authoring-skills
MUST BE USED when creating, improving, or reviewing Skills. Use PROACTIVELY when user mentions "create a skill", "new skill", "write a skill", "improve skill", "skill description", or "progressive disclosure". Guides through specification-compliant skill design with three-level progressive disclosure.
When & Why to Use This Skill
The Authoring Skills tool is a specialized meta-utility designed to standardize and streamline the creation, improvement, and auditing of AI agent skills. By enforcing the Universal Skills Specification and utilizing a three-level progressive disclosure framework, it ensures that skills are robust, secure, and optimized for autonomous agentic performance. This tool is essential for developers looking to build high-quality, maintainable, and specification-compliant AI capabilities.
Use Cases
- Skill Creation: Provides a structured workflow and templates to build new agent skills from scratch, ensuring they follow agentic patterns and security best practices.
- Compliance Auditing: Acts as a review mechanism to evaluate existing skills against a standardized checklist, identifying gaps in documentation, pathing, or instruction clarity.
- Instruction Optimization: Refines skill descriptions and workflows using imperative language and progressive disclosure to improve the reliability and efficiency of AI agents.
| name | authoring-skills |
|---|---|
| description | MUST BE USED when creating, improving, or reviewing Skills. Use PROACTIVELY when user mentions "create a skill", "new skill", "write a skill", "improve skill", "skill description", or "progressive disclosure". Guides through specification-compliant skill design with three-level progressive disclosure. |
| category | meta |
| keywords | skill, create, authoring, develop, progressive-disclosure, compliance |
Skill Development
Guides effective Skill creation following the Universal Skills Specification.
Overview
Skills enable agents to work autonomously without human intervention. This skill provides the methodology for creating specification-compliant skills using progressive disclosure, robust scripting, and agentic patterns.
When to Use
- User mentions "create a skill", "new skill", "write a skill", "improve skill"
- Reviewing or auditing existing skills for compliance
- Designing skill architecture with proper agentic patterns
Workflow
1. **Analyze & Plan**: - Determine Degrees of Freedom (Fragile vs Creative) using `references/agentic-patterns.md`. - Define Security Scope using `references/security-best-practices.md`. - Create `evaluation_plan.md` from `assets/templates/evaluation_plan.md`.Draft Content:
- Use
assets/templates/skill_structure.mdforSKILL.md. - Use
assets/templates/robust_script.pyfor any Python scripts. - Write frontmatter using
references/description-templates.md(CRITICAL).
- Use
Verify & Refine:
- Validate against
references/compliance-checklist.md. - Ensure all paths are relative to project root.
- Verify that instructions use imperative, third-person language (no "you").
- Validate against
Critical Rules
- **Frontmatter**: MUST use directive language (MUST BE USED, Use PROACTIVELY) in `description`. - **Language**: NEVER use second person ("You should"). Use imperative ("Do X"). - **Structure**: Keep `SKILL.md` body under 500 lines. Move details to `references/`. - **Pathing**: ALWAYS use project-root relative paths. - **Verification**: workflows MUST include self-correction loops.Resources
- Core Principles:
references/core-principles.md(Philosophy, Directory Structure, Detailed Workflow) - Templates:
- Description:
references/description-templates.md - Structure:
assets/templates/skill_structure.md - Script:
assets/templates/robust_script.py - Evaluation:
assets/templates/evaluation_plan.md
- Description:
- Guides:
references/compliance-checklist.mdreferences/agentic-patterns.mdreferences/security-best-practices.mdreferences/progressive-disclosure.md