accelint-skill-manager

Design agent skills with progressive disclosure by enforcing frontmatter metadata, token-efficient references, and a clear file structure. This approach allows AI agents to load guidance incrementally rather than consuming the entire context at once.

The workflow for creating and auditing agent skills follows a 4-step process using standardized templates and reference patterns. It enforces frontmatter metadata and best practices within a repository of reusable patterns.

Enforce frontmatter metadata and best practices by using standardized templates and centralized guidance during skill creation and refactoring. This ensures compliance with required file structures and token-efficient reference patterns.

Yes, standardized templates can be used to refactor existing agent skills. The progressive disclosure workflow explicitly supports auditing and refining skills to ensure they meet metadata and file structure best practices.

Organize SKILL.md and reference files by following centralized guidance that separates the main workflow, implementation rules, and token-efficient reference resources into a clear, progressive file structure.