What does the Retro skill do?

The Retro skill analyzes your active coding session to identify potential improvements, capture key lessons, and document critical decisions in a structured format.

What is the difference between Deep and Light mode?

Deep mode uses parallel sub-agents for a 7-section thorough analysis including expert feedback, while Light mode provides a faster, streamlined summary for routine or simple tasks.

Does this skill help with technical debt?

Yes, specifically through its 'Waste Reduction' and 'CDM Analysis' sections, which help identify inefficient patterns and better alternatives for future work.

How do I trigger a session retrospective?

You can trigger the skill by typing /retro, 'retrospective', or 'cwf:retro' in your Claude Code interface. Use the --light flag for a quick review.

Where are the retrospective results saved?

Retro outputs are saved as retro.md, plan.md, and lessons.md within your project's designated session artifact directory (usually under .cwf/projects/).

Session Retrospective (Retro)

Name: Session Retrospective (Retro)
Author: corca-ai

bycorca-ai

•

Productivity & Workflow

Conducts comprehensive session reviews to transform recent coding outcomes into durable process improvements and lessons learned.

Retro is a specialized Claude Code skill designed to facilitate rigorous end-of-session analysis, helping developers capture insights and optimize their future workflows. It features an adaptive depth system that can execute a deep-dive analysis of critical decisions, waste reduction, and expert-level architectural reviews, or a lightweight summary for routine tasks. By generating structured artifacts such as retro.md and lessons.md, it ensures that every session contributes to long-term project intelligence, team efficiency, and continuous delivery improvement.

Key Features

01Multi-agent analysis including Critical Decision Method (CDM) and Expert Lens reviews

02Automated evidence collection from session logs, file diffs, and live state

03Seamless integration with session artifact directories for project-wide tracking

04Adaptive analysis modes (Deep vs. Light) tailored to session complexity

05Automatic generation of persistent documentation and learning resources

0673 GitHub stars

Use Cases

01Summarizing key takeaways and external learning resources for team knowledge sharing

02Capturing technical debt and process waste after a complex feature implementation

03Synthesizing expert-level feedback on critical architectural decisions made during coding

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add corca-ai/claude-plugins retro

For use in Claude.ai and ChatGPT

Download Skill