Skills
100,928 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Design onboarding and activation flows that help users reach value quickly. Use when the user mentions onboarding, getting started, first-time users, activation, aha moments, or new user flows—not when the work is only the empty-state surface itself.
Use when a user asks to automatically generate a CLI command for a website. Takes a URL and optional goal, runs the full verified generation pipeline (explore, synthesize, cascade, verify), and returns a structured outcome. This is the primary entry point for "帮我生成 xxx.com 的 cli".
Turn a long video into short vertical clips ranked by a virality score, using OpenClip. Give it a podcast, webinar, interview, stream, or YouTube URL and get back the best moments with hooks, titles, social copy, and downloadable clip URLs. Use when asked to "clip this video", "find the viral moments", "make shorts from this", "cut this podcast into clips", "what are the best bits of this", "make TikToks from this stream", or "give me the top moments".
Pull still frames out of a video as thumbnail images with OpenClip, then compress, resize, crop, or reformat them to jpg, png, or webp. FREE with just an OpenClip account, no subscription. Use when asked for “a thumbnail from this video”, “a still frame”, “a poster image”, “preview frames”, “a contact sheet”, “grab a screenshot from the video”, or “resize this image”.
Auth check for the OpenFinance backend — confirms an API key is available before any other OpenFinance skill runs. Use FIRST whenever the user is about to call any /agent/* route (Polymarket, Hyperliquid, Relay), is hitting 401/412, or hasn't traded yet in this session. Triggers on "how do I get started", "API key is required", "Invalid API key", "401/412 from /agent/*", "set up OpenFinance", or any first call into a trading skill. Resolves the key from `OPENFINANCE_API_KEY` (or equivalent env / user-supplied value), confirms the format (`open_…`), verifies via GET /agent/wallets, and otherwise points the user to https://openfinance.tech to issue one.
Code review for the current change, before it is pushed. One command runs the security, secret, dependency and lint scanners that fit the changed files, a separate reviewer that checks every scanner finding and is given every changed line, and prints a short receipt with the path of the full report. Use before every git push, when asked for a code review, a security scan, or to review changes, a diff or a pull request, and when a push was blocked or warned by OpenQodex.
Implement decision models through OpenRouter's Decisions API, which returns probabilities instead of generated text: pick the primitive per judgment, build minimal state, write the questions, pin a model from the live catalog, gate in code, and probe thresholds on real inputs. Use when replacing a prompt-and-parse LLM call, a keyword or similarity heuristic, a human review queue, or a misfiring router, filter, or classifier with a decision model, when locating something in an image with a Box question, when moving TypeSafe SDK or System One code onto OpenRouter, when asked to find or implement decision-model use cases in an app or agent, and whenever decision models, Jev, TypeSafe, or System One come up. Covers routing, classification, moderation, guardrails, grounding checks, scoring, ranking, dedupe, escalation and approval gates, bounded extraction, and thresholds. Not for chat completions, text generation, or general OpenRouter setup.
Optimizes audio memory, CPU cost, and playback quality through import settings and mixer configuration. Use when asked to reduce audio memory or Audio Mixer CPU cost, change clip Load Types, sample rates or codecs, or force 3D audio to mono.
Optimizes TextMeshPro rendering, memory, and font setup, including font asset stacks, fallback atlases, SDF quality, and worldspace text. Use when the user asks about TextMeshPro font assets, CJK font fallbacks, or TMP memory, build size, quality or performance, not about translating text.
Use external Windows Python and the originpro package to automate Origin/OriginPro plotting, control figure hierarchy from runtime to export, map arbitrary data to Worksheet/Matrix/Layer/Plot objects, apply publication-quality templates and Nature-style rules, batch export figures, and save editable OPJU projects. For publication-style figures, follow the embedded palette/style and hierarchy-control rules.
Groundwater time series analysis and modelling using transfer function noise models. Use when the agent needs to: (1) Analyze groundwater level time series, (2) Model well responses to precipitation/pumping, (3) Calibrate aquifer parameters from head data, (4) Forecast or hindcast groundwater levels, (5) Decompose hydrological signals into components, (6) Compare response functions, (7) Perform model diagnostics and uncertainty analysis.
Set up the Paystack API client, environment variables, and TypeScript helpers for server-side payment integration. Use this skill whenever starting a new Paystack integration, configuring API keys, creating a reusable fetch wrapper for the Paystack REST API, or setting up the foundation for any Paystack feature. Also use when you see errors related to missing PAYSTACK_SECRET_KEY, authentication failures, or need to understand Paystack's base URL, response format, pagination, currencies, or amount subunit conversion.
A skill that analyses new feature requirements and ui-proto to generate a confirmed Spec that team-dev/team-e2e can execute without additional planning questions. Triggered by the keywords "Spec 만들어줘", "신규 기능 기획", "요구사항 정리". If ui-proto exists, extract the Spec from the screen plan with proto=<path>. If existing feature analysis is needed, use peach-doc-feature.
Generate and edit images from the CLI using picture-it. Use this skill whenever the user asks to create, edit, or manipulate images — blog headers, social cards, hero images, product comparisons, YouTube thumbnails, movie posters, magazine covers, Instagram edits, background removal, or any visual content. Also trigger when the user mentions picture-it by name, wants to composite images, apply color grading, add text to images, remove or replace backgrounds, crop/resize photos, or needs any kind of image generation or photo editing from the terminal. This skill covers multi-pass AI image editing workflows that chain composable operations together.
This skill should be used when the user asks to "integrate Plaid", "connect a bank account", "sync bank transactions", "add Plaid Link to my app", or "implement financial data access". Integrates the Plaid API for bank account connections and transaction syncing in TypeScript/Bun applications.
Use when operating Ultralytics Platform from a terminal or script with the ul cloud CLI (ul cloud RESOURCE OPERATION key=value, PyPI package ultralytics-platform) — listing, creating, updating, cloning, or moving projects, datasets, and models; inspecting dataset images, statistics, metadata, duplicates, and similar images; starting and monitoring cloud training runs; per-image model validation analysis; downloading weights; exports, deployments, uploads, trash and restore, storage integrations, and account usage, storage, and billing; and building Agents (YOLO, LLM, gate, and alert workflows). For the Platform web UI or local yolo commands, see yolo; for choosing what to train and how to improve it, see yolo-training and yolo-tuning.
Docs-PM decision support for the "Write the docs" authoring process — makes audience, stage, and cross-cutting scope calls during the Frame and Shape stages (including cross-repo product lookup via universe when accessible, else the public OSS path), and helps decide when a docs question needs to self-serve vs. escalate to a docs PM. Use when framing a new docs page or launch, deciding what product stage or audience a feature targets, judging whether a docs question needs PM sign-off, or confirming which product repos a launch spans.
A comprehensive guide for interacting with PostgreSQL 16 databases. Use this skill for constructing standard and advanced SQL queries, optimizing performance, debugging errors, managing schema objects, and introspecting database structure.
Use before emailing a PostHog customer, replying to one, or joining a call, and for any request to research a PostHog account. Fires on "/posthog-customer-deep-dive" and on natural language like "deep dive on X", "look up this customer", "help me reply to Y", "prep me for my call with Z", from an email address, domain, account name, or Vitally account id. Researches the account across Vitally and project 2 usage queries, then drafts an email (first touch, follow-up, or reply) or a call-prep brief, every recommendation carrying a live docs link.
Stateless review of an open GitHub PR. Standards hard-require /taste KISS + named principles (including honest names, trust the server, types tell the truth, and naming alignment) and /architecture, plus a correctness hunt and PR extras (body vs diff, migrations, breaking API). Triage historical finding threads, draft one-topic comments, post only after one publish decision.
Fetch tasks from PREQSTATION, execute work, and push updates/results back to PREQSTATION APIs. Use when asked to run a PREQ task, fetch todo tasks, or update execution status. Prefer the remote MCP endpoint at /mcp with OAuth login. Shell helper mode still uses PREQSTATION_API_URL and PREQSTATION_TOKEN. Tasks carry an `engine` field (`claude-code` | `codex` | `gemini-cli`) indicating which AI agent should execute them.
Transform AI agents from task-followers into proactive partners that anticipate needs and continuously improve. Now with WAL Protocol, Working Buffer for context survival, Compaction Recovery, and battle-tested security patterns. Part of the Hal Stack
When the user wants to create or update their product marketing context document. Also use when the user mentions 'product context,' 'marketing context,' 'set up context,' 'positioning,' or wants to avoid repeating foundational information across marketing tasks. Creates `.claude/product-marketing-context.md` that other marketing skills reference.
Workflows for investigating production issues in Honeycomb: context priming, broad query, BubbleUp, trace analysis, verification, and chaining results between steps to a root cause. Trigger phrases: "investigate production issue", "debug latency spike", "find root cause", "use BubbleUp", "analyze traces", "debug an outage", "why is my API slow", "errors are increasing", "health check", "SLO burning".
Access and ground work in a separate Project Brain knowledge repository through the authenticated GitHub CLI. Use at the start of substantive work in any Git repository to check its root for `.project-brain.json`; when the marker exists, using this skill and consulting the brain is mandatory before planning, implementation, diagnosis, review, documentation, or project answers. Also use when explicitly asked for Project Brain context, decisions, requirements, terminology, source-system records, or knowledge-graph output.
When the user wants to plan promotions, optimize trade spending, forecast promotional lift, or manage promotional supply chain. Also use when the user mentions "trade promotions," "promotional forecasting," "TPM," "TPO," "promotional uplift," "feature and display," "markdown optimization," or "promotional calendar." For pricing optimization, see markdown-optimization. For demand forecasting, see demand-forecasting.
Use when the user needs to find, prioritize, or engage B2B prospects — defining an ICP, tiering accounts, building or enriching lists, reading buying signals, designing multi-channel cadences, cold calling, LinkedIn outreach, routing inbound leads, or diagnosing weak pipeline generation. Also for live prospecting coaching. Not for closing (use sales-skill), copy (use copywriter-skill), or CAC analysis (use channel-expert).
Proves or disproves claims, or the test plan of a PR, by trying to falsify each and rating the strongest evidence reached (PROVEN, PARTIAL, DISPROVEN, UNPROVEN). Changes nothing. Use when the user says "prove it", or when asked to prove, verify, or fact-check claims or a PR test plan with evidence.
Linear operators for large-scale inverse problems with matrix-free representations. Use when the agent needs to: (1) Define linear operators for forward/adjoint operations, (2) Solve inverse problems (deconvolution, imaging, tomography), (3) Apply signal processing transforms (FFT, convolution, derivatives), (4) Compose operators for complex workflows, (5) Perform regularized inversion with smoothness or sparsity constraints, (6) Process seismic or image data at scale.
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.