Skills
98,639 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Memory forensics and anti-memory-forensics methodology. Understand from a blue team perspective how memory forensics uncovers malicious behavior (the Volatility3 analysis workflow), and from a red team perspective how to evade memory detection (process hiding, memory encryption, trace removal). Use when analyzing a memory dump or designing anti-forensics strategies
Meta Ads (Facebook/Instagram) management skill for B2B campaigns. Covers audience strategy, campaign structure, creative testing, and lead form optimization specifically for B2B SaaS on Meta's platform. MANDATORY TRIGGERS: Meta Ads, Facebook Ads, Instagram Ads, Facebook campaign, Meta campaign, Facebook advertising, Meta advertising, Facebook lead gen, Meta lead forms, Facebook retargeting, Meta remarketing, Facebook lookalike, Meta audience, Advantage Plus, Meta B2B, Facebook B2B
Discover DTC / e-commerce brands running video ads on Meta (Facebook/Instagram) via the ColdIQ marketplace Meta Ads endpoint, to expand Target Account Lists. Use when building net-new DTC/ e-commerce/subscription prospect lists, finding brands spending on Meta video ads, discovering companies in a vertical (baby, jewelry, wine, supplements), or finding companies with active creative teams. Triggers on "Meta ads", "Facebook Ad Library", "find DTC brands", "scrape ads", "ecommerce prospect list", "brands running video ads", "ad library search". Do NOT use for LinkedIn ads (use the linkedin-ad-library endpoint), firmographic database search (see apollo-search / coldiq-search-enrich), or contact enrichment (see contact-enrichment).
Apply rigorous, mathematically-grounded program construction and verification. Derive correct programs from formal pre/post specifications using axiomatic semantics, structural induction, recursive design with bounding functions, algorithm immersion, and iterative derivation with loop invariants. Language-agnostic. Use this skill whenever the user wants to implement a function correctly, reason about loops or recursion, write formal specifications, derive tests from postconditions, prove termination, design data structures algebraically, or ensure program correctness by construction — even if they don't use terms like "formal methods" or "specification".
MKD (`mkd` — Make Decision) — a local, browser-based decision-deck CLI the AI drives to collect structured human feedback: decisions with justified options, questions, report reactions and answerable tables, one screen at a time (Slack Catch-Up style). WHEN to use: the AI has MORE THAN ~3 decision points or a long multi-section report/plan that needs the user's verdict point by point. It beats `AskUserQuestion` (capped at ~4 questions × ~4 options, terminal-bound, no rich free-text, cannot show reference content) and beats an inline prose questionnaire (unanchored replies the AI must guess-map back). HOW it works: the AI writes a spec JSON of `items` (`decision` | `question` | `report` | `table`), runs `bun <skill-dir>/cli/index.ts <specPath>` (no install, no binary), and the CLI renders a self-contained deck page under `~/.mkd/`, opens the browser and EXITS immediately — the user answers at their own pace (progress rail, arrow keys, skip = decide later, live stats, localStorage persistence) and pastes the copied Result JSON back into the chat, which the AI treats as the execution contract. Inside Orca (multi-agent IDE) the default changes: the CLI serves the deck, opens it in a browser tab bound to the current worktree and BLOCKS until the user presses Send, then prints the Result JSON on stdout, so the AI must run it in the background (`--copy` restores the copy-paste flow). Add `--wait` outside Orca only when the answer is needed the same turn (blocking loopback server, Result JSON on stdout). In Claude Code with the `Artifact` tool, pass `--artifact`: the CLI renders a claude.ai Artifact page, the agent publishes it with `capabilities: {db: {}, comments: {}}` and arms a watch, and the user's Send to Claude wakes the session with the Result in the artifact's db (no localhost, no copy-paste); other harnesses keep the flows above. HARD RULE: every decision option MUST carry a written justification (value + cost), and the recommended option's justification must state WHY it is recommended — the validator rejects options without one. Triggers on: `mkd`, `make decision`, `decision deck`, `catch-up`, `catchup`, `decision form`, `feedback UI`, `ask me point by point`, `let me answer in the browser`, `review this plan point by point`, `react to each section`, `more than three questions`, `decisiones de auditoría`, `tomar decisiones`, `mkd artifact`, `answer in a claude.ai artifact`. Use this skill even when the user does not say `mkd` — if the task is to collect granular decisions or anchored feedback on many points, this is the right tool. Do NOT use for: a trivial 1-2 option pick or a single yes/no (use `AskUserQuestion`), or any non-interactive / CI / one-shot output where there is no human at a browser.
Searches and retrieves MLflow documentation from the official docs site. Use when the user asks about MLflow features, APIs, integrations (LangGraph, LangChain, OpenAI, etc.), tracing, tracking, or requests to look up MLflow documentation. Triggers on "how do I use MLflow with X", "find MLflow docs for Y", "MLflow API for Z".
Use CodexBar CLI local cost usage to summarize per-model usage for Codex or Claude, including the current (most recent) model or a full model breakdown. Trigger when asked for model-level usage/cost data from codexbar, or when you need a scriptable per-model summary from codexbar cost JSON.
Interact with Moltbook — the social network for AI agents. Post content, reply to discussions, browse feeds, upvote/downvote, join submolt communities, follow agents, search semantically, and track engagement. Use when the user wants to engage with Moltbook, check their feed, post, reply, or manage their agent presence.
Restore a prior business-state snapshot saved by /money-save, so you can continue from exactly where the last session left off. Loads the most recent snapshot for the current project (auto-detected from the working directory), or a specific one by index. Use when starting a new Claude Code conversation and the user wants to continue prior work. Triggered by: 'continue from last time', 'where did we leave off', 'pick up from before', 'resume', '接着上次', '续上', '之前的结论', '上次诊断到哪了'.
Review a business plan from the perspective of the solo founder who has to actually execute it. Asks the brutal operational questions: Can I build this with my time budget? Can I keep it running? Will my cashflow survive? Outputs a verdict — SHIPPABLE NOW / SHIPPABLE WITH DESCOPE / NEEDS HIRE / WRONG STACK. Use when a plan looks great on paper and you want to know whether a solo operator can actually execute it. Triggered by: 'operator review', 'can I actually do this solo', 'execution reality check', '操盘视角', '我能不能干得动'.
Review a business plan from the perspective of a hostile devil's advocate. Asks: what kills this in month 6? What's everyone politely not mentioning? Outputs a verdict — EXISTENTIAL RISK / SOLVABLE RISKS / LOW-RISK / WRONG QUESTION. Use after the investor/customer/operator reviews are done, as the final gauntlet before committing real time and money. Triggered by: 'skeptic review', 'devil's advocate', 'what would kill this', 'red team this', '泼冷水', '杀手在哪'.
Expand an image-generation request into a detailed MOSS-Image caption or spatial layout, and compile a local JSON draft into a ready-to-paste prompt. Use for prompt enhancement, image composition and text placement; this skill does not generate images.
Make a launch film for an app, mobile (inside an iPhone) or desktop (a window on a laptop's desktop, with a real pointer; web apps in a browser window), with its real UI rebuilt or captured in HTML, or a story told in panels with the brand's character and an end card. The film is cut on the beat to an original track made for it (or the user's), in a chosen visual style, with an optional voiceover in a local voice the user picks (Chatterbox or Kokoro), checked frame by frame and rendered to MP4, a muted loop, a poster and a single self-contained HTML. It also makes or cuts a track to the bar and the frame, writes and mixes a voiceover, and checks and installs the tools it needs after asking. Use when the user wants a launch video, promo, teaser or product film of their iOS, Android, macOS, Windows, Linux, Electron, Tauri or web app, wants to change such a film's story, style, pacing, device, music or voiceover, wants music made or a track cut to a film's beat or narration mixed over it, or asks what the tool needs to run.
Automate maintaining a Motoko mops package — upgrade dependencies, fix compiler warnings, run tests and benchmarks, fix breakages, update CHANGELOG, review doc strings and docs, run the formatter, and prepare a patch release on a new git branch.
Complete guide to MUI V7 for React + TypeScript. Teaches V6→V7 migration, advanced responsive design, professional theming, component patterns, and automatic detection of old syntax. Use when creating dashboards, modern interfaces, or working with Material-UI/MUI.
Panel of experts — consults multiple agents from the ecosystem in parallel for multi-perspective analysis of any topic. Activates personas, experts and technical agents simultaneously, each with its own unique perspective, and consolidates them into a final decision-oriented synthesis.
When the user wants to optimize multiple conflicting objectives, find Pareto-optimal solutions, or balance trade-offs between cost, service, quality, and sustainability. Also use when the user mentions "multi-objective," "Pareto optimization," "NSGA-II," "trade-off analysis," "scalarization," "weighted objectives," "goal programming," or "multiple criteria optimization." For single objective, see optimization-modeling.
Using the dbt-materialize adapter to manage Materialize streaming pipelines with dbt. Covers materializations (source, source_table, view, materialized_view, sink, table, seed), profile configuration, index creation, blue/green deployments (deploy_init, deploy_promote, deploy_cleanup), cluster management, strict mode, and testing. Use this skill whenever the user asks about writing dbt models for Materialize, configuring dbt profiles for Materialize, running dbt against Materialize, blue/green or zero-downtime deployments with dbt, creating sources or sinks in dbt, troubleshooting dbt-materialize issues, or migrating existing dbt projects to Materialize. Also trigger when the user mentions dbt-materialize, materialized_view materialization, deploy_init, deploy_promote, deploy_await, strict_mode, refresh_interval, retain_history, or partition_by in a dbt context.
Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learning transformers. By Andrej Karpathy. Perfect for understanding GPT architecture from scratch. Train on Shakespeare (CPU) or OpenWebText (multi-GPU).
Create, revise, audit or polish submission-grade manuscript figures for Nature-family and other high-impact journals using Python or R. Covers multi-panel scientific plots, matplotlib/seaborn, ggplot2/patchwork/ComplexHeatmap and journal-ready SVG/PDF/TIFF exports. Also use for OpenRouter GPT Image 2 graphical abstracts, concept or mechanism schematics when the user explicitly requests AI generation. Establish the figure conclusion, evidence logic and exports before execution; honor the chosen or saved plotting backend. Trigger on paper figures, scientific/academic plotting, manuscript data visualization, 论文配图、学术写作配图、科研绘图、科研作图、画图、作图、出图、论文图表、可视化、论文示意图、机制示意图、图形摘要. Not for dashboards or Illustrator/Figma-first infographics.
NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.
Comprehensive toolkit for creating, analyzing, and visualizing complex networks and graphs in Python. Use for network/graph data structures, analyzing relationships between entities, computing graph algorithms (shortest paths, centrality, clustering), detecting communities, generating synthetic networks, or visualizing network topologies. Applicable to social networks, biological networks, transportation systems, citation networks, and any domain involving pairwise relationships.
Provides expert guidance on Next.js rendering behavior, including static vs dynamic rendering, streaming, Suspense boundaries, cache components, and the prospective render system. Use when working with Next.js SSG, ISR, streaming, server components, or debugging rendering modes.
Generates NG-ZORRO code and guidance for Angular projects using `ng-zorro-antd`. Covers installation, LESS-variable theming (default, dark, compact, aliyun presets), the experimental dynamic CSS-variable theme, component usage across all categories (general, layout, navigation, data entry, data display, feedback), ReactiveForms / Signal Forms integration via `nz-form`, internationalisation (`provideNzI18n`), accessibility, and the module → standalone migration. Use when the user mentions NG-ZORRO, ng-zorro, ng-zorro-antd, Ant Design for Angular, `nz-*` components, `Nz*Module`, OR when the Angular project's `package.json` contains `ng-zorro-antd`. Pairs with `angular-developer` for Angular fundamentals.
Answer questions from Nietzsche's perspective — the Übermensch philosophy, will to power, revaluation of values and eternal recurrence. Use when the user explicitly asks for this figure's perspective, mentions "nietzsche-perspective", or wants to apply his core method to analyze a problem; do not trigger automatically for general questions.
Builds Databricks data products from live web data, end to end: discovers the right Nimble web-data agents, scrapes into Delta tables, and produces an AI/BI dashboard and/or a deployed Databricks App — a table → dashboard → app workflow, for production data products or quick demos. Use whenever a request pairs live or scraped web data WITH a Databricks destination — e.g. "scrape Amazon/Walmart prices into a Delta table and build a dashboard", "load Zillow/Instagram/Maps/search results into Databricks and build a dashboard or app", "showcase Nimble + Databricks to a prospect". Prefer it over nimble-web-expert or competitor-intel when the data lands in Databricks. Do NOT use for one-off web fetches or CSV exports with no Databricks destination — use nimble-web-expert instead. Do NOT use for competitor or company research briefings — use competitor-intel or company-deep-dive instead. Do NOT use for generic Databricks work with no Nimble/web-data angle — use the official databricks-* skills instead.
Port scanning and service identification with nmap. Use when precise port scanning, service version detection, OS fingerprinting, or NSE script vulnerability scanning is needed against a target. nmap is the most classic network scanner, supports multiple scan modes such as SYN/TCP/UDP/ACK, and ships with 600+ NSE scripts. This skill should be used in any scenario involving port scanning, service identification, vulnerability script scanning, or OS fingerprinting. Slower than naabu, but far more capable than naabu
Use when needing to reorder, split, drop, or amend git commits that are not the top commit, without interactive editor access. Covers programmatic rebase via GIT_SEQUENCE_EDITOR, commit splitting with automated hunk selection, and metadata changes (author, message, dates) on any commit in a range.
Nuxt 3 full-stack Vue framework. Covers file-based routing, data fetching, server routes, and auto-imports. Use when building Nuxt applications. USE WHEN: user mentions "Nuxt", "Nuxt 3", asks about "useFetch", "Nuxt server routes", "Nitro", "Nuxt auto-imports", "Nuxt composables", "Nuxt middleware", "SSR with Vue" DO NOT USE FOR: Vue without Nuxt - use `frontend-vue` instead; Next.js - use `nextjs-app-router` instead; Nuxt 2 - consult KB for migration; Astro - use `astro` instead
Use when turning a folder of research PDFs into Obsidian notes, even if Obsidian is not named. Writes one templated literature note per paper and extracts cross-linked atomic concept notes, never overwriting existing ones. A .bib or Zotero library is /lit-sync.
This skill should be used when the user asks to maintain an Obsidian knowledge base for a research project, import an existing research repository into Obsidian, keep project memory or daily notes synchronized, summarize project context into durable notes, or update experiments, results, papers, writing, and plans in an Obsidian vault without requiring MCP.
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.