skill-token-refactor

To refactor markdown instructions for token optimization, you compress existing Skill documentation into dense, machine-readable packages. This restructuring maximizes token efficiency while preserving behavioral integrity and safety rules.

Compressing agent capability libraries requires refactoring instructions into classified components like scripts and schemas. This structural optimization reduces noise while ensuring safety rules, stop conditions, and failure modes remain intact.

Token-efficient refactoring preserves agent behavior by maintaining compliance with authoring contracts during the compression process. It ensures semantic versioning and rigorous validation of changed constraints so original execution logic is not altered.

Yes, you can automate markdown refactoring for legacy skill packages. The process automatically classifies content into scripts, references, schemas, or evals to transform verbose formats into modern, high-ROI packages.

Refactoring instructions into machine-readable formats requires no external dependencies. It operates directly on existing Markdown-based Skill instructions to validate changed constraints and enforce structured output.

You should not use automated token optimization when your priority is human readability over machine parsing. The refactoring process deliberately packs instructions into dense formats, which increases structural density and reduces verbose explanations.