بنقرة واحدة
create-skill
// Create new Agent Skills following the agentskills.io specification. Use when the user wants to create, scaffold, or design a new skill for AI agents. Handles SKILL.md generation, directory structure setup, and validation.
// Create new Agent Skills following the agentskills.io specification. Use when the user wants to create, scaffold, or design a new skill for AI agents. Handles SKILL.md generation, directory structure setup, and validation.
| name | create-skill |
| description | Create new Agent Skills following the agentskills.io specification. Use when the user wants to create, scaffold, or design a new skill for AI agents. Handles SKILL.md generation, directory structure setup, and validation. |
| license | MIT |
| metadata | {"author":"njzjz-bot","version":"1.0"} |
This skill helps you create new Agent Skills that follow the agentskills.io specification.
When asked to create a new skill:
pdf-processing, data-analysis)SKILL.md with proper frontmatterskill-name/
├── SKILL.md # Required: main skill file
├── scripts/ # Optional: executable code
├── references/ # Optional: additional documentation
└── assets/ # Optional: static resources
---
name: your-skill-name
description: What this skill does and when to use it. Be specific and include keywords that help agents identify relevant tasks. Max 1024 characters.
license: MIT
compatibility: Optional - environment requirements if any
metadata:
author: your-name
version: "1.0"
allowed-tools: Optional - pre-approved tools (experimental)
---
# Skill Title
Brief introduction to the skill.
## Usage
Step-by-step instructions on how to use this skill.
## Examples
Example inputs and outputs.
## Notes
Common edge cases and tips.
---Valid: pdf-processing, data-analysis, code-review-2
Invalid: PDF-Processing, -pdf, pdf--processing
Good: "Extracts text and tables from PDF files. Use when working with PDF documents, extracting content from PDFs, or processing scanned documents." Poor: "Helps with PDFs."
MIT, Apache-2.0, Proprietary. LICENSE.txt has complete termsauthor, version, tagsBash(git:*) Bash(jq:*) ReadDesign for efficient context usage:
Keep SKILL.md under 500 lines. Move detailed content to references/.
SKILL.md focused on core instructionsreferences/REFERENCE.mdassets/scripts/Use relative paths from skill root:
See [the reference guide](references/REFERENCE.md) for details.
Run: scripts/process.py
Keep references one level deep. Avoid deeply nested chains.
After creating a skill, validate it:
# Install skills-ref if needed
npm install -g @agentskills/skills-ref
# Validate the skill
skills-ref validate ./your-skill-name
When asked to create a skill for X:
your-skill-name/SKILL.md with:
scripts/ for helper scriptsreferences/ for detailed docsassets/ for templates/dataskills-ref validateJust a SKILL.md with instructions:
my-skill/
└── SKILL.md
For skills that run code:
my-skill/
├── SKILL.md
└── scripts/
└── helper.py
For detailed documentation:
my-skill/
├── SKILL.md
└── references/
├── REFERENCE.md
└── examples.md
my-skill/
├── SKILL.md
├── scripts/
│ ├── setup.sh
│ └── process.py
├── references/
│ ├── API.md
│ └── FORMATS.md
└── assets/
├── template.json
└── schema.json