skills.rest
uv-xiao
uv-xiaoCOMMUNITY
Explore all skills by this author →
Website𝕏@youwei_xiaoGitHub

uv-skill-evolution-manager

Persist session feedback into evolution.json and stitch learnings into SKILL.md.

1|Updated Feb 6, 2026
Search Tags:#automation#skill management#feedback loop#skill improvement#AI evolution#LLM operations
One-click install
npx skills add https://github.com/uv-xiao/pkbllm --skill uv-skill-evolution-manager
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: uv-skill-evolution-manager
Source: https://github.com/uv-xiao/pkbllm/tree/main/bootstrap/skill-evolution-manager
Command: npx skills add https://github.com/uv-xiao/pkbllm --skill uv-skill-evolution-manager

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill helps you continuously improve AI skills based on real-world session feedback, ensuring that valuable learnings are captured and applied without losing them during skill updates.

Core Features & Use Cases

  • Structured Learning Persistence: Captures feedback into evolution.json for structured, reusable learnings.
  • Idempotent Skill Updates: Safely stitches learned sections into SKILL.md, allowing re-application without duplication.
  • Use Case: After a session where a skill repeatedly failed due to a specific platform quirk, you can use this skill to record that quirk and the fix in evolution.json, then apply it to the skill's SKILL.md so future users are warned or the skill behaves correctly.

Quick Start

Apply the latest learnings from the provided JSON to the 'uv-hands-on-learning' skill.

Frequently Asked Questions about uv-skill-evolution-manager

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I persist AI skill improvements from session feedback without losing them during updates?▼

To persist AI skill improvements safely, structured learnings from session feedback are captured into an evolution.json file, ensuring valuable insights survive skill updates without arbitrary loss. This approach anchors changes in evidence.

What is the best way to apply new learnings to SKILL.md without creating duplicate entries?▼

The best way to apply new learnings to SKILL.md without duplicates is using idempotent stitching, which safely inserts feedback sections and allows re-application without duplication. This ensures clean, non-arbitrary prompt injection.

Can I update local installed skill copies or do I need to modify the canonical PKB_PATH skills directly?▼

You can update both canonical PKB_PATH skills and local installed skill copies. The evolution manager supports applying structured learnings to either target, ensuring flexibility across different development and deployment environments.

Do I need global installs to manage the safe evolution of AI skills?▼

No, you do not need global installs to manage the safe evolution of AI skills. The process handles structured learning persistence and idempotent updates locally, focusing on evidence-first application without requiring global dependencies.

Why does my AI skill repeatedly fail on specific platform quirks after sessions?▼

AI skills fail on platform quirks when session learnings are not persisted. By recording the quirk and its fix into evolution.json and stitching it into SKILL.md, future sessions are warned or corrected automatically.

Related Skills

Explore More Software Engineering Skills→
shuff57shuff57

session-end-wrap

Orchestrate end-of-session workflows with memory pruning and skill evolution.

Community
Advanced
danger-Englishdanger-English
2

skill-evolution-manager

Record runtime errors and improvements into evolution.json and SKILL.md.

Community
Intermediate
bahayonghangbahayonghang
16

skill-evolution-manager

Convert conversation feedback into structured evolution JSON for Skill improvements.

Community
Advanced
Yaron9Yaron9
18

skill-creator

Automate Claude Code skill creation, evolution, and .skill packaging.

Community
Advanced