How much overhead does this skill add?

This skill typically adds 2-3 hours of overhead per experiment but is designed to save dozens of hours in the long run through optimized workflow performance.

What is the difference between Universal and Domain agents?

Universal agents (like coders) are 100% transferable across any project, while Domain agents (like coverage-analyzers) are 60-80% transferable within specific fields like testing or security.

What are Meta-Agent (M0) capabilities?

M0 capabilities are the core orchestration functions including observation, planning, execution, reflection, and evolution that manage the specialized agent set.

When should I create a specialized agent?

According to the skill's decision tree, you should specialize when a generic agent shows a performance gap of at least 5x and the task is recurring enough to justify the creation cost.

Agent Prompt Evolution

Name: Agent Prompt Evolution
Author: yaleh

byyaleh

•

生産性とワークフロー

Tracks and optimizes the emergence of specialized AI agents to improve workflow performance and prompt reusability.

概要

Agent Prompt Evolution is a meta-cognition framework for Claude Code designed to systematically monitor how generic agents evolve into specialized experts during methodology development. It helps developers identify performance gaps—often exceeding 5x speedups—to justify the creation of specialized sub-agents. By tracking agent set evolution, calculating ROI on specialization decisions, and categorizing prompts by their transferability, this skill enables the creation of a stable, high-performance AI architecture that matures across multiple experiments.

主な機能

Agent Set Evolution (An) tracking to monitor specialization iterations
Specialization Decision Tree for calculating performance ROI
Reusability assessment for Universal, Domain, and Task-specific agents
Meta-Agent (Mn) capability tracking and optimization
Cross-experiment analysis to build a reusable agent library
15 GitHub stars

ユースケース

Scaling generic AI agents into high-performance specialized workflows
Building a library of reusable agent prompts for cross-domain projects
Systematically improving the base meta-capabilities of AI orchestrators

概要

主な機能

Agent Set Evolution (An) tracking to monitor specialization iterations
Specialization Decision Tree for calculating performance ROI
Reusability assessment for Universal, Domain, and Task-specific agents
Meta-Agent (Mn) capability tracking and optimization
Cross-experiment analysis to build a reusable agent library
15 GitHub stars

ユースケース

Scaling generic AI agents into high-performance specialized workflows
Building a library of reusable agent prompts for cross-domain projects
Systematically improving the base meta-capabilities of AI orchestrators