cc-subagents

Prompt file imported from evgenygurin/r2r-fastmcp (.claude/commands/cc-subagents.md). Copyright stays with the author.

Subagents Guide

Comprehensive subagents documentation at @docs/claude_code/06-subagents.md

What are Subagents?

Subagents are specialized Claude instances that run autonomously to complete specific tasks in parallel.

Built-in Subagent Types

Code Analysis

• Explore - Fast codebase exploration and search
• Code Explorer - Deep dive into code structure
• Plan - Design implementation approach

Code Quality

• Code Reviewer - Review code for quality and issues
• Silent Failure Hunter - Find bugs that fail silently
• PR Test Analyzer - Analyze test coverage in PRs

Architecture

• Code Architect - Design system architecture
• general-purpose - Multi-step complex tasks

Key Features

Parallel Execution - Multiple subagents run simultaneously Autonomy - Subagents work independently with their own tools Specialization - Each type optimized for specific tasks Context Sharing - Results returned to main Claude

Instructions

1. Read the full subagents documentation:

Read docs/claude_code/06-subagents.md

2. Focus on explaining:

   • Subagent types and their capabilities
   • When to use each subagent
   • Parallel execution patterns
   • Creating custom subagents
   • Tool access per subagent
   • Best practices and workflows

3. Show practical examples:

   • Codebase exploration
   • Code review workflow
   • Architecture design
   • Parallel task execution

When to Use Subagents

Use Explore:

• "Find all API endpoints"
• "Understand authentication flow"
• Quick codebase searches

Use Plan:

• "Design database schema for users"
• "Plan refactoring approach"
• Architecture decisions

Use Code Reviewer:

• After completing features
• Before commits/PRs
• Quality checks

Use Silent Failure Hunter:

• Debug mysterious issues
• Find edge cases
• Audit error handling

Parallel Execution

Launch multiple subagents simultaneously:

# Claude will create multiple Task tool calls in one message:
Task(subagent_type="explore", prompt="Find auth code")
Task(subagent_type="code-reviewer", prompt="Review security")
Task(subagent_type="plan", prompt="Design API")

All three run in parallel and report back.

Creating Custom Subagents

Custom agent definitions in .claude/agents/:

---
name: my-agent
description: Custom agent for specific task
tools: [Read, Grep, Glob]
---

# Agent Instructions

Your custom agent prompt and instructions...

See .claude/agents/ for examples:

• research-assistant.md
• doc-analyst.md
• knowledge-explorer.md

Subagent vs Agent vs Skill

Subagent - Autonomous subprocess with own tools (via Task tool) Agent - Persistent definitions in .claude/agents/ Skill - Lightweight capability without subprocess

Hooks - Different! Hooks run shell commands at lifecycle events. Commands - Different! Commands are slash shortcuts for prompts.

Next Steps

• Read full subagents documentation
• Check .claude/agents/ for custom agents
• Practice parallel execution
• Create custom agents for your workflows
• Use /cc-commands for slash commands
• Use /cc-hooks for lifecycle automation