with one click
agent-name
Technical reference for implementing subagents. Use when asked about frontmatter fields, system prompt structure, or subagent file format. Provides essential specs with references for details.
Menu
Technical reference for implementing subagents. Use when asked about frontmatter fields, system prompt structure, or subagent file format. Provides essential specs with references for details.
| name | Subagent Implementation |
| description | Technical reference for implementing subagents. Use when asked about frontmatter fields, system prompt structure, or subagent file format. Provides essential specs with references for details. |
| version | 0.2.0 |
Subagent files are Markdown with YAML frontmatter:
---
name: agent-name
description: When to use this agent...
model: inherit
color: blue
tools: ["Read", "Grep"]
---
You are [agent role]...
**Responsibilities:**
1. [Task 1]
2. [Task 2]
Location: .claude/agents/ or .claude-plugin/agents/
Naming: kebab-case, 3-50 chars (e.g., code-reviewer)
| Field | Format | Example |
|---|---|---|
name | kebab-case, 3-50 chars | code-reviewer |
description | Text with 2-4 examples | See below |
model | inherit/haiku/sonnet/opus | inherit |
color | blue/cyan/green/yellow/magenta/red | blue |
Include 2-4 <example> blocks:
description: |
Use when reviewing code for quality. Examples:
<example>
Context: User wrote new code
user: "Review this"
assistant: "I'll use the code-reviewer"
<commentary>Code review is core function</commentary>
</example>
| Field | Purpose | Example |
|---|---|---|
tools | Restrict tool access | ["Read", "Grep", "Glob"] |
permissionMode | Permission handling | acceptEdits |
skills | Preload skills | ["api-conventions"] |
Permission modes: default, acceptEdits, dontAsk, bypassPermissions, plan
See references/frontmatter-reference.md for complete field documentation.
You are a [role] specializing in [domain].
**Responsibilities:**
1. [Primary task]
2. [Secondary task]
**Process:**
1. [Step 1]
2. [Step 2]
3. [Verification]
**Output Format:**
[Specify exact output structure]
**Edge Cases:**
- [Case 1]: [How to handle]
DO:
DON'T:
See references/system-prompt-patterns.md for detailed templates.
For complete documentation:
references/frontmatter-reference.md - All fields with examplesreferences/system-prompt-patterns.md - Detailed templatesexamples/complete-analyzer-example.md - Working exampleFor distribution via marketplaces, see the Anthropic docs on Plugin Marketplaces.
This skill should be used when the user asks to "secure Notion MCP", "protect OAuth token", "MCP security best practices", "prevent token exposure", "credential management", or needs guidance on securing Notion MCP integration. Covers credential storage, validation, and security considerations.
This skill should be used when the user asks to "set up Notion MCP", "explain OAuth flow", "understand Notion MCP", "how to authenticate with Notion", or wants to learn about the Notion MCP setup process. Provides understanding of what Notion MCP is and why OAuth-based setup is required.
This skill should be used when the user asks to "what Notion MCP tools are available", "how to use read_page", "Notion MCP tool reference", "when to use query_database", "Notion MCP capabilities", or needs a guide to what operations are possible with Notion MCP. Provides reference to all available tools and when to use each.
Core principles for designing effective subagents. Use when asked to "design a subagent", "structure a subagent", "subagent best practices". Provides essential guidance with references for deep dives.