Skip to content
Back to skills

Create Agent

ASecurity

Create a new Claude Code agent with proper YAML frontmatter structure. Use when the user wants to add a specialized agent to a plugin. Handles agent file creation with name, description, tools, model selection, and color configuration.

  • 15 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added May 27, 2026
data-aigobashcode-reviewsecurityperformancedocumentation

Works with

  • claude code

Security analysis

A100/100

Scanned May 27, 2026

npx -y skills add jpoutrin/product-forge --skill create-agent --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Create Agent?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Create Agent
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/jpoutrin-create-agent/badge)](https://www.skillsdirectory.com/skills/jpoutrin-create-agent)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: create-agent
description: Create a new Claude Code agent with proper YAML frontmatter structure. Use when the user wants to add a specialized agent to a plugin. Handles agent file creation with name, description, tools, model selection, and color configuration.
---

# Create Agent Skill

Create new Claude Code agents with proper configuration and structure.

## Agent File Format

Agents are markdown files in the `agents/` directory with YAML frontmatter:

```markdown
---
name: agent-name
description: Full description of what the agent does and when to use it
tools: Glob, Grep, Read, Write, Edit, Bash, WebFetch, WebSearch, TodoWrite
model: sonnet
color: green
---

# Agent Title

Agent prompt content goes here...
```

## Required Frontmatter Fields

| Field | Description | Example |
|-------|-------------|---------|
| `name` | Unique identifier (kebab-case) | `code-reviewer` |
| `description` | One-line summary of agent's purpose | `Reviews code for quality and best practices` |
| `tools` | Comma-separated list of available tools | `Glob, Grep, Read, Write, Edit, Bash` |
| `model` | Model to use: `opus`, `sonnet`, or `haiku` | `sonnet` |
| `color` | Status line color | `green`, `blue`, `purple`, `red`, `orange`, `cyan`, `magenta`, `yellow`, `pink`, `teal`, `violet` |

## Model Selection Guidelines

- **opus**: Strategic decisions, complex architecture, security-critical, compliance, executive-level analysis
- **sonnet**: Technical implementation, code writing, general development tasks (most common)
- **haiku**: Quick lookups, simple transformations, fast responses

## Available Tools

Common tool combinations by agent type:

### Code Development Agents
```
Glob, Grep, Read, Write, Edit, Bash, TodoWrite
```

### Research/Analysis Agents
```
Glob, Grep, Read, WebFetch, WebSearch, TodoWrite
```

### Full-Featured Agents
```
Glob, Grep, Read, Write, Edit, Bash, WebFetch, WebSearch, TodoWrite
```

## Creation Process

1. Determine the agent's purpose and expertise area
2. Choose appropriate model based on complexity
3. Select tools needed for the agent's tasks
4. Write clear, actionable agent prompt
5. Save to `plugins/<plugin-name>/agents/<agent-name>.md`

## Example Agent

```markdown
---
name: code-reviewer
description: Reviews code changes for quality, security, and best practices. Use when user wants code reviewed, needs security audit, or asks about code quality.
tools: Glob, Grep, Read, TodoWrite
model: sonnet
color: yellow
---

# Code Reviewer Agent

You are an expert code reviewer focused on quality, security, and maintainability.

## Review Checklist

- Code correctness and logic
- Security vulnerabilities
- Performance considerations
- Code style and consistency
- Test coverage
- Documentation

## Output Format

Provide structured feedback with:
1. Summary of changes
2. Issues found (categorized by severity)
3. Recommendations
4. Approval status
```

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.