Optimizes AI-human collaboration by identifying and resolving workflow friction through post-feature reflection and history analysis.
The Retro skill provides a structured framework for Claude to analyze recent development sessions and identify recurring friction points. By evaluating interrupted actions, requested clarifications, and context slips, the skill proposes high-leverage refinements to project documentation, custom skills, and communication norms. It features an autonomous history search capability to find patterns across multiple sessions, ensuring that process improvements are based on actual evidence rather than speculation, ultimately reducing future cognitive overhead and token usage.
Características Principales
01Evidence-based workflow refinement proposals
02Configurable Lite and Full retrospective modes
035 GitHub stars
04Automated session history analysis via subagents
05Context-aware documentation and skill updates
06Friction signal identification and assessment
Casos de Uso
01Reducing repetitive clarifications during long development sessions
02Improving AI alignment after a complex feature implementation
03Streamlining project-specific communication norms in CLAUDE.md