mcpbeat

Sub Agents

parcadei/sub-agents

Create and configure Claude Code sub-agents with custom prompts, tools, and models

928 tokens
context cost
the whole folder, loaded on every use
1
files
instructions only
0
copies elsewhere
how many repositories repackaged it
3880
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/parcadei/Continuous-Claude-v3 --skill sub-agents

The instruction itself

19 sections, as written by the author

Sub-Agents Reference

Create specialized AI agents with isolated contexts for specific tasks.

When to Use

  • "How do I create a sub-agent?"
  • "Configure agent tools"
  • "What built-in agents exist?"
  • "Agent model selection"
  • "Agent chaining patterns"

Quick Start

/agents

Opens menu to create, edit, and manage agents.

Manual Creation

mkdir -p .claude/agents
cat > .claude/agents/reviewer.md << 'EOF'
---
name: reviewer
description: Code review specialist. Use proactively after code changes.
tools: Read, Grep, Glob, Bash
model: sonnet
---

You are a senior code reviewer focusing on quality and security.

## Review Checklist
- Code clarity and naming
- Error handling
- Security vulnerabilities
- Test coverage
EOF

CLI-Based

claude --agents '{
  "reviewer": {
    "description": "Code reviewer",
    "prompt": "Review for quality and security",
    "tools": ["Read", "Bash"],
    "model": "sonnet"
  }
}'

Agent File Format

---
name: agent-name
description: When/why to use this agent
tools: Read, Edit, Bash      # Optional, inherits all if omitted
model: sonnet                 # sonnet, haiku, claude-opus-4-5-20251101, inherit
---

System prompt content here...

Configuration Fields

| Field | Required | Options |

|-------|----------|---------|

| name | Yes | lowercase, hyphens |

| description | Yes | When to use |

| tools | No | Tool list (inherits all if omitted) |

| model | No | sonnet, haiku, claude-opus-4-5-20251101, inherit |

Built-In Agents

| Agent | Model | Tools | Purpose |

|-------|-------|-------|---------|

| General-purpose | Sonnet | All | Complex multi-step tasks |

| Plan | Sonnet | Read-only | Plan mode research |

| Explore | Haiku | Read-only | Fast codebase search |

Model Selection

| Model | Speed | Best For |

|-------|-------|----------|

| Haiku | Fastest | Search, quick lookups |

| Sonnet | Fast | Most tasks (default) |

| Opus | Slower | Complex reasoning |

Tool Combinations

# Code Reviewer (read-only)
tools: Read, Grep, Glob, Bash

# Debugger
tools: Read, Edit, Bash, Grep, Glob

# Implementer
tools: Read, Write, Edit, Bash, Glob

Example Agents

Code Reviewer

---
name: code-reviewer
description: Reviews code for quality and security. Use after code changes.
tools: Read, Grep, Glob, Bash
model: sonnet
---

Review code for:
- Security vulnerabilities
- Code quality issues
- Missing error handling
- Test coverage gaps

Output findings by priority: Critical > Warning > Suggestion

Debugger

---
name: debugger
description: Debug errors and test failures.
tools: Read, Edit, Bash, Grep, Glob
model: inherit
---

Debugging process:
1. Capture error details
2. Identify failure location
3. Form hypotheses
4. Test and verify
5. Implement fix

File Locations

| Type | Location | Priority |

|------|----------|----------|

| Project | .claude/agents/ | Highest |

| User | ~/.claude/agents/ | Lower |

Advanced Patterns

Resumable Agents

[Agent returns agentId: "abc123"]

# Later: resume with context
claude -r "abc123" "Continue analysis"

Agent Chaining

Use code-analyzer to find issues,
then use optimizer to fix them

Best Practices

  • Single responsibility - One clear purpose per agent
  • Restrict tools - Only grant what's needed
  • Clear descriptions - Action-oriented, include "proactively"
  • Version control - Check .claude/agents/ into git

How to use it

Copy the folder

Take parcadei/sub-agents from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.