skill-creator

You create modular skills by running scripts/init_skill.py to scaffold a new skill, then customize the mandatory SKILL.md YAML frontmatter and optional resource folders before packaging. Running scripts/package_skill.py validates the structure and produces a distributable .skill artifact.

A modular skill package requires a SKILL.md file with YAML frontmatter containing name and description fields. It optionally supports scripts/, references/, and assets/ directories to bundle domain-specific workflows, tooling, and rich context resources.

Yes, pyyaml is a required dependency for packaging modular skills. The packaging and validation scripts rely on pyyaml to parse the mandatory YAML frontmatter in SKILL.md and ensure the structural integrity of the .skill artifact.

Yes, you can include scripts and references in a self-contained skill pack. The skill structure supports optional resource directories for scripts, references, and assets, enabling deterministic task execution and rich context loading on demand.

The best way to standardize agent capability packaging is enforcing a mandatory SKILL.md file with YAML frontmatter for name and description. This approach standardizes creation, enables predictable loading, and supports progressive disclosure of bundled resources.

You should avoid using optional resource directories when a skill requires only basic metadata without bundled scripts, references, or assets. Keeping the package minimal without these directories reduces complexity if deterministic tasks and rich context are unnecessary.