, a chat assistant/bot, or human-in-the-loop review. Covers Agent-vs-chain-vs-classifier choice, the model/memory/tools/outputParser slots, tool names/descriptions as prompt, structured output with autoFix, memory, RAG, human review, and chat topologies.
Write Prisma Next queries for Postgres, SQLite, or Mongo — pick a lane (Postgres/SQLite `db.orm.<Model>` + `db.sql.<table>`; Mongo `db.orm.<root>` + `db.query.from(...)` pipeline builder), filter / project / sort / paginate, eager-load with `.include(...)`, Postgres/SQLite `db.transaction(...)`, Postgres/SQLite ORM `.aggregate(...)`, Mongo aggregations via query builder, namespace-aware accessors (`db.orm.<ns>.<Model>`, `db.sql.<ns>.<table>`). Triggers: query, where, match, select, project, orderBy, take, skip, include, lookup, first, all, count, aggregate, group, create, update, delete, upsert, returning, transaction, db.close, script teardown, variant, polymorphism, drizzle-style, kysely-style. Notes: `.all()` is a Thenable (just `await` it), iterators are single-use (`RUNTIME.ITERATOR_CONSUMED`), Postgres `count` is `number` while sum/avg/min/max are `number | null`, ranges use chained `.where()` or `and(...)` (no `.between(...)`).
Creates and reviews infographics with Nano Banana 2 via OpenRouter. Use for statistical summaries, timelines, comparisons, processes, and visual explanations with supplied data or optional Sonar research. Supports ten layouts, eight style presets, reference images, and accessible palette starting points.
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.