writing-agents

Design Claude agents with standardized frontmatter, templates, and coordination guidelines.

125|16|Updated Feb 6, 2026
One-click install
npx skills add https://github.com/aaddrick/claude-pipeline --skill writing-agents-aaddrick
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: writing-agents
Source: https://github.com/aaddrick/claude-pipeline/tree/main/.claude/skills/writing-agents
Command: npx skills add https://github.com/aaddrick/claude-pipeline --skill writing-agents-aaddrick

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Writing and maintaining Claude agents can be error-prone without standardized templates. This skill provides canonical frontmatter, templates, and coaching on persona, scope, anti-patterns, and coordination to ensure agents behave consistently across projects.

Core Features & Use Cases

  • Agent templates for common specialist, orchestrator, and reviewer styles to accelerate creation.
  • Explicit persona and scope guidance to prevent role creep and ensure safe, predictable behavior.
  • Project context and coordination protocols to enable multi-agent collaboration and deferral rules.

Quick Start

Use this skill to design a new agent by copying a template from .claude/skills/writing-agents/agent-templates.md, populate the frontmatter, and create the agent file under .claude/agents/.md. Then restart Claude Code to load the agent.

Frequently Asked Questions about writing-agents

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

FAQPage Schema
How do I create Claude agents with standardized frontmatter and templates?

To create Claude agents with standardized frontmatter, copy a specialist, orchestrator, or reviewer template from the agent-templates.md file, populate the frontmatter fields, and save the agent file under the .claude/agents/ directory. Restart Claude Code to load the new agent.

What is the best way to prevent role creep in multi-agent Claude workflows?

The best way to prevent role creep in multi-agent Claude workflows is to define explicit persona and scope boundaries within each agent's frontmatter, establishing anti-patterns and deferral rules for safe, predictable behavior across projects.

How do you coordinate multiple Claude agents in a single project?

You coordinate multiple Claude agents by establishing project context and coordination protocols, defining deferral patterns that specify how agents interact and hand off tasks within the .claude/agents and .claude/skills directories.

When do I need to use deferral patterns in Claude agent design?

You need deferral patterns in Claude agent design when building scalable multi-agent workflows, allowing specialized agents to hand off specific tasks to other agents rather than handling out-of-scope actions themselves.

Can I use custom agent templates for specific project contexts in Claude Code?

Yes, you can use custom agent templates for specific project contexts by documenting project context within the agent files and referencing existing templates to ensure agents behave consistently across your projects.