Your Skill Name¶

One-line summary of what this skill does.

Keep this file under ~500 lines. Focus on patching hallucinations and filling knowledge gaps: API gotchas, deprecated patterns, and non-obvious conventions. It should be opinionated. Don't restate general knowledge or upstream docs.

Structure for progressive disclosure: put essentials here in SKILL.md, split detailed references or advanced workflows into separate files that the agent loads only when needed.

Instructions¶

Core procedure for the skill. Only include what the model gets wrong or can't know from its training data.

Use prose for flexible tasks; numbered steps for fragile sequences where order matters. Explain why a rule exists; LLMs follow reasoning better than bare directives.

Add an example to illustrate the goal.

Resources¶

• Primary docs

References¶

This template references Anthropic's official Skills docs:

• Agent Skills overview
• Quickstart
• Skill authoring best practices
• Skills in the API
• How to create custom Skills (Help Center)