utilities

Create Architecture Decision Records (ADR) in MADR format with Skills integration. Records architecture decisions with context and rationale. Auto-numbering (0001, 0002, ...), saves to docs/adr/.

Generate project-specific skill from ADR for context-aware implementation guidance. Creates skill with references, examples, and validation. Saves to .claude/skills/, enables automatic discovery. Provides implementation guidance based on architecture decisions. Use when ADR needs to guide implementation with comprehensive context and examples. ADRからプロジェクト固有のスキルを生成し、コンテキストに応じた実装ガイダンスを提供。

Automatically execute tests after file modifications and invoke /fix command if tests fail using SlashCommand tool. Streamlines test-fix cycle with automatic conditional execution. Can be triggered via hooks in settings.json. Use after code changes to verify functionality and automatically attempt fixes on failure. ファイル変更後に自動的にテストを実行し、失敗時に/fixコマンドを呼び出す。

Analyze current Git changes and suggest appropriate branch names following conventional patterns (feature/fix/chore/docs). Uses branch-generator agent to analyze diff and file patterns. Provides 3-5 naming suggestions with rationale. Use before creating a new branch when you need help with naming conventions. Git差分を分析して適切なブランチ名を自動生成。慣習的なパターン（feature/fix/chore/docs）に従う。

Implement code following TDD/RGRC cycle (Red-Green-Refactor-Commit) with real-time test feedback and quality checks. Use for feature implementation, refactoring, or bug fixes when you have clear understanding (≥70%) of requirements. Applies SOLID principles, DRY, and progressive enhancement. Includes dynamic quality discovery and confidence scoring.

This module defines completion criteria and decision support.

This module handles quality verification during implementation.

This module provides detailed TDD implementation guidance with Baby Steps methodology.

This module handles automatic detection and loading of spec.md for implementation guidance.

This module handles automatic Stories generation from spec.md Component API sections.

This module handles pre-TDD test generation from specifications.

Analyze Git diff and generate Conventional Commits format messages automatically. Uses commit-generator agent. Detects commit type (feat/fix/chore/docs), scope, breaking changes. Focuses on "why" rather than "what". Use after staging changes when ready to commit. Git差分を分析してConventional Commits形式のメッセージを自動生成。型、スコープ、破壊的変更を検出。

Diagnose current context usage and provide token optimization recommendations. Displays token usage, file count, session cost. Helps identify context-heavy operations. Use when context limits are approaching or to optimize session efficiency. 現在のコンテキスト使用状況を診断し、トークン最適化の推奨事項を提供。

Rapidly fix small bugs and minor improvements in development environment with dynamic problem detection and parallel quality verification. Use for well-understood (≥80%) small-scale fixes in development only. NOT for production emergencies (use /hotfix instead). Applies Occam's Razor (simplest solution), Progressive Enhancement (CSS-first), and TIDYINGS (clean as you go). 開発環境で小さなバグの修正や軽微な改善を素早く実行。よく理解された小規模な修正に使用。本番緊急時は/hotfixを使用。

Orchestrate complete development cycle through SlashCommand tool integration, executing from research through implementation, testing, and validation. Chains multiple commands: /research → /think → /code → /test → /review → /validate with conditional execution and error handling. TodoWrite integration for progress tracking. Use for comprehensive feature development requiring full workflow automation.

Create custom hooks to prevent unwanted behaviors

Configure hookify rules interactively

List all hookify rules

Emergency fixes for critical production issues ONLY. For production-impacting problems, security vulnerabilities, or immediate deployment needs. 5-min triage, 15-min fix, 10-min test. Minimal process overhead with required rollback plan. NOT for development fixes (use /fix instead). High severity (critical/security) only. 本番環境の緊急対応が必要な重大な問題を修正。本番影響、セキュリティ脆弱性、即座のデプロイが必要な場合のみ。

Analyze branch changes and generate comprehensive PR description automatically. Uses pr-generator agent. Examines all commits from branch divergence, not just latest. Creates summary, test plan, and checklist. Use when ready to create pull request and need description text. ブランチの変更内容を分析して包括的なPR説明文を自動生成。分岐点からのすべてのコミットを検査。

Perform project research and technical investigation without implementation. Explore codebase structure, technology stack, dependencies, and patterns. Use when understanding is low (≥30%) and you need to learn before implementing. Documents findings persistently for future reference. Uses Task agent for complex searches with efficient parallel execution. プロジェクト理解と技術調査を行う（実装なし）。コードベース構造、技術スタック、依存関係、パターンを探索。

Orchestrate multiple specialized review agents with dynamic context analysis, hierarchical task decomposition, and confidence-based filtering. Use after code changes or when comprehensive quality assessment is needed. Includes security, performance, accessibility, type safety, and more. All findings include evidence (file:line) and confidence markers (✓/→/?) per Output Verifiability principles.

Generate project rules from ADR automatically and integrate with CLAUDE.md. Converts decision into AI-executable format. Saves to docs/rules/, auto-integrates with .claude/CLAUDE.md. Enables AI to follow project-specific decisions. Use when ADR decision should affect AI behavior and enforce project-specific patterns. ADRからプロジェクトルールを自動生成し、CLAUDE.mdに統合。決定内容をAI実行可能形式に変換。

Display current SOW progress status, showing acceptance criteria completion, key metrics, and build status. Read-only viewer for active work monitoring. Lists and views Statement of Work documents stored in workspace. Use to check implementation progress anytime during development. SOW文書の一覧表示と閲覧。受け入れ基準の完了状況、主要メトリクス、ビルドステータスを表示。

Run project tests and validate code quality through comprehensive testing. Automatically discovers test commands from package.json, README, or project configuration. Handles unit, integration, and E2E tests with progress tracking via TodoWrite. Includes browser testing for UI changes when applicable. Use after implementation to verify functionality and quality standards. プロジェクトのテストを実行し、包括的なテストでコード品質を検証。ユニット、統合、E2Eテストに対応。

Create a comprehensive Statement of Work (SOW) for feature development or problem solving. Use when planning complex tasks, defining acceptance criteria, or structuring implementation approaches. Ideal for tasks requiring detailed analysis, risk assessment, and structured planning documentation.

Validate implementation against SOW acceptance criteria with L2 (practical) validation level. Checks acceptance criteria, coverage, and performance. Pass/fail logic with clear scoring. Identifies missing features and issues. Use when ready to verify implementation conformance. SOWの受け入れ基準に対して実装を検証。受け入れ基準、カバレッジ、パフォーマンスをチェック。

Create reusable browser automation workflows using Chrome DevTools MCP through interactive step recording. Triggers interactive workflow builder, executes steps in real browser, saves as discoverable slash command. Use when creating E2E tests, monitoring, or automation workflows.

Diagnose current context usage and provide optimization suggestions. Displays token usage, file counts, session cost estimation. Analyzes context efficiency and suggests improvements. 現在のコンテキスト使用状況を診断し、最適化提案を行う。トークン使用量、ファイル数、セッションコストを表示。