skill-creator
Generates new agent skills following the official Agent Skills specification. Use when you need to teach the AI a new capability, create reusable workflows, or document best practices in a structured format.
When & Why to Use This Skill
The Skill Creator is a meta-tool designed to automate the generation, validation, and structuring of AI agent capabilities. It ensures that new workflows and instructions strictly adhere to the official Agent Skills specification, facilitating the creation of reusable, standardized, and high-quality skills for agentic ecosystems.
Use Cases
- Standardizing custom prompts: Transform loose instructions or complex prompts into formal, structured skill files that are easy for AI agents to parse and execute.
- Automated validation: Verify that new skills meet strict naming conventions, YAML frontmatter requirements, and directory structures before deployment.
- Workflow documentation: Convert manual SOPs and best practices into a machine-readable format to teach AI agents new, repeatable capabilities.
- Scaling agent ecosystems: Rapidly generate the necessary file architecture (.github/skills/) for large-scale agentic projects while maintaining consistency.
| name | skill-creator |
|---|---|
| description | Generates new agent skills following the official Agent Skills specification. Use when you need to teach the AI a new capability, create reusable workflows, or document best practices in a structured format. |
Skill Creator Skill
This skill helps you generate, validate, and structure new capabilities (Skills) for the agent, ensuring strict adherence to the Agent Skills Specification.
When to Use This Skill
Use this skill when you want to:
- Create a new skill directory and
SKILL.mdfile - Convert a set of instructions or a prompt into a reusable skill
- Standardize existing documentation into the Skill format
- Validate that a skill follows naming and formatting rules
Skill Generation Workflow
1. Planning
Before generating a skill, define:
- Name: 1-64 chars, lowercase
a-z, numbers, hyphens. No start/end hyphen. Match directory name. - Purpose: What specific task does it solve?
- Trigger: When should the agent use it?
- Context: What dependencies or tools does it need?
2. File Structure
Always create this structure:
.github/skills/
└── <skill-name>/
└── SKILL.md # Required: Frontmatter + Instructions
Optional subdirectories (only if needed):
scripts/(executable code)references/(lookup tables, extended docs)assets/(templates, images)
3. Frontmatter Template
Every SKILL.md MUST start with this YAML block:
---
name: skill-name-here # REQUIRED: Matches directory, valid formatting
description: Description. # REQUIRED: Max 1024 chars. What + When to use.
license: Apache-2.0 # OPTIONAL
compatibility: ... # OPTIONAL: Environment requirements
---
4. Body Content Template
Use this structure for the Markdown body:
# [Skill Name]
[Brief introduction]
## When to Use This Skill
- [Condition 1]
- [Condition 2]
## Step-by-Step Instructions
1. [Step 1]
2. [Step 2]
## Examples
#### Example 1: [Scenario]
[Description or input/output]
## Best Practices / Notes
- [Tip 1]
- [Tip 2]
Naming Rules (Strict)
- ✅
pdf-processing,data-analysis,code-review-v2 - ❌
PDF-Processing(no uppercase) - ❌
-pdf(no starting hyphen) - ❌
pdf--processing(no consecutive hyphens) - ❌
pdf processing(no spaces)
Generation Prompts
When asking the agent to create a skill, provide:
- Name: "Create a skill named
git-bisect-helper" - Goal: "It should guide the helping user find bugs using git bisect"
- Details: "Include steps for starting, marking good/bad, and resetting"
Validation Checklist
- Directory name matches exactly
namein frontmatter - Frontmatter is valid YAML
- Description explains "What" and "When"
- References to files use relative paths (e.g.,
scripts/run.py) - Main
SKILL.mdis concise (progressive disclosure)