Analyzes development sessions to automatically improve your AI toolkit, documentation, and workflow automation.
Reflect is a meta-improvement skill designed to optimize your development environment by reviewing past sessions and identifying patterns of friction. It helps automate repetitive tasks, update documentation like AGENTS.md, refine permission settings, and capture institutional knowledge. By transforming manual discoveries into permanent system improvements, it ensures that your Claude Code experience becomes more efficient and tailored to your specific project needs over time through structured post-session analysis.
Key Features
01Streamlined toolkit updates for AGENTS.md and local skills
020 GitHub stars
03Identification of permission gaps and documentation needs
04Persistence of learnings via session summaries in history/reflections/
05Prioritized improvement suggestions (P0-P3)
06Automated session analysis and pattern recognition
Use Cases
01Automating multi-step workflows by proposing new specialized skills based on observed session history.
02Performing a post-sprint review to document architectural patterns and edge cases discovered during coding.
03Reducing repetitive manual approvals by identifying and whitelisting frequently used bash commands.
