skills.rest
LumenWipe
LumenWipeOFFICIAL
Explore all skills by this author →
WebsiteGitHub

optimize-skills

Analyze local usage traces to refine installed skill triggers and instructions.

3|1|Updated May 12, 2026
Search Tags:#prompt optimization#skill optimization#skill evaluation#few-shot examples#usage traces#local learning loop
One-click install
npx skills add https://github.com/LumenWipe/lumenwipe --skill optimize-skills-lumenwipe
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: optimize-skills
Source: https://github.com/LumenWipe/lumenwipe/tree/main/.claude/skills/optimize-skills
Command: npx skills add https://github.com/LumenWipe/lumenwipe --skill optimize-skills-lumenwipe

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill turns real-world skill usage traces into sharper, more reliable installed skills without sending personal data outside the local environment.

Core Features & Use Cases

  • Trace Analysis: Groups local JSONL records into multi-turn skill usage spans and evaluates their outcomes.
  • Skill Compilation: Improves trigger descriptions, operational instructions, and learned examples using successful usage patterns.
  • Safe Reversibility: Creates timestamped backups before editing skills and supports restoring the previous version.
  • Profile Refreshing: Distills durable user preferences from traces into a concise local profile for future interactions.
  • Use Case: After repeatedly using several skills, run this Skill to identify successful patterns, refine the installed instructions, and preserve the original files for rollback.

Quick Start

Use the optimize-skills skill to analyze my local usage traces, improve eligible installed skills, refresh my learning profile, and report the changes.

Frequently Asked Questions about optimize-skills

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

FAQPage Schema
How do I refine installed AI skills from local usage traces?▼

To refine installed AI skills from local usage traces, you need a tool that analyzes multi-turn JSONL records, evaluates outcomes, and updates trigger descriptions. This skill compiles successful patterns into improved instructions locally.

Can I extract few-shot examples from previous prompt usage spans?▼

You can extract few-shot examples from previous prompt usage spans by analyzing local interaction history. This skill mines successful work products and generalized examples from at least five eligible usage spans to improve future triggers.

Does skill optimization require sending my local prompt data outside the environment?▼

Skill optimization does not require sending local prompt data outside the environment. This skill analyzes local JSONL records and installed skill files directly on your machine, ensuring personal usage traces remain completely private.

What happens to my original files when an optimization process edits the instructions?▼

When an optimization process edits the instructions, your original files are preserved using timestamped backups. This skill maintains local backup copies and optimization records, allowing you to safely restore the previous version if needed.

How many usage spans are needed before an installed skill becomes eligible for optimization?▼

Five usage spans are needed before an installed skill becomes eligible for optimization. This skill requires a minimum of five recorded multi-turn interaction spans to accurately evaluate outcomes and mine generalized examples for improvements.

How do I refresh durable user preferences from past AI interactions?▼

To refresh durable user preferences from past AI interactions, you analyze local usage traces to distill consistent behavioral patterns. This skill compiles those preferences into a concise local profile for future interactions.

Related Skills

Explore More Software Engineering Skills→
rysweetrysweet
1

session-replay

Analyze claude-trace JSONL files for session health, token usage, and error patterns.

Community
Intermediate
eddie-292eddie-292
52

skill-creator

Create, test, and iterate on Claude skills with evaluation workflows.

Community
Advanced
OpenCovenOpenCoven
20

daily-dev-agentic

Automate daily.dev feed creation, tag following, and scheduled learning loops.

Official
Intermediate
memorysavermemorysaver
21

skill-capability-matcher

Match user requirements to available Looplia skills and export a structured data-flow specification.

Community
Advanced