skills-best-practices

Best practices for SKILL.md descriptions involve validating and tuning frontmatter constraints to ensure consistent triggers. You should apply structured discovery, classify the skill type, and iterate with user feedback until the description activates reliably across various prompts.

To structure an Agent Skill workflow, start by interviewing the user to classify the skill type. Then draft the SKILL.md using official templates, apply progressive disclosure to bundle references on demand, and iterate with feedback to refine the workflow.

Progressive disclosure in skill building is the practice of bundling references, templates, and assets to load on demand rather than embedding everything in the main file. Use it to keep SKILL.md concise while retaining robust workflow support.

Yes, the structured discovery process supports creating, reviewing, and refining existing Agent Skills. It enforces frontmatter constraints and references official templates to validate and tune your SKILL.md files across different domains and tools.

The best way to validate frontmatter constraints is through a repeatable review process that checks triggers and references official templates. Description optimization ensures the frontmatter aligns with best practices and triggers reliably across prompts.

Agent Skills fail to trigger consistently when frontmatter is not properly validated or tuned. You need to apply description optimization best practices, iterate with user feedback, and enforce frontmatter constraints to ensure reliable activation across various prompts.