Back to Skills

create-skill

verified

Create and manage Claude Code skills following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter (including context fork, allowed-tools, agent), trigger types (keywords, intent patterns, file paths, content patterns), enforcement levels (block, suggest, warn), hook mechanisms (UserPromptSubmit, PreToolUse), session tracking, and the 500-line rule.

View on GitHub

Marketplace

fullstack-dev

runxgalee/fullstack-dev

Plugin

development

development

Repository

runxgalee/fullstack-dev

plugins/development/skills/create-skill/SKILL.md

Last Verified

January 20, 2026

Install Skill

Select agents to install to:

Scope:
npx add-skill https://github.com/runxgalee/fullstack-dev/blob/main/plugins/development/skills/create-skill/SKILL.md -a claude-code --skill create-skill

Installation paths:

Claude
.claude/skills/create-skill/
Powered by add-skill CLI

Instructions

## Purpose

Help developers create and manage Claude Code skills following Anthropic best practices. This skill provides comprehensive guidance for:

- **Skill Structure**: YAML frontmatter configuration, content organization, and file placement
- **Trigger Patterns**: Keywords, intent patterns, file paths, and content patterns for automatic activation
- **Advanced Features**: Context forking, tool restrictions, agent delegation, and session tracking
- **Enforcement Levels**: Block, suggest, and warn behaviors for different use cases

## When to Use

Use this skill when you need to:

- **Create a new skill** from scratch with proper structure and configuration
- **Understand trigger patterns** - how skills are automatically activated based on user input
- **Configure skill behavior** - context fork, allowed-tools, agent specification
- **Debug skill activation** - troubleshoot why a skill isn't triggering as expected
- **Implement guardrails** - create blocking skills that prevent common mistakes
- **Design domain skills** - create advisory skills for specific technical areas
- **Follow the 500-line rule** - structure large skills with progressive disclosure

## Key Information

### Skill File Structure

Skills are stored in `.claude/skills/{skill-name}/SKILL.md` with required YAML frontmatter:

```yaml
---
name: skill-name              # Required: Unique identifier
description: Keywords here    # Required: Max 1024 chars, include trigger terms
context: fork                 # Optional: Run in isolated sub-agent context
agent: specialist-name        # Optional: Delegate to specific agent
allowed-tools:                # Optional: Restrict available tools
  - Read
  - Grep
  - Glob
user-invocable: true          # Optional: Show in slash command menu (default: true)
---
```

### Trigger Pattern Best Practices

The `description` field is critical for skill activation. Include:

1. **Primary keywords**: Direct terms users might use (e.g., "skill", "create skill")
2. **Rel

Validation Details

Front Matter
Required Fields
Valid Name Format
Valid Description
Has Sections
Allowed Tools
Instruction Length:
6570 chars