How does this skill integrate with my existing git workflow?

It analyzes your git diffs to identify which subsystems were touched, ensuring the reflection is grounded in actual file changes and verified invariants.

When should I trigger the reflection skill?

You should use it whenever you finish a task, find yourself saying 'what did we learn?', or want to capture a specific friction point encountered during coding.

What is a 'frontier-shifting' learning in this context?

A frontier-shifting learning is an improvement that enhances your underlying engineering capability, allowing both the volume and quality of work to increase simultaneously rather than trading one for the other.

Where are the reflection artifacts stored?

The skill generates structured Markdown files within a dedicated './compound-discoveries/' directory, creating a permanent audit trail of your project's evolution.

Engineering Reflection Craft

Name: Engineering Reflection Craft
Author: mberto10

bymberto10

•

生産性とワークフロー

Transforms engineering sessions into structured, actionable learnings and discovery artifacts to improve development cycles.

Engineering Reflection Craft provides a systematic framework for post-session debriefing within the Claude Code environment. By analyzing touched subsystems, git diffs, and friction points, it distills technical work into testable rules, features, and specification updates. This skill is designed to power a 'compound engineering' cycle, ensuring that every coding session contributes to a higher-quality codebase and improved agent capabilities by shifting the productivity frontier through structured reflection.

主な機能

01Subsystem-aware debriefing based on session history and git diffs

021 GitHub stars

03Automated creation of structured discovery artifacts in Markdown

04Frontier-check logic to prioritize high-impact architectural improvements

05Categorization of insights into rules, features, fixes, and spec updates

06Distillation of technical friction into one-line actionable and testable rules

ユースケース

01Documenting discovered subsystem dependencies and spec gaps during debugging

02Building a permanent repository of technical learnings to prevent recurring engineering friction

03Capturing architectural insights and new invariants after a complex refactoring session

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mberto10/mberto-compound compound-engineering-reflection-craft

For use in Claude.ai and ChatGPT

主な機能

01Subsystem-aware debriefing based on session history and git diffs

021 GitHub stars

03Automated creation of structured discovery artifacts in Markdown

04Frontier-check logic to prioritize high-impact architectural improvements

05Categorization of insights into rules, features, fixes, and spec updates

06Distillation of technical friction into one-line actionable and testable rules

ユースケース

01Documenting discovered subsystem dependencies and spec gaps during debugging

02Building a permanent repository of technical learnings to prevent recurring engineering friction

03Capturing architectural insights and new invariants after a complex refactoring session

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mberto10/mberto-compound compound-engineering-reflection-craft

For use in Claude.ai and ChatGPT