Skills directory · page 3

並行リポジトリ対応のレビューパスを使用して、スキル、コマンド、ルール、フック、エクストラを DAILY と LIBRARY のバケットに分類することで、特定のリポジトリ向けのエビデンスに基づいた ECC インストール計画を構築します。プロジェクトが完全なバンドルをロードする代わりに実際に必要なものに ECC をトリミングする必要がある場合に使用します。

Build an evidence-backed ECC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when ECC should be trimmed to what a project actually needs instead of loading the full bundle.

Build an evidence-backed ECC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when ECC should be trimmed to what a project actually needs instead of loading the full bundle.

Verify supply chain integrity for AI agent plugins, tools, and dependencies. Use this skill when: - Generating SHA-256 integrity manifests for agent plugins or tool packages - Verifying that installed plugins match their published manifests - Detecting tampered, modified, or untracked files in agent tool directories - Auditing dependency pinning and version policies for agent components - Building provenance chains for agent plugin promotion (dev → staging → production) - Any request like "verify plugin integrity", "generate manifest", "check supply chain", or "sign this plugin"

Tools are how AI agents interact with the world. A well-designed tool is the difference between an agent that works and one that hallucinates, fails silently, or costs 10x more tokens than necessary. This skill covers tool design from schema to error handling. JSON Schema best practices, description writing that actually helps the LLM, validation, and the emerging MCP standard that's becoming the lingua franca for AI tools. Key insight: Tool descriptions are more important than tool implementa

Use when architecting a multi-step LLM or agent pipeline, choosing single vs multi-agent, or fixing context bloat and broken handoffs. Scaffolds the pattern, handoff contracts and failure handling with workflow_scaffolder.py.

Run several agents on one task in parallel and keep the best result: faster code, copy variants, research angles. Use whenever the user wants multiple attempts compared or best of N; spawns agents in git worktrees, ranks and merges the winner.

Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing. Use when planning or executing engineering work that agents will carry out end to end.

Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing. Use when AI agents perform most implementation work and humans enforce quality and risk controls.

Claude Code 上に永続的なマルチエージェントオペレーティングシステムを構築します。カーネルアーキテクチャ、スペシャリストエージェント、スラッシュコマンド、ファイルベースのメモリ、スケジュールされた自動化、外部データベースなしの状態管理をカバーします。

Build persistent multi-agent operating systems on Claude Code. Covers kernel architecture, specialist agents, slash commands, file-based memory, scheduled automation, and state management without external databases. Use when building a persistent multi-agent system on Claude Code with its own memory, commands, and scheduling.

Use your assigned AgentMail inbox to read email tasks, explicitly send or reply, and check delivery. Provided automatically by your inbox assignment.

Autonomous AI agent platform for building and deploying continuous agents. Use when creating visual workflow agents, deploying persistent autonomous agents, or building complex multi-step AI automation systems.

Maintain README.md, AGENTS.md/CLAUDE.md, skills, and repo context. Create, scaffold, or initialize AGENTS.md files and project-local agent skills. Use task evidence to delete obsolete skills, merge overlapping skills, and capture recurring workflows.

Multi-agent orchestration framework for autonomous AI collaboration. Use when building teams of specialized agents working together on complex tasks, when you need role-based agent collaboration with memory, or for production workflows requiring sequential/hierarchical execution. Built without LangChain dependencies for lean, fast execution.

Use for current official documentation about Codex, Codex CLI, or Claude Code behavior, configuration, prompting, skills, permissions, tools, surfaces, capabilities, troubleshooting, hooks, app-server, or hook trust. Consult the relevant official source before answering.

Use local Codex/Claude Code transcripts to assess recurrence risk for agent behavior. Recommend durable fixes to AGENTS.md or skills from that evidence. Not for routine post-success skill-evolution or agent self-improvement reviews.

Framework for building LLM-powered applications with agents, chains, and RAG. Supports multiple providers (OpenAI, Anthropic, Google), 500+ integrations, ReAct agents, tool calling, memory management, and vector store retrieval. Use for building chatbots, question-answering systems, autonomous agents, or RAG applications. Best for rapid prototyping and production deployments.

Data framework for building LLM applications with RAG. Specializes in document ingestion (300+ connectors), indexing, and querying. Features vector indices, query engines, agents, and multi-modal support. Use for document Q&A, chatbots, knowledge retrieval, or building RAG pipelines. Best for data-centric LLM applications.

Redesign the assigned repository or repositories for agent understanding, control, and cumulative learning. Use when asked to make a system agent-intuitive, agent-ergonomic, or agent-accretive, then revise its design documents and plans.

Use for backlog work: writing user stories and acceptance criteria, splitting epics, estimating or prioritizing for a sprint. Generates stories with user_story_generator.py.

Agile product ownership toolkit for Senior Product Owner including INVEST-compliant user story generation, sprint planning, backlog management, and velocity tracking. Use for story writing, sprint planning, stakeholder communication, and agile ceremonies.

Use when the user asks about aging or longevity research - what the evidence says on a drug, supplement, diet, gene or process (rapamycin, metformin, senolytics, NAD+, GLP-1 drugs, caloric restriction), biomarkers of aging such as epigenetic clocks, clinical trials in aging and age-related disease, new papers and preprints in the field, research grants, longevity industry news or drug programs in development - or asks to follow a topic on AgingResearch.News. Answers with the AgingResearch.News connector's tools, which read new papers on aging every day and judge each one for its link to aging.

AI agent payment infrastructure — ACTP escrow, x402 instant payments, USDC settlement on Base L2. Interactive onboarding: asks your preferences, generates customized agent code, and verifies setup. Covers provider/requester patterns, adapter routing, 8-state machine, pricing, disputes, and identity.

EU AI Act readiness: risk tier, duties, deadlines. Load whenever AI ships in the EU or the user asks if the Act applies.

Expert in designing and building autonomous AI agents. Masters tool use, memory systems, planning strategies, and multi-agent orchestration. Use when: build agent, AI agent, autonomous agent, tool use, function calling.

Hand a bounded coding or review task to an external autonomous coding-agent CLI in its own git worktree, then verify the diff, the tests and every acceptance claim yourself before integrating. Use when work should outlive the current turn, run isolated from the operator's checkout, or be compared across agents. Not for a quick bounded subtask in the current turn, and not for work with no written task spec or acceptance command.

Use before letting AI or an agent act on a marketing or GTM task without human review, or when writing an AI review policy. Places each workflow in AI-primary, collaborative or human-primary by stakes and reversibility.

Lists an AI product on free AI directories, launch sites and MCP or agent directories, one at a time, from FalcoScan's checked list of free directories, and keeps a record of each submission. Use when the user wants to list, launch or promote an AI tool, SaaS, MCP server or agent on directories, or asks where to submit their product.

Engineering operating model for teams where AI agents generate a large share of implementation output. Use when setting team process, review gates, or ownership rules for a codebase largely written by agents.

Every product will be AI-powered. The question is whether you'll build it right or ship a demo that falls apart in production. This skill covers LLM integration patterns, RAG architecture, prompt engineering that scales, AI UX that users trust, and cost optimization that doesn't bankrupt you. Use when: keywords, file_patterns, code_patterns.

Prune agent scratch directories on macOS. Move clearly outdated `.ai/` files into `.ai/archive/YYYY-MM-DD/` and trash clearly outdated `.cache/` entries. Leave `.ai/todos` untouched.

Regression testing strategies for AI-assisted development. Sandbox-mode API testing without database dependencies, automated bug-check workflows, and patterns to catch AI blind spots where the same model writes and reviews code. Use when adding regression coverage to AI-assisted code, or when the same model both wrote and reviewed a change.

Regression-testing strategies for AI-assisted development. Covers sandbox-mode API testing without database dependencies, automated bug-check workflows, and patterns for catching AI blind spots when the same model writes and reviews code.

AI 支援開発のためのリグレッションテスト戦略。データベース依存なしのサンドボックスモード API テスト、自動化されたバグチェックワークフロー、同じモデルがコードを書いてレビューする AI のブラインドスポットを捕捉するパターン。

Daily digest of AI releases from Anthropic, Meta, Vercel Labs, OpenAI and Google GitHub repos, the Claude Cowork changelog, the claude.dev blog and Hacker News, deduplicated against a local state file and sent by email. Use when the user asks for a daily AI news digest, to scan AI labs for new releases, or to set up a scheduled release-hunter job.

Use when assessing an AI/ML system or LLM agent for prompt injection, jailbreaks, model inversion, data poisoning or tool abuse. Maps findings to MITRE ATLAS and scores robustness with ai_threat_scanner.py.

Create AI-powered social media content for TikTok, Instagram, YouTube, Twitter/X. Generate: images, videos, reels, shorts, thumbnails, captions, hashtags. Tools: FLUX, Veo, Seedance, Wan, Kokoro TTS, Claude for copywriting. Use for: content creators, social media managers, influencers, brands. Triggers: social media content, tiktok, instagram reels, youtube shorts, twitter post, content creator, ai influencer, social content, reels, shorts, viral content, thumbnail generator, caption generator, hashtag generator, ugc content

Checklist of patterns that make writing sound AI-generated (filler openers, stacked connectors, rule of three, hedging). Use when reviewing or de-AI-ing a draft (newsletters, social posts, blog posts, emails) or when the user says a text sounds like AI.

Verified configuration reference for the three AI tools Aide supports: Claude Code, GitHub Copilot and OpenAI Codex. Covers instruction files, skill discovery paths, frontmatter field support, rules, agents, hooks, MCP and config formats, with a comparison matrix and source links. Use when: answering which tool reads which path, changing an installer, adding or moving a skill, checking frontmatter field support, updating the support matrix. Do NOT use for: news about new releases (use check-news), general web questions about the tools.

How Claude Code and the Copilot and Codex CLIs are kept up to date on a machine, and how to install the daily upgrade job — plain cron on an always-on machine, a launchd agent on a laptop (cron silently skips runs the machine sleeps through). Use when: setting Aide up on a new machine, the AI tools have gone stale, choosing between cron and launchd, editing core/scripts/upgrade-ai-tools. Do NOT use for: checking for news (use check-news).

Expert in building products that wrap AI APIs (OpenAI, Anthropic, etc.) into focused tools people will pay for. Not just 'ChatGPT but different' - products that solve specific problems with AI. Covers prompt engineering for products, cost management, rate limiting, and building defensible AI businesses. Use when: AI wrapper, GPT product, AI tool, wrap AI, AI SaaS.

Analyze the codebase for a spec. Maps affected files with file:line references, grades the change's complexity (LOW/MEDIUM/HIGH) by its risk, and creates an implementation plan with TDD. Use when: analyzing the codebase for an existing task, filling in 2-analysis.md and 3-solution.md, needing an overview of affected files and API impact. Do NOT use for: creating a new task (use aide-create), implementation (use aide-implement).

Archive a finished spec and feed its durable knowledge back into the project's living documentation — resolving the branch's merge conflict with the default branch first, if there is one. Use when: a spec is done, closing it out, cleaning up the specs root, a spec's branch conflicts with the default branch. Do NOT use for: creating specs (use aide-create), unfinished work, deleting specs.

Close a spec that is not going to work: record the reason, move it to archive/, and delete its code branch without merging it — resolving the specs branch's merge conflict with the default branch first, if there is one. Use when: a spec's idea does not hold, its description asked for something that cannot be built, a work round needs to end without being finished. Do NOT use for: finished work (use aide-archive), another round on the same spec (change an acceptance criterion and run analyze again), unfinished work that is still worth finishing.

Create a spec — a title and a description — as the 4-file spec structure (description, analysis, solution, status). Use when: creating a new spec, starting new work that needs documentation. Do NOT use for: analysis (use aide-analyze), implementation (use aide-implement), code review.

No-stakes thinking partner before /aide-create: explore a problem or idea, weigh approaches, sharpen the scope — without creating any files. Use when: an idea is not ready for a spec yet, weighing whether/how to do something, unclear scope, "what would it take to ...". Do NOT use for: creating specs (use aide-create), analysis of an existing spec (use aide-analyze), implementation.

Implement the solution for a spec with Test-Driven Development (RED → GREEN → VERIFY). Reads the existing analysis and plan, writes tests first, implements, and runs the quality check. Use when: implementing a solution with TDD, having a completed analysis and implementation plan, coding based on 3-solution.md. Do NOT use for: creation (use aide-create), analysis (use aide-analyze).

Take an archived spec back into the active list for another round, keeping every file as it is, or, when asked, resetting the analysis, the plan and the status as well. Keeps the description and the archive trail, and marks the point after which nothing counts as having run. Use when: archived work has to be done again, a spec was archived too early, a shipped change has to be redone from its own description. Do NOT use for: creating a new spec (use aide-create), analysis (use aide-analyze), archiving (use aide-archive).

How to add a new skill, update a shared rule, or change the 4-file spec structure in the Aide repo itself, and how install.sh/uninstall.sh must stay mirrored. Use when: adding or editing a skill or rule under core/, changing spec-structure.md, or touching install.sh/uninstall.sh. Do NOT use for: ordinary spec work (create/analyze/implement/archive) in any project Aide is installed into — this is about maintaining Aide's own source.