Does this skill require manual data entry?

No, it automatically analyzes the session transcript and provides a comprehensive report without requiring the user to manually track events.

What metrics does this skill evaluate?

It scores dimensions such as tool precision, iteration efficiency, context utilization, and autonomy span, providing both narrative feedback and structured data.

What is the learn-from-session Claude Code skill?

It is a specialized tool that analyzes your current AI coding session transcript to provide a detailed review of agent performance, tool precision, and adherence to project standards.

How does it improve future AI coding sessions?

It identifies specific points of friction and suggests actionable updates for your CLAUDE.md or AGENTS.md files, helping the AI avoid making the same mistakes twice.

Can the session review be shared with my team?

Yes, the skill includes built-in commands to post the full session analysis, including a structured JSON block, as a comment on an active GitHub Pull Request.

Session Analytics & Improvement

Name: Session Analytics & Improvement
Author: ClipboardHealth

byClipboardHealth

•

生産性とワークフロー

Analyzes Claude Code sessions to optimize agent efficiency, code quality, and project-specific automation rules.

The learn-from-session skill acts as a retrospective agent for your AI coding sessions, evaluating how effectively Claude utilized tools, followed project standards, and maintained autonomy. By identifying friction points and successful patterns, it generates actionable improvements such as CLAUDE.md updates and workflow optimizations. This skill is essential for teams looking to refine their AI development environment by turning every interaction into a learning opportunity that reduces future human intervention and improves output quality.

主な機能

01Automated generation of ready-to-paste diffs for updating project instructions.

02Detailed compliance auditing against CLAUDE.md and project-specific coding standards.

03Structured JSON data emission for tracking agent performance metrics over time.

04Direct integration with GitHub PRs to share session insights and improvements with the team.

0530 GitHub stars

06Quantitative efficiency scoring for tool usage, autonomy levels, and iteration cycles.

ユースケース

01Onboarding new team members by documenting discovered code patterns and environment quirks.

02Refining project instructions by identifying recurring agent errors or context gaps.

03Auditing agent autonomy during complex refactoring tasks to identify workflow bottlenecks.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add clipboardhealth/core-utils learn-from-session

For use in Claude.ai and ChatGPT

Download Skill