Where does the skill store captured information?

Findings are routed based on their scope to locations such as CLAUDE.md (project-wide), .claude/rules/ (scoped), or ~/.claude/CLAUDE.md (user global).

What is feedback compaction?

Feedback compaction is a feature that summarizes long-term session history into aggregate statistics, allowing the skill to understand which suggestions you typically accept or skip over time.

How does Dev Wrap-Up improve my coding sessions?

It identifies patterns, corrections, and new conventions from your current session and suggests saving them into your project's documentation, ensuring the AI adapts to your specific requirements.

Will this skill overwrite my files automatically?

No, the skill operates in a review-only mode. It presents findings for your review and requires explicit confirmation before writing any changes to your files.

When should I run the dev-wrapup skill?

You should run this skill at the end of a working session, after completing features or bug fixes, to ensure all relevant learnings and corrections are captured.

Dev Session Wrap-Up

Name: Dev Session Wrap-Up
Author: andreaserradev-gbj

byandreaserradev-gbj

•

生产力与工作流

Analyzes development sessions to extract and persist project conventions, technical gotchas, and personal preferences into documentation.

Dev Session Wrap-Up is a specialized skill for Claude Code designed to capture and institutionalize knowledge gained during active development sessions. It systematically scans conversation history for assistant corrections, architectural decisions, and workflow friction points, then routes these insights to the appropriate documentation files like CLAUDE.md or scoped rules. By leveraging a feedback compaction system to track long-term user preferences, this skill ensures that the AI's behavior evolves alongside the project, reducing repetitive errors and maintaining high alignment with team-specific standards across multiple sessions.

主要功能

012 GitHub stars

02Automated feedback compaction for long-term preference tracking

03Smart quality filtering to ignore session-specific noise and duplicates

04Intelligent conversation scanning for technical corrections and patterns

05Strict review-only mode requiring user confirmation for all changes

06Multi-tier routing to project-wide, scoped, or global documentation

使用场景

01Capturing project-specific naming conventions and architectural patterns discovered during implementation

02Recording personal preferences and 'do-not-do' lists to ensure consistent AI behavior in future sessions

03Persisting complex workarounds for environment-specific issues to prevent future debugging time

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add andreaserradev-gbj/dev-workflow dev-wrapup

For use in Claude.ai and ChatGPT

主要功能

012 GitHub stars

02Automated feedback compaction for long-term preference tracking

03Smart quality filtering to ignore session-specific noise and duplicates

04Intelligent conversation scanning for technical corrections and patterns

05Strict review-only mode requiring user confirmation for all changes

06Multi-tier routing to project-wide, scoped, or global documentation

使用场景

01Capturing project-specific naming conventions and architectural patterns discovered during implementation

02Recording personal preferences and 'do-not-do' lists to ensure consistent AI behavior in future sessions

03Persisting complex workarounds for environment-specific issues to prevent future debugging time