Skip to content

Latest commit

 

History

History
40 lines (28 loc) · 1.6 KB

File metadata and controls

40 lines (28 loc) · 1.6 KB

Skill Template

Copy the block below into domains/<domain>/skills/<skill-name>/skill.md and replace the placeholders. See Authoring a skill for the canonical layout and the full frontmatter schema.


---
name: <skill-name>
description: <one or two sentences, including when-to-use cues so agents can match this skill; keep it well under the operator limits, around 1,000 characters>
maturity: experimental
---

# <Skill Title>

## When To Use

- <the concrete condition that should trigger this skill>
- <another triggering condition>

## Workflow

1. **<First step>.** State what the agent should do and how to tell it is done.
2. **<Second step>.** Keep steps imperative and verifiable.
3. **<Continue as needed>.** Reference supporting material in `references/` rather than inlining long content.

## Common Pitfalls

| Pitfall | Rule |
|---|---|
| <a tempting wrong move> | <the rule that avoids it> |
| <another common mistake> | <its rule> |

Notes for authors

  • Keep name unprefixed and matching the skill's directory name. The installer adds the mms- prefix to the generated outputs.
  • maturity is optional and defaults to stable. Use experimental for new skills and deprecated on the way out.
  • Put supporting docs in references/, which the skill reads on demand, rather than expanding the description. The description is always loaded into the agent's discovery surface, while the body and references are not.
  • Add a repos/<repo>.md overlay only when a consuming repo needs repo-specific guidance. It is merged into the skill body at install time.