How does the QC process work for new sub-agents?

The skill delegates quality control to a specialized 'qa-skill-qc' sub-agent which evaluates the new agent's criteria, skill settings, and I/O definitions against a standardized design guide.

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

You should generate a sub-agent when a task is context-independent, requires specialized domain knowledge, needs to perform web searches asynchronously, or involves parallel brainstorming and validation.

What is a sub-agent in Claude Code?

A sub-agent is a specialized configuration designed to handle a specific task or domain independently, often allowing for parallel execution and specialized tool use without cluttering the primary agent's context.

What types of agents can this skill create?

It provides optimized templates for Research-agents (data collection), Feedback-agents (reviews), Ideation-agents (brainstorming), Validation-agents (feasibility), and Parallel-check-agents (multi-file verification).

Subagent Architect

Name: Subagent Architect
Author: matsuni-kk

bymatsuni-kk

•

생산성 및 워크플로

Generates and orchestrates specialized, domain-specific sub-agents to handle complex parallel tasks and expert reviews.

The Subagent Architect skill automates the creation of specialized Claude sub-agents, enabling parallel processing and domain-specific expertise within your development workflow. By analyzing existing project skills and requirements, it identifies opportunities for task delegation and generates structured agent definitions in the .claude/agents/ directory. This skill is essential for scaling complex projects that require independent research, multi-perspective brainstorming, or rigorous technical validation without overloading the main agent's context window.

주요 기능

01Template-based generation for Research, Feedback, and Ideation agents

028 GitHub stars

03Domain-specific skill mapping and 'portable skill' assignment

04Automated sub-agent feasibility assessment

05Standardized markdown output for seamless agent integration

06Integrated Quality Control (QC) loop using specialized validation agents

사용 사례

01Generating specialized feedback agents to provide objective code reviews based on specific quality criteria

02Designing independent ideation agents to brainstorm multiple architectural approaches simultaneously

03Creating parallel research agents to perform asynchronous web searches and documentation audits

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add matsuni-kk/agent_template_public subagent-generate

For use in Claude.ai and ChatGPT

Download Skill