executing-ai-development-workflow

What problem does it solve?

Managing a high-quality, AI-driven development workflow with multiple review stages, automated fixes, and human oversight is complex. This Skill provides a structured 10-step workflow, ensuring robust implementation and quality assurance for new features and refactorings.

Core Features & Use Cases

• Multi-layer Review: Integrates reviews from multiple Sub-agents, Claude Code /review, and CodeRabbit CLI for comprehensive feedback.
• Priority-Based Fixes: Automatically addresses Critical/High priority issues, while seeking human confirmation for Medium priority items.
• Full Documentation: Saves plans and review results to _docs/ for transparency and traceability.
• Use Case: When implementing a security-critical feature, this Skill ensures a rigorous development process, from detailed planning and Sub-agent-driven implementation to multi-layered security audits and automated fixes, culminating in a well-reviewed Pull Request.

Quick Start

Example: Implement a new user authentication feature

"AI開発ワークフローでユーザー認証機能を実装して"

Claude will initiate the 10-step workflow, starting with planning, then implementation by Sub-agents, multi-layer review, automated fixes, and PR creation.