Verified:GitHub

evolve-system

Analyzes system health and generates actionable recommendations for improving agent performance and efficiency.

1|1|Updated Jan 22, 2026

Search Tags:#prompt engineering #root cause analysis #ai optimization #performance analysis #system evolution #agent tuning

Ready to run

npx skills add https://github.com/csuzngjh/principles --skill evolve-system

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: evolve-system
Source: https://github.com/csuzngjh/principles/tree/main/claude/skills/evolve-system
Command: npx skills add https://github.com/csuzngjh/principles --skill evolve-system

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill addresses the need for continuous improvement and self-optimization within AI systems by analyzing performance metrics and identifying systemic issues.

Core Features & Use Cases

Performance Analysis: Calculates agent win rates and identifies recurring error patterns from issue logs.
Root Cause Diagnosis: Pinpoints the source of inefficiencies in agent prompts, hooks, or rules.
Optimization Proposals: Generates concrete suggestions for improving system components.
Use Case: An AI agent is consistently failing a specific type of task. This Skill analyzes its performance, identifies a vague prompt as the root cause, and proposes a revised prompt to improve success rates.

Quick Start

Run the evolve-system skill to analyze system performance and propose optimizations.

Frequently Asked Questions about evolve-system

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

FAQPage Schema

How do I optimize AI agent prompts to improve task success rates?▼

AI agent prompt optimization is achieved by analyzing performance metrics and issue logs to pinpoint systemic issues. This process calculates win rates, identifies recurring error patterns, and proposes revised prompts to improve task success rates.

What is second-order observation in AI system evolution?▼

Second-order observation in AI system evolution is a mechanism that analyzes agent performance metrics and diagnostic logs to evaluate and propose optimizations for system-level components like hooks and rules.

How do I find the root cause of recurring errors in AI agents?▼

Root cause analysis for AI agents is performed by parsing issue logs and performance scorecards to diagnose inefficiencies. It identifies vague prompts or flawed rules as the source of recurring error patterns and generates optimization proposals.

Can I automatically modify agent rules in the .claude directory?▼

Modifying agent rules in the .claude directory requires explicit user confirmation. The system performs analysis and proposes concrete optimizations for agents, hooks, and rules, but it does not apply system file modifications automatically.

What format do performance metrics need to be in for AI performance analysis?▼

AI performance analysis requires metrics to be stored in AGENT_SCORECARD.json and issue logs in ISSUE_LOG.md. These specific file formats are parsed to calculate agent win rates and diagnose systemic issues.

Does agent tuning work for fixing hooks and rules within the .claude directory?▼

Agent tuning works for hooks and rules within the .claude directory by diagnosing system inefficiencies from issue logs. It evaluates performance metrics and proposes targeted optimizations to improve the behavior of these system components.