AI Skills Hub - Browse 180,000+ Agent Skills

Invoke this skill when: - Pod status is `CrashLoopBackOff`, `Error`, or `ImagePullBackOff` - User mentions "pod crashing", "deployment failing", or "container restarting" - Kubernetes deployment not reaching Ready state

Invoke this skill when: - `npm run build` or `next build` fails - `pip install` or Python dependency issues occur - TypeScript compilation errors appear - User mentions "build failing" or "deployment error"

How to create a StateChanger. useful for when creating a StateChanger

.NET test execution patterns and diagnostics. Use when running tests, analyzing test failures, or configuring test options.

Follow these patterns when designing guardrails validation contracts in OptAIC. Use for signal bounds, dataset schemas, portfolio constraints, PIT validation, and other domain-specific validation rules. Covers the "Law vs Police" architecture where Definitions contain contracts and the Engine enforces them.

Guide for designing Definition resources in OptAIC (the "Law"). Use when creating PipelineDef, StoreDef, AccessorDef, OpDef, MLModuleDef, or PortfolioOptimizerDef. Covers interface_spec, input/output schemas, compatibility rules, and embedded guardrail contracts.

Run comprehensive quality gates (mypy, ruff, pytest) before committing code. Use after completing implementation tasks and before marking phases complete.

Create CLI commands using Typer framework with Rich formatting, proper error handling, and user-friendly output. Use when implementing CLI tasks (T024, T030, T036, T042, etc.).

Create production-grade MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Embodies expertise of a Senior Backend Engineer specializing in API design and LLM integrations. Use when: user requests building MCP servers, LLM tool integrations, API wrappers for AI agents, or Model Context Protocol implementations. Outputs: Complete MCP server projects (TypeScript or Python) with tools, resources, and evaluation tests. Keywords: MCP, Model Context Protocol, LLM tools, API integration, agent tools, FastMCP, MCP SDK, tool server

Comprehensive Word document creation, editing, and analysis with support for tracked changes, comments, formatting preservation, and text extraction. Use when: Claude needs to create new Word documents, modify existing content with tracked changes, add comments, work with professional documents, or perform document analysis. Outputs: Word documents (.docx), tracked changes, comments, formatted text, markdown conversions. Keywords: docx, word, document, tracked changes, comments, ooxml, professional documents

Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

Test an implementation by running the local dev environment and performing E2E testing through the browser. Use when the user asks to test a feature, debug an issue, or verify a fix works correctly. (project)

Extract defensible conclusions from experiment results without overclaiming. Use when reading test results, interpreting p-values or intervals, or deciding what can actually be said with confidence.

This skill should be used when invoked by project-orchestrator at Gate 6, or when the user asks to "review security", "audit vulnerabilities", "harden the application", or needs security assessment. Identifies vulnerabilities and blocks release on Critical/High security issues.

This skill should be used when invoked by project-orchestrator at Gate 5 (before QA), or when the user asks to "verify the implementation", "smoke test", "check if it runs", or needs pre-QA validation. Smoke tests infrastructure, routes, and database to catch "it doesn't even start" issues.

This skill should be used when invoked by project-orchestrator at Gate 3, or when the user asks to "set up Docker", "configure CI/CD", "create infrastructure", or needs development environment setup. Sets up docker-compose, env vars, and CI/CD so developers can start coding.

Implement, refactor, or review worker-based multithreading in TypeScript/JavaScript apps (browser or Node). Use when offloading CPU-heavy loops to Web Workers, setting up a worker pool, moving data across threads, or diagnosing worker crashes, slowdowns, or synchronization bugs.

通过关键字、时间范围或 ID 搜索过去的对话。在"搜索会话"、"找到那场对话"、"我们做了什么"、"查找关于 X 的过去讨论"、"我们在哪"、"我学了什么"、"知识缺口"、"回顾"时触发。也在以下情况触发："我们讨论了..."、"你提到过..."、指代先前工作的过去时动词、没有上下文的物主代词（"我的项目"、"我的认证系统"）、引用未共享上下文的假设性问题。提取干净的对话数据并应用分析透镜以获得结构化见解。

Automatically refresh AWS SSO authentication tokens when encountering expiration errors. Use when AWS MCP tools fail due to expired SSO sessions.

Performs SEO keyword research using DataForSEO API. Use when user asks about keyword ideas, search volume, CPC, competition data, or needs keywords for blog posts, landing pages, or content strategy.

Optimizes images for web by converting to WebP format with optional resizing. Use when user needs to optimize images, generate favicons, create icon sets, make social media cards (og:image, twitter), or create thumbnails.

Generate or edit images using Google Gemini. Use for general-purpose image generation including photos, illustrations, artwork, visual assets, concept art, and any image that isn't a technical diagram or schematic. For flowcharts, circuits, pathways, and technical diagrams, use the scientific-schematics skill instead.

Remove backgrounds from images using AI segmentation. Use when user asks to remove, delete, or make transparent backgrounds from photos or images.

Implement iOS UIContentConfiguration and UIHostingConfiguration for modern UIKit architecture. Use when working with cells, custom views, or replacing legacy view configuration code.

This skill helps launch and configure the Chrome DevTools MCP server, giving Claude visual access to a live browser for debugging and automation. Use when the user asks to set up browser debugging, launch Chrome with DevTools, configure chrome-devtools-mcp, see what my app looks like, take screenshots of my web application, check the browser console, debug console errors, inspect network requests, analyse API responses, measure Core Web Vitals or page performance, run a Lighthouse audit, test button clicks or form submissions, automate browser interactions, fill out forms programmatically, simulate user actions, emulate mobile devices or slow networks, capture DOM snapshots, execute JavaScript in the browser, or troubleshoot Chrome DevTools MCP connection issues. Supports Windows, Linux, and WSL2 environments.

Deploy applications to Coolify self-hosted PaaS. Use when deploying to Coolify, configuring docker-compose for production, setting up environment variables, or troubleshooting deployment issues.

Research agent for external documentation, best practices, and library APIs via MCP tools

Analyze Claude Code sessions via Braintrust

AST-based code search and refactoring via ast-grep MCP

Use this skill when a multi-step task has a validated design and you need a detailed, reviewable implementation plan before touching code.

--- name: architecture-testing description: > Use this skill whenever defining or reviewing architectural boundaries, layering, or project dependency rules to ensure structural integrity and prevent drift. ---

Use this skill before any creative or multi-step work to clarify intent, explore alternatives, and produce a validated design or specification.

Use this skill when adding or standardizing application logging and distributed tracing to ensure structured logs, OWASP-aligned security hygiene, and Correlation-Id propagation across the system.

Use this skill whenever defining, modifying, or reviewing structure and boundaries to ensure related code, tests, docs, templates, and scripts are co-located at the narrowest meaningful scope (slice, shared library, monorepo, or API version), with shared extraction only when reuse is proven.

--- name: testcontainers-integration-tests description: > Use this skill whenever implementing or reviewing integration tests that require real infrastructure (databases, queues, caches) for single-component validation. ---

Read frontmatter metadata from .ai-docs documents without loading full content. Use when asked to scan documents for topics, check what sections are available, find documents by keywords, or discover content overview. Triggers on phrases like "read the frontmatter", "read frontmatter", "check the frontmatter", "show me the frontmatter", "get the frontmatter", "what topics are in", "scan the frontmatter", "what sections does it have", "check if it contains", "does it contain topics about", "tell me if it contains", "what's the metadata".

Extract specific sections from .ai-docs documents using line numbers. Use when asked to read a specific section, get content from certain lines, extract a topic, or retrieve partial document content. Triggers on phrases like "read the section", "read the section about", "extract the section", "extract lines", "read lines", "get the content from", "get content from lines", "show me the section", "show me the section on", "show me lines", "what does the section say", "what's in the section", "read the dependency injection section".

Search past work for relevant decisions, patterns, and approaches. Auto-triggers when discussing similar work, debugging recurring issues, or asking "what did we do before?"

This skill should be used when adding or updating Homebrew formulas in this tap. Invoked for creating new formulas, updating versions, fixing formula issues, or managing multi-platform binary distributions.

Data collector specs for lcaudit — required vs optional tiers, Promise.allSettled patterns, screenshot storage, SERP provider. Use when building or modifying any collector.

API routes, error handling, retry logic, rate limiting, and tracing patterns for lcaudit. Use when building API endpoints, server-side logic, or integrations.

Convert Markdown files or folders to standard multi-file format. Splits content by H1 headers into separate chapter files with a table of contents. Use when preparing Markdown content for audioread processing.

Convert EPUB ebook files to standard Markdown format. Extracts content by chapters, separates images, and generates a structured output with a table of contents file and individual chapter files. Use when converting EPUB to Markdown for further processing.

클로바노트(Clova Note) 음성 기록 전문을 바탕으로 회의록을 요약하는 스킬. 사용자가 클로바노트 텍스트 파일(.txt)을 제공하면, 구조화된 회의록 요약을 마크다운으로 생성한다. 트리거: 회의록 요약, 클로바노트, 미팅노트 정리, 회의 정리, 또는 클로바노트 형식의 .txt 파일이 제공될 때.

A powerful RSS subscription manager and reader. Use it to (1) Import/export OPML files, (2) Manage RSS feeds (add, remove, categorize), (3) Validate feed connectivity, (4) Schedule periodic updates via cron, (5) Fetch and summarize articles with a progressive disclosure approach.

Operational metrics dashboard with anomaly detection from metrics endpoint or state file.

Preview release — version bump, changelog, and flagged commits since last tag.

This skill should be used when the user asks to "list screens", "design screen flow", "create screen inventory", "plan navigation", or "map user journeys". Creates screen inventory and transition diagrams for Wave A parallel execution.

This skill should be used when the user asks to "design API", "create REST endpoints", "document API specifications", "define API contracts", "plan external integrations", or "create OpenAPI spec". Designs RESTful APIs and external system integration specifications for Wave B parallel execution.

This skill should be used when the user asks to "analyze implementation gap", "evaluate feasibility", "assess codebase compatibility", "compare requirements with existing code", or "plan implementation approach". Analyzes gaps between requirements and existing codebase to inform implementation strategy.

Design patterns and conventions for the vm0 CLI user experience

This document records common patterns and best practices when using ccstate in the vm0 platform.

Create command aliases for skills with arguments

Ethereum Beacon Chain explorer workflows using the beaconcha.in API (V1 and V2). Use when the user mentions "validator", "beacon chain", "staking rewards", "attestation", "epoch", "slot", "beaconcha.in", "BeaconScore", "validator efficiency", "staking performance", "consensus layer", "CL rewards", "execution layer rewards", "validator balance", "validator status", "block proposal", "sync committee", "slashing", "validator queue", "activation queue", "exit queue", "withdrawal", "ETH staking", "validator dashboard", "validator monitoring", or asks to look up, monitor, analyze, or report on Ethereum validators or network health. Also use when the user provides a validator index, public key, or withdrawal address and wants data about it. Do NOT use for general Ethereum execution layer queries unrelated to staking or the beacon chain.

Implements Manus-style file-based planning for complex tasks. Creates task_plan.md, findings.md, and progress.md. Use when starting complex multi-step tasks, research projects, or any task requiring >5 tool calls.

Start a new OpenSpec change using the experimental artifact workflow. Use when the user wants to create a new feature, fix, or modification with a structured step-by-step approach.

Use Cocoon for confidential AI inference via TEE-protected workers on the TON blockchain. OpenAI-compatible API. Use when the user wants to run inference through Cocoon, check Cocoon status, list Cocoon models, or use decentralized/confidential/TEE compute.

Generate markdown documentation for a module or feature

Generate AI music with ElevenLabs Music API. Use for: background music, soundtracks, jingles, theme songs, instrumental tracks, AI music composition.

This skill should be used when the user asks to "share a command", "make a command public", "publish a command", "share my slash command", or mentions making a Claude Code command available publicly. Publishes user-defined commands from ~/.claude/commands/ to the HartreeWorks/claude-commands repository.

Upload Claude Code skills to claude.ai for use in Claude Desktop. Use when the user says "sync skill to desktop", "upload skill to claude.ai", "sync all skills to desktop", or wants to make their Claude Code skills available in Claude Desktop.

This skill should be used when the user asks to "share a script", "make a script public", "publish a script", "share my script", or mentions making a Claude Code script available publicly. Publishes scripts from ~/.claude/scripts/ to the HartreeWorks/claude-scripts repository.

Analyze MongoDB collection schemas, infer data types, detect inconsistencies, and recommend schema design improvements. Use when the user asks about database structure, data modeling, schema validation, or wants to understand their MongoDB collections.

Orchestrate feature implementation through planning, agent delegation, and user checkpoints. Use when starting work on a Jira ticket or implementing a feature with structured review gates. Triggers on "implement feature", "work on ticket", "orchestrate", or ticket IDs like ENG-12345. CRITICAL - This skill coordinates agents, it does NOT implement directly.

Use when needing to connect to a running Elixir/BEAM node via AWS ECS exec for live inspection and debugging. Triggers on "connect to the node", "check running processes", "inspect memory", "remote console", "ECS exec", "check the node", or any live BEAM debugging request.

BEAM VM and OTP patterns including processes, GenServer, supervision, Registry, ETS, and production debugging

Manage Mihomo proxy on Raspberry Pi — update subscriptions, switch nodes, and diagnose network issues via the auto-mihomo MCP HTTP API.

Create atomic git commits with heuristic analysis, conventional-commit formatting, staging rules, optional deep analysis, Graphite stack commits, and optional push. Use when the user asks to craft a commit message, commit changes, stage/commit only session edits, or run a commit workflow with flags like --all, --deep, --push, or --stack.

Use when starting new feature development to create isolated workspace without affecting main branch

Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

Run the autonomous review loop to address PR comments, fix issues, and push changes iteratively until all comments are resolved.

Write and maintain Gherkin `.feature` behavior specifications (behavior specs) for features and core system modules. Use AFTER completing requirement clarification and writing-plans (docs/plan ready), BEFORE implementation begins. Triggers writing behavior specs, defining acceptance criteria, updating system behavior definitions, creating feature specs, reviewing core module behavior changes, Gherkin specs, BDD specs, 行为规范, 行为定义, 验收标准, 系统行为. Outputs `.feature` files to docs/spec/feature-spec/ and docs/spec/core-spec/ and lints generated specs with gherkin-lint. Also serves as agent memory for understanding current system behavior across modules.

飞书（Lark）全通道自动化。使用 lark-mcp 工具操作飞书多维表格（Bitable）、发送消息、管理文档、创建群组、自动化工作流等。当用户需要操作飞书平台、同步数据到飞书表格、发送飞书通知、管理飞书文档或自动化飞书业务流程时使用此技能。

Convert GSD codebase mapping to Autocoder app_spec.txt. This skill should be used when the user has run /gsd:map-codebase and wants to use Autocoder on an existing project. Triggers: "convert to autocoder", "gsd to spec", "create app_spec from codebase", "use autocoder on existing project", after /gsd:map-codebase completion.

Execute an approved plan step by step, verifying each step before proceeding.

Set up a working branch and start the dev environment. Be fast and opinionated — no unnecessary questions.

Create a pull request from the current branch with an auto-generated title and description based on the commits and changes.

This skill should be used when the user asks to "create an API", "design REST endpoints", "implement API routes", "add an endpoint", "create a REST service", or when working on HTTP API code. Provides API design patterns including layered architecture and service functions.

QQ 桌面端自动回复技能（macOS）。通过 AppleScript + 截图实现 QQ 消息读取和自动回复。支持激活 QQ、截取聊天窗口、搜索联系人、发送消息。agent 通过截图分析消息内容，然后调用 reply 命令发送回复。

Production incident response procedures for Python/React applications. Use when responding to production outages, investigating error spikes, diagnosing performance degradation, or conducting post-mortems. Covers severity classification (SEV1-SEV4), incident commander role, communication templates, diagnostic commands for FastAPI/ PostgreSQL/Redis, rollback procedures, and blameless post-mortem process. Does NOT cover monitoring setup (use monitoring-setup) or deployment procedures (use deployment-pipeline).

Deployment procedures and CI/CD pipeline configuration for Python/React projects. Use when deploying to staging or production, creating CI/CD pipelines with GitHub Actions, troubleshooting deployment failures, or planning rollbacks. Covers pipeline stages (build/test/staging/production), environment promotion, pre-deployment validation, health checks, canary deployment, rollback procedures, and GitHub Actions workflows. Does NOT cover Docker image building (use docker-best-practices) or incident response (use incident-response).

Sync and query CalDAV calendars (iCloud, Google, Fastmail, Nextcloud, etc.) using vdirsyncer + khal. Works on Linux.

def export_all_fbx(output_dir): for obj in bpy.data.objects: if obj.type == 'MESH': bpy.ops.object.select_all(action='DESELECT') obj.select_set(True) bpy.context.view_layer.objects.active = obj filepath = os.path.join(output_dir, f"{obj.name}.fbx") bpy.ops.export_scene.fbx( filepath=filepath, use_selection=True, apply_scale_options='FBX_SCALE_ALL' ) ```

ANIMATIONS = { 'idle': (1, 60), 'walk': (61, 120), 'run': (121, 180), 'attack_01': (181, 210), 'death_01': (211, 260), }

Persist agent memory, guidelines, and knowledge seeds in MongoDB. Use for: remember/save/store, search knowledge, add guideline, seed/export/import, migrate workspace state, persistent cross-session context. Supports local, remote, Atlas, VPN+TLS.

Implementation execution for context-driven development. Trigger with ci, /conductor-implement, or /conductor-* commands. Use when executing tracks with specs/plans. For design phases, see designing skill. For session handoffs, see handoff skill.

Capture and verify all agreements before your agent accepts them. Provides audit trails, risk analysis, and dispute evidence for autonomous transactions.

Standardize LeetCode solution additions with JavaDoc, tests, and README index updates for global and LCR packages.

Generate detailed implementation checklists from specification documents. Use when asked to create an implementation plan, impl plan, or implementation checklist for a spec. Transforms architecture/design specs into phased, actionable tasks with file listings, verification steps, and progress tracking. Triggers on requests like "create an implementation plan for this spec", "break down this spec into tasks", or "write an impl.md for this feature".

Run end-to-end MCP protocol tests against the Narra server

Guide for planning complex software architectures using a 4-layer hierarchical approach. Use when designing system architecture, planning technical structure, creating design documents, or reviewing existing architectures. Helps prevent context exhaustion and hallucination by breaking architecture into manageable layers with validation gates.

Research assistant that generates structured `references.md` reports for thesis chapters. It identifies reviews, key papers, and section-specific additional sources via NotebookLM.

Structured messaging protocols for agent team communication including message type selection, plan approval, shutdown procedures, and anti-patterns to avoid. Use this skill when establishing team communication norms, handling plan approvals, or managing team shutdown.

Commit changes and create or update a pull request following project standards. Use when the user wants to commit work, create a PR, update an existing PR, or use the /pr command.

Initialize a fresh Hetzner Ubuntu server with a non-root user configured for Docker application deployment. Sets up secure SSH access, Docker permissions, and a dedicated data directory.

새 감정 카테고리를 Emova MCP 서버의 refine-task 도구에 추가합니다.

Context-aware routing to code generation workflows including SwiftGen, Sourcery, Feature Flags, and Protobuf. Use when running make generate, adding feature flags, or working with generated files.

Context-aware routing to the Anytype iOS localization system. Use when working with .xcstrings files, Loc constants, hardcoded strings, or user-facing text.

Access team credentials stored in GitLab CI/CD variables - servers, databases, APIs

Create clear, maintainable documentation: READMEs, runbooks, architecture docs, API docs, and onboarding guides.