skill-authoring-and-validation

Agent skill authoring requires a SKILL.md file with YAML frontmatter for discovery and a body defining execution rules. Proper structure ensures the skill is reusable, correctly routed, and bounded to avoid overlap across agents.

To prevent over-triggering, routing-focused frontmatter descriptions must clearly state when the skill should trigger and what it should not overlap with, ensuring boundaries are checked against existing repository skills.

Validate trigger routing by running prompt-based trigger checks and boundary checks using realistic scenarios. This ensures the skill activates only under intended conditions before finalizing important or ambiguous workflows.

Overlap occurs when skill scope is overly broad or routing descriptions lack specific trigger boundaries. Perform repository fit and overlap checks against existing skills to avoid duplication and ensure proper workflow standardization.

No, skill authoring does not assume platform-specific runtime contracts. The validation workflow relies on prompt-based trigger checks and realistic scenarios to standardize creation without binding to specific platform runtimes.

Workflow standardization defines required SKILL.md structure and repository-specific authoring rules for OpenCode Agent Skills. It ensures workflows are reusable, properly routed, and bounded to avoid overlap or over-triggering across agents.