Reflect

About

Reflect provides a structured framework for continuous improvement within the Claude Code environment by analyzing conversation signals such as user corrections, successes, and edge cases. It utilizes a validated workflow—including a specialized critic subagent—to suggest high-confidence updates to your local skills and memories, ensuring your AI assistant evolves and adapts to your specific workflow and project requirements over time.

Key Features

• 0 GitHub stars
• Automated session analysis for feedback and edge case detection
• Persistent memory management for cross-skill preferences
• Performance tracking and effectiveness metrics
• Skill validation against 12-factor agent principles
• Evidence-based improvement proposals with confidence ranking

Use Cases

• Optimizing custom skills based on real-world usage and failures
• Capturing and automating project-specific coding preferences
• Auditing existing skills for architectural compliance and reliability