skill-creator_skill

Overview

This skill helps create new AI agent skills following a clear project spec and naming conventions. It guides when to create a skill, how to structure files and assets, and what mandatory metadata to include. The outcome is consistent, discoverable skills that encode reusable patterns and project-specific rules.

How this skill works

It inspects recurring patterns, project conventions, and complex workflows to generate a new skill scaffold with required frontmatter and minimal examples. It advises whether to include code templates or local documentation links and provides a checklist to validate the new skill before registering it. The result is a concise, rule-driven skill file plus optional assets and local references.

When to use it

• A pattern repeats and the AI needs explicit guidance or constraints
• Project conventions differ from generic best practices
• Workflows require step-by-step instructions or decision trees
• You need to provide reusable code templates, schemas, or configs
• You want to encode decision logic for choosing approaches

Best practices

• Start with the most critical patterns the AI must follow
• Keep examples minimal, focused, and copy-paste ready
• Place code templates and JSON schemas in assets/ when needed
• Point to local documentation in references/ rather than external URLs
• Ensure frontmatter includes name, description with trigger, author, license, and version

Example use cases

• Create a skill that standardizes test runner setup and common commands
• Add a project-specific formatting and linting skill with config templates
• Document a multi-step deployment workflow with decision tables
• Provide JSON schemas and example configs for configuration-heavy components
• Encode naming conventions and registration checklist for new skills