How do I manage the context between different agents?

The skill provides access to context management patterns, including retrieval and summarization strategies, to ensure information is passed efficiently between agents.

What kind of validation does this skill provide?

It includes specific shell scripts to validate agent markdown files and check that frontmatter metadata is correctly formatted and complete.

How does this skill help with security and permissions?

It follows a least-privilege access strategy, providing guidelines and templates to ensure each sub-agent only has access to the tools strictly necessary for its role.

When should I use a sub-agent instead of a standard skill?

Sub-agents should be used for high-complexity tasks (scoring 7-8 on the complexity framework) that require distinct roles, specific model settings, or isolated tool access. Simpler tasks are better handled by skills.

Does it support different types of agent autonomy?

Yes, it includes templates for various levels of autonomy, including Level 3 autonomous agents, read-only agents, and agents with specific skill/hook integrations.

Implement Sub-Agents

Name: Implement Sub-Agents
Author: yu-iskw

byyu-iskw

0•

开发者工具

Creates and validates specialized Claude sub-agents with precise role definitions, tool permissions, and delegation boundaries.

This skill provides a structured framework for building and managing multi-agent architectures within Claude Code. It streamlines the creation of specialized sub-agents by offering standardized templates for various complexity levels, ranging from read-only assistants to fully autonomous agents. By enforcing least-privilege tool access and clear delegation patterns, the skill ensures that complex tasks are decomposed effectively while maintaining robust context management and evaluation strategies. It includes built-in validation scripts to verify frontmatter and file structures, aligning your implementation with Anthropic's recommended best practices for agentic workflows.

主要功能

01Granular tool configuration guidance based on least-privilege principles

02Standardized templates for autonomous, read-only, and skill-based sub-agents

03Decision framework for determining when to use sub-agents versus skills

040 GitHub stars

05Automated validation scripts for sub-agent frontmatter and file integrity

06Comprehensive evaluation planning for success and failure mode testing

使用场景

01Architecting a multi-agent system to handle complex software engineering tasks

02Standardizing agent definitions across a development team using shared markdown templates

03Restricting agent authority through explicit permission boundaries and tool limits

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yu-iskw/claude-code-plugin-template implement-sub-agents

For use in Claude.ai and ChatGPT

主要功能

01Granular tool configuration guidance based on least-privilege principles

02Standardized templates for autonomous, read-only, and skill-based sub-agents

03Decision framework for determining when to use sub-agents versus skills

040 GitHub stars

05Automated validation scripts for sub-agent frontmatter and file integrity

06Comprehensive evaluation planning for success and failure mode testing

使用场景

01Architecting a multi-agent system to handle complex software engineering tasks

02Standardizing agent definitions across a development team using shared markdown templates

03Restricting agent authority through explicit permission boundaries and tool limits