skills.rest
TorchedHat
TorchedHatOFFICIAL
Explore all skills by this author →
GitHub

agent-writer

Create and update Claude Code agent definition files with YAML frontmatter and validation.

6|8|Updated May 7, 2026
Search Tags:#claude code#subagents#delegation#multi-agent systems#agent design#yaml frontmatter#workflow validation
One-click install
npx skills add https://github.com/TorchedHat/ai-marketplace --skill agent-writer-torchedhat
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-writer
Source: https://github.com/TorchedHat/ai-marketplace/tree/main/ai-writer/skills/agent-writer
Command: npx skills add https://github.com/TorchedHat/ai-marketplace --skill agent-writer-torchedhat

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you create and maintain well-structured Claude Code agent definitions without overlooking essential frontmatter, delegation, tooling, workflow, or validation requirements.

Core Features & Use Cases

  • Agent Definition Design: Structure focused agents with clear identities, scopes, workflows, deliverables, and guardrails.
  • Configuration Guidance: Choose appropriate tools, models, skills, scopes, callable agents, and handoff patterns using least-privilege principles.
  • Validation and Troubleshooting: Check YAML frontmatter, automatic delegation triggers, context isolation, structured outputs, and common discovery problems.
  • Use Case: Design a read-only PyTorch code-review agent, define its tool allowlist and response format, and validate that it can be discovered and invoked reliably.

Quick Start

Use the agent-writer skill to create a focused Claude Code agent for reviewing PyTorch changes with read-only tools and structured findings.

Frequently Asked Questions about agent-writer

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

FAQPage Schema
How do I create a Claude Code subagent with valid YAML frontmatter?▼

To create a Claude Code subagent, structure your agent definition files with valid YAML frontmatter, focused instructions, explicit deliverables, and least-privilege tool access to ensure reliable automatic delegation and discovery.

What is the best way to design multi-agent systems for task delegation?▼

Designing multi-agent systems requires defining focused agent identities, explicit handoff protocols, context isolation, and validation procedures to ensure reliable task delegation and orchestration across callable agents.

Can I configure least-privilege tool access for Claude Code agents?▼

Yes, you can configure least-privilege tool access by specifying a strict tool allowlist in the agent definition, ensuring the subagent only has permissions necessary for its focused scope and structured outputs.

How do I troubleshoot automatic delegation triggers in multi-agent workflows?▼

Troubleshooting automatic delegation triggers involves validating your YAML frontmatter, checking context isolation boundaries, verifying handoff protocols, and ensuring the agent's scope and identity match the intended discovery workflows.

Does agent design require explicit guardrails and validation procedures?▼

Yes, reliable agent design requires explicit guardrails and validation procedures to enforce structured outputs, maintain context isolation, and ensure subagents operate safely within their defined scopes and tool allowlists.

Why does my Claude Code subagent fail to be discovered and invoked?▼

Subagent discovery fails when YAML frontmatter is invalid, delegation triggers are misconfigured, or the agent lacks explicit scoping and deliverables required for reliable invocation within multi-agent orchestration workflows.

Related Skills

Explore More Software Engineering Skills→
iiammaeiiammae

designing-workflow-skills

Design workflow-based Claude Code skills with structured patterns and progressive disclosure.

Community
Advanced
suhwansuhwan

orchestrate

Coordinate multiple AI subagents with dependency-aware execution and result aggregation.

Community
Advanced
zuzu59zuzu59

security

Enforce security best practices for AI agent operations.

Community
Intermediate
ybbms777ybbms777
1

setup

Automate review agent selection for compound-engineering projects using ce:review.

Community
Basic