Skills
98,067 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Deeply grill an idea, decision, design, plan, or implied approach through evidence, failure modes, and user-owned decisions. Use discovery mode for raw ideas, clarification, "grill me", or an unfinished Product Brief; use audit mode for a formed target that needs a verdict or revision. Continue every reply in an active deep-grill session until its Product Brief or verdict is confirmed, without requiring the user to name the skill again.
Use this skill when the user's prompt requires (1) researching a topic across multiple sources, comparing options or alternatives, analyzing trends or history, understanding markets or industries, or reviewing literature or studies and (2) synthesizing that research into a comprehensive, narrative report. If you're planning to search the web or internal knowledge bases, consider using this skill. This skill coordinates research subagents, so use it only when you have a tool for spawning subagents (the Agent or Task tool); otherwise, research the question directly.
Export raw OpenTelemetry traces from an AI application to Confident AI's Observatory. TRIGGER when the user wants to send OpenTelemetry or OTLP traces/spans from an LLM app, agent, RAG pipeline, or chatbot to Confident AI; configure the Confident AI OTLP endpoint; set confident.span.* or confident.trace.* attributes; export AI-app traces without the deepeval Python package; wire an OTLPSpanExporter, OpenTelemetry Collector, or vendor-neutral OTel SDK to Confident AI; or pick the US vs EU OTLP endpoint. Language-agnostic: the mechanism is OTLP attribute keys plus an exporter endpoint. DO NOT TRIGGER for building DeepEval pytest eval suites, datasets, goldens, metrics, or deepeval test run (use the `deepeval` skill); for instrumenting with the DeepEval SDK's @observe decorator or framework integrations (use the `deepeval-tracing` skill); or for non-AI software such as web servers, CRUD backends, or infrastructure: the confident.* attributes describe AI components only.
Guidance for Microsoft Defender for Business (MDB) — the SMB-segment endpoint security product (≤300 employees), bundled with Microsoft 365 Business Premium and available as a standalone SKU. Covers what's included vs MDE Plan 1/2 (next-gen AV, EDR with simplified configuration, ASR, automated investigation and response, vulnerability management, web content filtering, attack surface reduction, mobile threat defense add-on), wizard-driven onboarding, server add-on for SMB servers, the simplified portal experience vs full Defender XDR, transition path to MDE P2 / E5 as the org grows, and integration with Microsoft 365 Lighthouse for MSP delivery. WHEN: Defender for Business, MDB, SMB endpoint security, M365 Business Premium security, MDB server add-on, simplified EDR small business, MSP Defender for Business, M365 Lighthouse Defender, SMB upgrade to MDE Plan 2. DO NOT USE for enterprise EDR (use defender-for-endpoint), Defender XDR cross-product investigation (use defender-xdr), or Defender for Servers / Cloud (use defender-for-servers).
Guidance for Microsoft Defender for Cloud — AI workload protection (AI-SPM and runtime threat detection for generative AI). Covers AI Security Posture Management (discovery of Azure OpenAI / Azure AI Foundry / Amazon Bedrock / Google Vertex AI resources, identification of grounding data exposure, model deployment posture), runtime threat detection on Azure OpenAI (prompt injection / jailbreak attempts, sensitive data leakage in prompts/responses, wallet abuse, credential leakage), integration with Azure AI Content Safety Prompt Shields, attack path analysis for AI workloads, alert investigation in Defender XDR, and pairing with Purview DSPM for AI (user side) and Azure AI Content Safety (model side). WHEN: Defender for Cloud AI, AI-SPM, AI workload protection, Azure OpenAI threat detection, prompt injection alert, jailbreak alert Azure OpenAI, AI wallet abuse, AI workload posture, Amazon Bedrock posture, Vertex AI posture, generative AI security Azure. DO NOT USE for end-user AI usage governance (use purview-ai-hub), content moderation API (use azure-ai-content-safety), or M365 Copilot rollout (use copilot-for-m365-readiness).
Extract clean markdown content from web pages using Defuddle CLI, removing clutter and navigation to save tokens. This skill MUST be used instead of WebFetch when the user provides a URL to read or analyze, for online documentation, articles, blog posts, or any standard web page. MUST NOT be used for content already served as Markdown; identify it by the response content type and actual body rather than the URL suffix, and read it directly with WebFetch or curl.
How to delegate work to another AI agent (Pi, Codex, Claude Code, Hermes) — picking the right agent, sending prompts to TUI agents, polling progress. Read BEFORE any `cmux send`/`tmux send-keys` to an agent, or whenever delegating, relaying, spawning, or orchestrating agent-to-agent work.
Master dispatcher for deobfuscation workflows. Loads the right combination of deobf and reverse-engineering skills based on target type — native binary, JavaScript, VM-protected, packed, or CTF. Use when you need comprehensive deobfuscation capabilities for any target.
Periodic dependency review for Node.js/pnpm projects — outdated package triage, security audit, update batching strategy (patch/minor/major), validation checklist. Run monthly or before major releases. Use when asked to review or update dependencies in a repo.
Extract design systems from reference UI images and generate implementation-ready UI design prompts. Use when users provide UI screenshots/mockups and want to create consistent designs, generate design systems, or build MVP UIs matching reference aesthetics.
Use for Desktop Commander MCP capabilities — persistent shells and REPLs, long-running processes, filesystem beyond the workspace, structured files (.xlsx, .docx, .pdf, images) and large local data files such as CSVs, ripgrep search at scale, SSH, or cross-turn state.
Fully autonomous DTF vault management on Solana. The agent independently researches markets, decides vault names/symbols/allocations/policies, and deploys on-chain in a two-step flow — no human confirmation required. The policy engine is the guardrail, not a human approval step. Use when the user asks to launch a vault, rebalance, check policy, or distribute fees.
Methodology for disk forensics and anti-disk forensics. Understand how blue teams recover deleted files from disk, extract timelines and analyze file system artifacts, and how red teams perform anti-forensics (secure deletion, timestamp tampering, trace removal). Use when you need to evade disk forensics or perform disk analysis
Use when working on BK-CI build scheduling, such as build machine selection, queue advancement, quota control, third-party build machine scheduling, and Docker/Kubernetes scheduling. Prefer this when the user wants to change task allocation strategies rather than Agent/Worker execution details.
Audit, organize, and redesign developer documentation information architecture, including content inventory, navigation, landing pages, sequences, hierarchies, webs, breadcrumbs, sidebars, metadata, redirects, migration, platform/tool selection, and maintainable IA decisions. Use when restructuring docs sites, improving findability, planning doc migration, evaluating docs tooling, or organizing large documentation sets.
Create optimized, secure multi-stage Dockerfiles for .NET Core and ASP.NET Core applications. Use when (1) creating a new Dockerfile for a .NET project, (2) containerizing a .NET application, (3) optimizing an existing .NET Dockerfile, (4) setting up Docker for .NET 6/7/8+ apps, or (5) user mentions .NET and Docker/container together.
When the user is doing development, writing code, writing docs, looking up APIs, or troubleshooting deployment or operations issues related to internal projects at dida (道旅科技), proactively search the Dragon private knowledge base (via the KnowledgePlugin__knowledge_base_search tool of the dragon-knowledges MCP) to obtain internal technical docs, architecture designs, API specs, field definitions, business processes, error codes, configuration and operations material. Be sure to trigger proactively in the following scenarios, without the user explicitly saying "use Dragon" or "search the knowledge base": the user refers to internal assets with "our", "the company's", "internal", "this project's", "production", etc.; the user mentions any dida internal system name (such as Beaver, Phoenix, LionKing, Shopping, EBooking, Wolf, Wolf 2.0, AngryBirds, CRM, Analysis, Whale, Admin, Marketing, Tiger, Bingo, Kangaroo, Finance, Content, Dolphin, Grafana, Penguin, Panda, AutoLine, FalconService, f-operating, f-supplier, f-analysis, f-order, f-policy, f-profit, f-monitor, EBookingBookingAPI, Wuying/无影, customer service system, API channel system, API supplier system, Supplier Service/SS, BDS); the user asks about business rules, product logic, approval flows, alert thresholds, or the meaning of error codes; the user asks about an API field, callback structure, or internal SDK usage and the field name does not look like an open source framework; the user asks about release/canary/deployment/configuration/monitoring tied to a specific internal environment; the user writes code, fixes bugs, writes comments or writes docs in the company's internal code repository; the user mentions the didainternal.com domain or a similar internal domain. When in doubt, lean toward searching — the cost of one internal search is far lower than the cost of "answering wrong or too generically". Use this skill aggressively whenever the user is doing internal dida company development work, even without explicit mentions of Dragon/knowledge base.
Build, review, debug, configure, migrate, teach, or plan Drizzle ORM 1.0 RC (not 0.x) with kit, seed, validators, and Effect drivers. Use for drizzle-orm@rc, drizzle-kit@rc, drizzle-seed@rc, defineRelations RQBv2, pgTable/mysqlTable/sqliteTable, generate migrate push pull studio, drizzle-orm/zod|valibot|typebox|arktype|effect-schema, drizzle-orm/effect-postgres and other effect-* drivers on Effect 4, kit --output json/--hints/drizzle-kit/cli SDK, codecs, snakeCase, InferSelectModel, and 0.x-to-1.0 upgrades.
Type-safe SQL ORM for TypeScript with zero runtime overhead. Use when code imports drizzle-orm, drizzle-kit, or drizzle-orm/pg-core, user asks about Drizzle schema design, queries, relations, migrations, or database management with Drizzle ORM. Covers PostgreSQL focus with pgTable, pgEnum, pgSchema, pgView, and drizzle-kit CLI.
CLI for Drupal.org issue lifecycle management. Use when fetching issue details, generating patches/interdiffs, listing project or maintainer issues, looking up releases, or working with GitLab merge requests on issue forks. Also supports projects that have migrated to GitLab work items. Pass --format=llm to every read command for structured XML output optimised for agent consumption.
After coding work, sweep comments in changed files and sync only related durable guidance through WORTH → PLACE → SHAPE. TRIGGER when: user asks to update/sync durable docs, code comments, or AGENTS.md after work; an executor skill reaches close-out.
Extract construction intelligence from AutoCAD DWG files by compiling libredwg from source, converting DWG to DXF, and parsing all entity types (survey points, utility structures, contours, property boundaries, construction keynotes, grading data). Use this skill whenever the user uploads or references a .dwg file, mentions AutoCAD drawings, Civil 3D files, grading plans, site plans in DWG format, or wants to extract data from CAD files. Also triggers when processing project documents that include .dwg files, or when the user says "read the DWG", "extract from CAD", "parse the drawing file", "process the civil plans", "what's in the DWG", "survey points from the drawing", "utility data from CAD". This skill handles the entire pipeline from binary DWG to structured JSON data.
Analyze business cycle phases (expansion, peak, contraction, trough) and their implications for business strategy and policy response. Use this skill when the user needs to identify the current economic phase, anticipate cyclical turning points, or adapt business strategy to macroeconomic cycles — even if they say 'are we heading into a recession', 'how should we prepare for a downturn', or 'when will the economy recover'.
ElevenLabs speech-to-text with Scribe models and forced alignment via inference.sh CLI. Models: Scribe v1/v2 (98%+ accuracy, 90+ languages). Capabilities: transcription, speaker diarization, audio event tagging, word-level timestamps, forced alignment, subtitle generation. Use for: meeting transcription, subtitles, podcast transcripts, lip-sync timing, karaoke. Triggers: elevenlabs stt, elevenlabs transcription, scribe, elevenlabs speech to text, forced alignment, word alignment, subtitle timing, diarization, speaker identification, audio event detection, eleven labs transcribe
ElevenLabs voice isolator - remove background noise and isolate vocals from audio via inference.sh CLI. Capabilities: noise removal, voice extraction, audio cleanup, background removal. Use for: podcast cleanup, interview audio, music vocals, noisy recordings, audio restoration. Triggers: voice isolator, noise removal, background removal, isolate voice, clean audio, remove background noise, audio cleanup, voice extraction, elevenlabs isolator, eleven labs noise, vocal isolation, denoise, audio restoration, voice separation
Email marketing design with layout patterns, subject line formulas, and deliverability rules. Covers welcome sequences, promotional emails, transactional templates, and mobile optimization. Use for: email marketing, newsletter design, drip campaigns, email templates, transactional emails. Triggers: email design, email template, email marketing, newsletter design, email layout, email campaign, drip campaign, welcome email, promotional email, transactional email, email subject line, email header image, email banner
Analyzes and improves LLM prompts and agent instructions for token efficiency, determinism, and clarity. Use when (1) writing a new system prompt, skill, or CLAUDE.md file, (2) reviewing or improving an existing prompt for clarity and efficiency, (3) diagnosing why a prompt produces inconsistent or unexpected results, (4) converting natural language instructions into imperative LLM directives, or (5) evaluating prompt anti-patterns and suggesting fixes. Applies to all LLM platforms (Claude, GPT, Gemini, Llama).
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.