Create Agent Skill
Purpose
Assist in creating new Agent Skills that follow the Agent Skills open standard specification and VS Code conventions.
When to Use
Use this skill when:
- Creating a new Agent Skill for the repository
- Need guidance on Agent Skills structure and format
- Updating or refactoring existing skills
Agent Skills Structure
Directory Layout
.github/skills/{skill-name}/SKILL.md
Requirements:
- Skills must be in
.github/skills/directory - Each skill gets its own subdirectory with a descriptive kebab-case name
- Skill content must be in a file named
SKILL.md(all caps)
File Format
YAML Frontmatter (required):
---
name: Human-Readable Skill Name
description: Brief one-line description of what the skill does
---
Markdown Content:
- Clear heading with skill name
- Purpose/overview section
- When to use this skill
- Step-by-step instructions
- Examples and code blocks
- Success criteria
- Related resources
Content Guidelines
Writing Style
- Clear and actionable: Each instruction should be unambiguous
- Command-focused: Include actual commands that can be run
- Context-aware: Explain when and why to use each step
- Complete: Don't assume prior knowledge
Structure Best Practices
- Purpose: One sentence explaining what the skill does
- When to Use: Specific triggers or scenarios
- Prerequisites: What needs to be in place first (if any)
- Steps: Numbered or sectioned workflow
- Expected Results: What success looks like
- Troubleshooting: Common issues and solutions
- Quick Reference: Copy-paste friendly commands
Code Blocks
- Use proper syntax highlighting (
bash,python, etc.) - Include expected output when helpful
- Provide full commands, not fragments
- Add comments for complex operations
Example Structure
---
name: My Skill Name
description: What this skill accomplishes in one line
---
# My Skill Name
## Purpose
Brief explanation of what this skill does and why it exists.
## When to Use
- Scenario 1
- Scenario 2
- Scenario 3
## Prerequisites
- Requirement 1
- Requirement 2
## Steps
### 1. First Action
\```bash
command --with-flags
\```
**Expected Result**: What should happen
### 2. Second Action
\```bash
another-command
\```
**Expected Result**: What should happen
## Success Criteria
- ✅ Criterion 1
- ✅ Criterion 2
## Troubleshooting
**Issue**: Common problem
**Solution**: How to fix it
## Quick Reference
\```bash
# All commands in sequence
command1 && command2 && command3
\```
Validation Checklist
Before committing a new skill, verify:
- Located in
.github/skills/{skill-name}/SKILL.md - Has YAML frontmatter with
nameanddescription - Name is clear and descriptive
- Description is concise (one line)
- Purpose section explains the "why"
- "When to Use" section lists specific scenarios
- Instructions are step-by-step and actionable
- Code blocks have proper syntax highlighting
- Success criteria are explicit
- No typos or formatting issues
- Skill is focused on one clear task
References
Quick Start Template
# Create new skill directory
mkdir -p .github/skills/{skill-name}
# Create SKILL.md with frontmatter
cat > .github/skills/{skill-name}/SKILL.md << 'EOF'
---
name: Skill Name
description: One-line description
---
# Skill Name
## Purpose
[What this skill does]
## When to Use
- [Scenario 1]
- [Scenario 2]
## Steps
### 1. [First Action]
\```bash
[command]
\```
### 2. [Second Action]
\```bash
[command]
\```
## Success Criteria
- ✅ [Criterion 1]
- ✅ [Criterion 2]
EOF
Post-Creation
After creating a skill:
- Test the instructions manually to verify accuracy
- Review frontmatter formatting
- Commit with message:
docs: add {skill-name} agent skill - Skills are automatically discovered by GitHub Copilot - no registration needed
Converted and distributed by TomeVault — claim your Tome and manage your conversions.