| name | your-skill-name |
| author | Your Name |
| license | MIT |
| description | Third-person description of what this skill does, when it should activate, and how to invoke it. Include concrete trigger phrases the agent can match against. Max 1024 characters. Example: "Reviews SQL migrations for destructive operations and suggests safer alternatives. Activates when the user asks to review a migration, or runs `/crystools-skills:your-skill-name`."
|
| metadata | {"version":"0.1.0","tags":"tag-one, tag-two, tag-three","github":"https://github.com/crystian/skills","linkedin":"https://www.linkedin.com/in/your-handle"} |
Your Skill Name
One-line elevator pitch. What problem does this skill solve?
A short paragraph of context: who the skill is for, when it fires, and what the user should expect.
When to use this skill
- Trigger phrase 1.
- Trigger phrase 2.
- Explicit invocation:
/crystools-skills:your-skill-name.
When NOT to use it
- Clearly state what's out of scope. This prevents overlap with other skills.
Execution
1. Resolve inputs
Explain how the skill picks up its inputs (conversation, arguments, files).
2. Do the work
Step-by-step instructions. Keep them imperative and concrete.
3. Present results
How the skill should present output to the user. Format, verbosity, confirmation prompts.
Guardrails
- Never edit files without user confirmation.
- Never run destructive operations.
- Add any skill-specific guardrails here.
References
Heavier material lives in references/ and is loaded on demand:
references/examples.md — read when the user asks for examples.
references/rubric.md — read when running the quality review.