How does the highlight skill improve future coding tasks?

It persists 'lessons learned' into a structured library, allowing the AI to reference successful patterns and avoid repeating past mistakes in subsequent development sessions.

Can I manually save an insight during a chat?

Yes, you can use the command '/task-ai:highlight' followed by a description of what you want to save to trigger an ad-hoc experience capture.

Does this skill require manual Git commits?

No, the highlight skill handles atomic writes and automatically generates standardized Git commit messages for every piece of knowledge it persists.

What is the Experience-to-Skill promotion feature?

It is an automated pipeline that monitors your knowledge library for high-quality, frequently used patterns and converts them into new, executable Claude Code skills.

What is semantic naming in the context of this skill?

Instead of using task IDs, it organizes experiences by their knowledge domain (e.g., 'react-virtual-scroll'), ensuring all related insights are consolidated into single, high-value files.

Experience Distiller & Knowledge Engine

Name: Experience Distiller & Knowledge Engine
Author: huacheng

byhuacheng

•

生産性とワークフロー

Captures, distills, and persists reusable coding experiences and thinking patterns into a searchable, structured knowledge library.

The highlight skill serves as a sophisticated long-term memory system for AI-driven development workflows. It systematically extracts lessons learned during the development lifecycle, organizes them using semantic naming to prevent knowledge fragmentation, and maintains a searchable index of insights. By distilling raw conversations and task outcomes into high-quality documentation, it ensures that hard-won architectural decisions and bug fixes become permanent assets that inform future AI logic and skill generation.

主な機能

01Structured 'naming-index' maintenance for deterministic knowledge retrieval

021 GitHub stars

03Automated comprehensive experience distillation upon task completion

04Semantic naming resolution to prevent duplicate knowledge domains

05Experience-to-Skill promotion pipeline for automating AI capability growth

06Ad-hoc conversation insight capture via natural language triggers

ユースケース

01Automatically drafting new Claude Code skills based on repeated implementation patterns

02Capturing a complex bug-fix logic to prevent regression in future related tasks

03Building a team-wide library of architectural patterns and best practices

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add huacheng/moonview highlight

For use in Claude.ai and ChatGPT