LLM & Agents

Sets up new projects and generates workflow files like CLAUDE.md, AGENTS.md, Plans.md. Use when user mentions セットアップ, setup, 初期化, initialize, 新規プロジェクト, ワークフローファイル生成. Do NOT load for: 実装作業, レビュー, ビルド検証, デプロイ.

Write tests following TDD principles. Use when implementing features, fixing bugs, or adding test coverage. Covers e2e, integration, and unit testing patterns.

Ingest new procedural memory (skills, patterns, docs) into the vector database.

Initialize or hydrate the agent's memory system and verify configuration.

Context-aware development assistant for AgenticFleet with auto-learning and dual memory (NeonDB + ChromaDB). Handles development workflows with intelligent context management.

End-to-end guide for creating React components in AgenticFleet's frontend, from design tokens through Tailwind styling, shadcn/ui integration, state management, to testing.

Complete guide to the AgenticFleet memory system. Read this first.

Expert prompt engineering skill that transforms Claude into "Alpha-Prompt" - a master prompt engineer who collaboratively crafts high-quality prompts through flexible dialogue. Activates when user asks to "optimize prompt", "improve system instruction", "enhance AI instruction", or mentions prompt engineering tasks.

Systematic approach for creating minimal-change, reversible implementation plans. Claude invokes this skill when transforming requirements/research into executable blueprints. Emphasizes simplicity, safety, and clear verification steps.

Systematic approach for gathering authoritative, version-accurate documentation. Claude invokes this skill when research is needed before implementation. Ensures truth over speed while achieving both.

Systematic methodology for identifying, capturing, and documenting reusable patterns from implementations. Enables automatic learning and knowledge-core.md updates. Claude invokes this after successful implementations to preserve institutional knowledge.

Guide to effective Claude Code skill authoring using TDD methodology and persuasion principles. Triggers: skill authoring, skill writing, new skill, TDD skills, skill creation, skill best practices, skill validation, skill deployment, skill compliance Use when: creating new skills from scratch, improving existing skills with low compliance rates, learning skill authoring best practices, validating skill quality before deployment, understanding what makes skills effective DO NOT use when: evaluating existing skills - use skills-eval instead. DO NOT use when: analyzing skill architecture - use modular-skills instead. DO NOT use when: writing general documentation for humans. YOU MUST write a failing test before writing any skill. This is the Iron Law.

Python testing with pytest, fixtures, mocking, and TDD workflows. Triggers: pytest, unit tests, test fixtures, mocking, TDD, test suite, coverage, test-driven development, testing patterns, parameterized tests Use when: writing unit tests, setting up test suites, implementing TDD, configuring pytest, creating fixtures, async testing DO NOT use when: evaluating test quality - use pensive:test-review instead. DO NOT use when: infrastructure test config - use leyline:pytest-config. Consult this skill for Python testing implementation and patterns.

Delegate tasks to external LLM services (Gemini, Qwen) with quota, logging, and error handling. Triggers: delegate task, external LLM, gemini, qwen, large context, offload Use when: tasks exceed context window or need cheaper processing DO NOT use when: task requires Claude's reasoning.

detailed hook evaluation framework for Claude Code and Agent SDK hooks. Triggers: hook audit, hook security, hook performance, hook compliance, SDK hooks, hook evaluation, hook benchmarking, hook vulnerability Use when: auditing existing hooks for security vulnerabilities, benchmarking hook performance, implementing hooks using Python SDK, understanding hook callback signatures, validating hooks against compliance standards DO NOT use when: deciding hook placement - use hook-scope-guide instead. DO NOT use when: writing hook rules from scratch - use hookify instead. DO NOT use when: validating plugin structure - use validate-plugin instead. Use this skill BEFORE deploying hooks to production.

Complete guide for writing Claude Code and SDK hooks with security-first design. Triggers: hook creation, hook writing, PreToolUse, PostToolUse, UserPromptSubmit, tool validation, logging hooks, context injection, workflow automation Use when: creating new hooks for tool validation, logging operations for audit, injecting context before prompts, enforcing project-specific workflows, preventing dangerous operations in production DO NOT use when: logic belongs in core skill - use Skills instead. DO NOT use when: complex multi-step workflows needed - use Agents instead. DO NOT use when: behavior better suited for custom tool. Use this skill BEFORE writing any hook. Check even if unsure.

Declarative release readiness checklist that mirrors GitHub checks, deployment issues, and documentation requirements. Triggers: release gates, release readiness, deployment checklist, release review, quality signals, rollout scorecard, QA handshake, deployment gates Use when: preparing releases, validating deployment gates, conducting release reviews, embedding release gate snippets in PRs DO NOT use when: weekly status updates - use github-initiative-pulse. DO NOT use when: code reviews - use pensive review skills. Standardizes release approvals with GitHub-aware checklists.

Retrospective workflow evaluation and improvement of skills, agents, commands, and hooks. Triggers: workflow improvement, retrospective, workflow efficiency Use when: workflow felt slow, confusing, or needs optimization DO NOT use when: implementing features - focus on feature work first.

Structured file enumeration and content analysis for understanding codebase structure before reviews or refactoring. Triggers: file analysis, codebase structure, directory mapping, hotspot detection, code exploration, file enumeration, structure mapping, module boundaries Use when: before architecture reviews to understand file organization, exploring unfamiliar codebases to map structure, estimating scope for refactoring or migration DO NOT use when: general code exploration - use the Explore agent. DO NOT use when: searching for specific patterns - use Grep directly. Provides structural context for downstream review and refactoring workflows.

Update and maintain tests following TDD/BDD principles with detailed quality assurance. Triggers: test updates, test maintenance, test generation, TDD workflow, BDD patterns, test coverage, pytest, test enhancement, quality assurance Use when: updating existing tests, generating new tests for features, enhancing test quality, ensuring detailed coverage, pre-commit validation DO NOT use when: auditing test suites - use pensive:test-review. DO NOT use when: writing production code - focus on implementation first. Run git-workspace-review first to understand which tests need updates.