Skill Authoring Workflow

What it does

This skill, DeanPeters' Skill Authoring Workflow, streamlines the creation and updating of AI skills. It transforms initial ideas, workshop content, or draft prompts into fully compliant SKILL.md assets ready for inclusion in a repository. The workflow ensures consistency and quality by enforcing standards and automated testing throughout the process.

When to use it

• When you have an idea for a new skill but need guidance on structuring it.
• When you already have existing content (like research or workshop materials) that needs to be formatted into a skill.
• When refining an existing skill and ensuring it meets all requirements.
• To avoid inconsistent quality ("looks good to me" roulette) when creating new skills.

Key capabilities

• Guided Wizard: Helps create a skill from initial ideas using interactive prompts (build-a-skill.sh).
• Content-first Generator: Converts existing source content into a skill format (add-a-skill.sh).
• Automated Validation: Includes scripts to check frontmatter validity, section order compliance, metadata limits, and cross-reference resolution (test-a-skill.sh, check-skill-metadata.py, check-skill-triggers.py).
• Workshop Facilitation Protocol: Defines a structured conversation flow for guided skill creation using numbered prompts and options.

Example prompts

• "Run the guided wizard to help me create a new skill." (build-a-skill.sh)
• "Convert my research document into a skill." (add-a-skill.sh research/your-framework.md)
• "Test the skill 'my-new-skill' for compliance." (test-a-skill.sh --skill my-new-skill --smoke)

Tips & gotchas

• Preflight is crucial: Use ./scripts/find-a-skill.sh to avoid creating duplicate skills.
• Follow the workflow sequence: This document defines the logic and outputs, taking precedence over other instructions.
• Strict validation: Always run the automated tests (test-a-skill.sh, check-skill-metadata.py) before committing changes.