skill

YAML-based skills are created and validated using a required four-file structure: skill.yaml, sharp-edges.yaml, validations.yaml, and collaboration.yaml. This enforces consistent metadata for authoring, validating readiness, and orchestrating cross-skill workflows within Ralph's ecosystem.

The required file structure for a YAML skill includes four files: skill.yaml, sharp-edges.yaml, validations.yaml, and collaboration.yaml. You can also include optional directories for scripts, references, and assets to support the skill's functionality.

Cross-skill workflow orchestration coordinates multiple YAML-defined skills during runtime. It uses frontmatter-driven metadata for discovery and leverages the collaboration.yaml file to manage interactions and dependencies between separate modular capabilities.

Yes, frontmatter-driven metadata is supported for skill discovery. It allows the runtime to list available skills and identify their capabilities based on the structured metadata defined within the skill.yaml file.

The four-file YAML structure requires strict adherence to skill.yaml, sharp-edges.yaml, validations.yaml, and collaboration.yaml. This constraint ensures structural validation and deployment readiness but limits skill definitions to this specific architectural format.

No dependencies are required to start authoring YAML skills. You can create a new skill by generating the four required YAML files and utilizing optional directories for scripts, references, and assets to modularize your AI workflows.