Agent skill

examples-git-fg-meta-plugin-manager

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/examples-git-fg-meta-plugin-manager

SKILL.md

Example: Basic Skill

This example shows a minimal, well-structured skill following all thecattoolkit_v3 standards.


File Structure

skills/hello-world/
└── SKILL.md

Complete SKILL.md

markdown
---
name: hello-world
description: "Greets users with a friendly message. Use when you want to say hello or demonstrate basic skill structure. Not for complex greetings or personalized messages."
---

# Hello World Skill

This skill demonstrates the basic structure of a well-formed skill.

## What It Does

Greets users with a simple, friendly message.

## Usage

Say "Hello, World!" to the user.

## Message

Hello, World! Welcome to the thecattoolkit_v3.

Breakdown and Explanation

Frontmatter

yaml
---
name: hello-world              # Max 64 chars, lowercase, hyphens only
description: "Greets users... Use when... Not for..."
---

Key Points:

  • name: Uses kebab-case (lowercase with hyphens)
  • description: Non-spoiling, includes "Use when" and "Not for"
  • No context: fork (uses default shared context)
  • No agent (not needed without context: fork)

Frontmatter Fields Explained

Field Value Why
name hello-world Lowercase, hyphens, descriptive
description "Greets users... Use when... Not for..." Non-spoiling, clear triggers and exclusions

Body Structure

markdown
# Hello World Skill      <- Clear title

This skill demonstrates...  <- Brief purpose statement

## What It Does          <- What section
Greets users...

## Usage                 <- How section
Say "Hello..."

## Message               <- Content
Hello, World!

Key Points:

  • Clear headings for organization
  • Brief purpose statement
  • Separation of "what" and "how"
  • Actual content/message at the end

Quality Checklist Applied

Frontmatter

  • name present: hello-world
  • description present: Non-spoiling
  • description has "Use when": "Use when you want to say hello..."
  • description has "Not for": "Not for complex greetings..."

Naming

  • Gerund form: hello-world (noun phrase acceptable for simple names)
  • Lowercase only: Yes
  • Hyphens as separators: Yes
  • Not reserved words: Yes
  • Not vague: hello-world is specific

Description Quality

  • Third person: "Greets users" not "I will greet"
  • Non-spoiling: Doesn't reveal the exact message
  • Trigger conditions: "Use when you want to say hello..."
  • Exclusions clear: "Not for complex greetings..."

Progressive Disclosure

  • Core in SKILL.md: All content in main file
  • No references/ folder: Not needed for simple skill
  • No scripts/: Not needed

How to Invoke

User Invocation

bash
# From command palette
/hello-world

# Or from chat
Use the hello-world skill

Agent Invocation

markdown
# In another skill or command
Skill(hello-world)

What Happens When Invoked

  1. Metadata loaded: name and description already in system prompt
  2. SKILL.md read: Body content loaded into context
  3. Skill executes: Agent reads and follows instructions
  4. Result: "Hello, World!" message displayed

Common Variations

With argument-hint

yaml
---
name: greet
description: "Greets users with a custom message. Use when you want to say hello with a name. Not for complex greetings."
argument-hint: "<name>"
---

With user-invocable: false

yaml
---
name: internal-helper
description: "Internal helper skill. Use when performing internal operations. Not for user-facing tasks."
user-invocable: false
---

Result: Skill is hidden from / menu, but Claude can still invoke via Skill(internal-helper).

With disable-model-invocation

yaml
---
name: dangerous-operation
description: "Performs a dangerous operation. Use when user explicitly requests this operation. Not for automated tasks."
disable-model-invocation: true
---

Result: Skill won't auto-invoke; requires explicit user intent.


Next Steps

After understanding this basic skill:

  1. Study worker-skill.md for isolated context
  2. Review loadout-agent.md for multi-skill bundling
  3. See quality-gate.md for validation patterns

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results