Skills（SKILL.md）は、AIエージェント（Claude Code、Cursor、Codexなど）に特定の能力を追加するための設定ファイルです。
詳しく見る →Python coding guidelines and best practices. Use when writing, reviewing, or refactoring Python code. Enforces PEP 8 style, syntax validation via py_compile, unit test execution, modern Python versions only (no EOL), uv for dependency management when available, and idiomatic Pythonic patterns.
Orchestrate multi-agent teams with defined roles, task lifecycles, handoff protocols, and review workflows. Use when: (1) Setting up a team of 2+ agents with different specializations, (2) Defining task routing and lifecycle (inbox → spec → build → review → done), (3) Creating handoff protocols between agents, (4) Establishing review and quality gates, (5) Managing async communication and artifact sharing between agents.
Short video script generator. 短视频脚本生成器、视频脚本、抖音文案、抖音脚本、快手脚本、口播稿、视频拍摄脚本、YouTube脚本、YouTube Shorts脚本、B站脚本、bilibili脚本、分镜脚本、视频大纲、视频文案、短视频创作、Reels脚本、TikTok脚本、vlog脚本、带货脚本、种草视频脚本、系列视频规划、视频数据复盘、完播率分析、前3秒钩子。Generate complete video scripts with hooks, outlines, titles, tags, CTA, storyboards, series planning, and data review. Use when: (1) creating short video scripts for any platform, (2) writing口播稿/talking-head scripts, (3) generating viral video titles, (4) planning video outlines and storyboards, (5) writing opening hooks (first 3 seconds), (6) generating CTA/ending prompts, (7) planning video series, (8) reviewing video performance data. 适用场景：写短视频脚本、拍摄脚本、口播文案、视频策划、爆款标题、开场钩子、结尾引导、完整分镜、系列规划、数据复盘。 Triggers on: video script creator.
Collect all customer reviews from an Amazon product URL or product-reviews URL through a logged-in Chrome session on port 9222, export a 14-column factual workbook, optionally fill translations through DeepLX, and then help the model tag the rows into a final delivery-ready spreadsheet. Use when the user sends an Amazon link and wants review scraping, competitor review analysis, review export, or a delivery-ready spreadsheet with usernames, review links, review time, helpful votes, translation, summary, sentiment, categories, and tags.
Place outbound phone calls via Mobayilo with safe defaults (preview mode by default) and explicit live execution.
Ship code with oneshot CLI. One command that plans, executes, reviews, and opens a PR. Runs over SSH or locally. Use when the user wants to ship code changes, automate PRs, or run a coding pipeline with Claude and Codex.
AetherCore v3.3.4 - Security-focused final release. High-performance JSON optimization with universal smart indexing for all file types. All security review issues fixed, ready for production.
>
Stateful cron system for OpenClaw with persistent memory, change detection, smart routing, token budget tracking, and self-healing. Requires 'openclaw' and 'gog' CLIs. Broad I/O capabilities (file reads, HTTP requests, shell execution). Must be reviewed carefully before use. See security notes below.
Local-first TikTok Growth OS for strategy, hooks, scripts, retention design, and analytics feedback. Use when the user mentions TikTok, short-form video, hooks, scripts, retention, virality, content pillars, series planning, account positioning, or performance review. Generates execution-ready outputs and stores all assets locally. No API, no posting, no platform automation.
Personal execution engine for tasks, projects, reminders, commitments, follow-ups, and next actions. Use whenever the user mentions something they need to do, remember, plan, follow up on, prioritize, or make progress on. Also use when the user feels overwhelmed, brain-fogged, unsure what to do next, or needs to offload mental strain. This skill captures natural language, turns vague intentions into structured action, tracks momentum, and surfaces the most human-friendly next step. Local-only storage.
Use when designing, reviewing, or refactoring a CLI to reliably serve AI agents alongside humans — including evaluating stdout contracts, schema introspection, dry-run, exit codes, auth delegation, safety tiers, and self-description. Use when converting an API or SDK into an agent-usable CLI interface.
Use when evaluating biology grant ideas in the Chinese funding context (NSFC, MOST, etc.) — diagnosing project legitimacy, mechanism-centered scientific questions, reviewer-aware logic, innovation discipline, feasibility, and scope control across funding levels (youth, general, key).
Use when evaluating grant ideas, diagnosing proposal logic, framing fundable projects, strengthening reviewer-aware arguments, or preparing to write any section of a research proposal.
>
Apply these principles when writing, reviewing, or refactoring code. They are not rules to follow blindly — use judgment, but default to clean.
The most comprehensive proactive life management skill for AI agents. Covers morning intelligence briefings, email and message triage, follow-up tracking, subscription and bill monitoring, file organization, meeting preparation, health habit nudges, and end-of-day review. Built for busy professionals, parents, entrepreneurs, and anyone who wants their AI to run the operational layer of daily life without being micromanaged. No technical setup required.
Real estate transaction support with affordability analysis, property evaluation, and offer strategy. Use when user mentions buying a home, selling property, house hunting, mortgages, inspections, or rental agreements. Calculates true affordability, evaluates properties systematically, builds offer strategies, reviews contracts, and tracks transaction milestones. NEVER provides investment advice.
Solo business management with dashboard, pipeline tracking, invoicing, and weekly reviews. Use when user mentions solo business, clients, revenue, pipeline, invoices, or priorities. Produces business dashboard, tracks prospects from lead to close, drafts invoices, prioritizes actions, and runs weekly business reviews. All business data stays private and local.
Local shell copilot for command planning, safe execution, preview-first workflows, output summarization, privacy-aware history controls, and step-by-step terminal help. Use whenever the user wants to run terminal commands, inspect files, debug shell issues, automate local tasks, or translate natural language into shell actions. Prefer safe preview before mutation. Require explicit confirmation for destructive commands. Local-only.
Nightly memory consolidation and self-reflection for OpenClaw agents. Your agent dreams — reviewing sessions, organizing memories, pruning stale info, and reflecting on its own behavior. Works with any OpenClaw agent. Features: 5-phase dream cycle, safe 2-pass deletion, automatic backup, change gates (>50% blocked), gate check (24h + 5 sessions), growth notifications, old memory resurface, zero-config setup. Inspired by Claude Code Dream but open-source with real self-awareness. Use when: dream, memory, consolidation, self-reflection, agent identity, persistent memory, long-term memory, memory organization, nightly cleanup, memory management.
Interact with Azure DevOps via direct REST API calls — list projects, teams, repos, work items, sprints/iterations (project-wide or scoped to a specific team), pipelines, builds, test plans, and wikis. Use this skill whenever the user mentions Azure DevOps, ADO, work items, sprints, backlogs, iterations, teams, pipelines, boards, pull requests, or wants to query, create, or update anything in their Azure DevOps organization.
UI/UX design intelligence. 50 styles, 21 palettes, 50 font pairings, 20 charts, 9 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient. Integrations: shadcn/ui MCP for component search and examples.
>
Audit another Codex skill for structural compliance, trigger quality, instruction clarity, reuse of scripts or references, and overall maintainability. Use when Codex is given a skill folder and needs to judge whether the skill is qualified, explain why it passes or fails, and summarize strengths, weaknesses, blockers, and improvement ideas across multiple dimensions.
Analyze data with `adme-property-predictor` using a reproducible workflow, explicit validation, and structured outputs for review-ready interpretation.
Generate ARRIVE 2.0 compliant animal research protocols with structured
Use blind-review-sanitizer for academic writing workflows that need structured anonymization, explicit assumptions, and clear output boundaries for double-blind submission.
Use when identifying seminal papers in a research field, mapping research lineage and intellectual heritage, discovering related work through reference tracking, or finding potential collaborators through co-citation analysis. Maps citation networks to trace research evolution, identify influential papers, and discover hidden connections in scientific literature. Supports systematic reviews, bibliometric analysis, and research planning through comprehensive citation tracking.
Check for co-authorship conflicts between authors and suggested reviewers
Simulates NIH study section peer review for grant proposals. Triggers
Assists researchers with Institutional Review Board (IRB) application tasks, including drafting informed consent documents, reviewing research protocols for compliance, generating application forms, and preparing submission checklists. Use when the user mentions IRB, Institutional Review Board, research ethics, human subjects research, protocol review, informed consent, or needs help preparing or reviewing an IRB application or submission.
Use when creating forest plots for meta-analyses, visualizing effect sizes across studies, or generating publication-ready meta-analysis figures. Produces high-quality forest plots with confidence intervals, heterogeneity metrics, and subgroup analyses.
Analyze data with `metagenomic-krona-chart` using a reproducible workflow, explicit validation, and structured outputs for review-ready interpretation.
Assist in drafting professional peer review response letters. Trigger
Analyze data with `phylogenetic-tree-styler` using a reproducible workflow, explicit validation, and structured outputs for review-ready interpretation.
Analyze data with `pseudotime-trajectory-viz` using a reproducible workflow, explicit validation, and structured outputs for review-ready interpretation.
Build complex Boolean query strings for precise PubMed/MEDLINE literature retrieval. Trigger when user needs MeSH term mapping, Boolean query construction, advanced PubMed filters, citation searching, systematic review search strategy, or clinical query optimization.
Polishes response letters by transforming defensive or harsh language.
Analyze data with `smiles-de-salter` using a reproducible workflow, explicit validation, and structured outputs for review-ready interpretation.
Analyze data with `survival-curve-risk-table` using a reproducible workflow, explicit validation, and structured outputs for review-ready interpretation.
Automated abstract screening tool for systematic literature reviews with PRISMA workflow support.
Analyze data with `toxicity-structure-alert` using a reproducible workflow, explicit validation, and structured outputs for review-ready interpretation.
Create and manage stock crypto portfolios with live AISA pricing and P&L tracking. Use when: the user wants to add holdings, inspect portfolio performance, rename portfolios, or review current profit and loss.
Extract clean article content from URLs through the AISA Tavily extract endpoint. Use when: the user already has URLs and needs readable page content for summarization, comparison, or evidence review.
Analyze any GitHub repository in AI-friendly format. Convert entire repos to single markdown documents, generate architecture diagrams with Mermaid, inspect structure trees, language breakdowns, and recent activity. Includes GitHub URL tricks, API shortcuts, and advanced search techniques. Read-only analysis — never executes code from repositories. Built for AI agents — Python stdlib only, no dependencies. Use for repository analysis, code architecture review, open source research, GitHub intelligence, repo documentation, and codebase understanding.
Security audit and quarantine system for third-party OpenClaw skills. Use when evaluating, reviewing, or installing any skill from ClawHub or external sources. Automatically triggered before any skill installation.
AI founder coaching system — daily decision journal, accountability tracking, weekly strategy reviews, and AI-era specific questions on moat, commoditization risk, and outcome-based pricing. Not generic startup advice. Use for founder productivity, decision tracking, and strategic reflection.
Multi-model consensus system — send a query to 3+ different LLMs via OpenRouter simultaneously, then a judge model evaluates all responses and produces a winner, reasoning, and synthesized best answer. Like having a board of AI advisors. Use for important decisions, code review, research verification.
AI research pipeline with automatic logging. Search via Perplexity, auto-save results to SQLite with topic and project metadata, full Langfuse tracing. Never lose a research session again. Use when conducting research, competitive analysis, or building a knowledge base.