---

name: skill-name-in-kebab-case description: >- REQUIRED AND MOST IMPORTANT DESCRIPTION IN THIS FILE. It is the only part read by the agent's "router" to decide whether this skill is relevant to the user's request. Write in the format: "Does X, Y, and Z. Use when the user asks for A, mentions B, or needs C." Be specific, use active verbs, and include realistic trigger phrases (what the person would actually type). Avoid vague or generic descriptions.

Optional fields below — remove any you don't need.

license: Apache-2.0

metadata:

author: your-name-or-org

version: "1.0.0"

when_to_use: |

- More detailed list of triggers, if the description isn't enough

- E.g.: user mentions "review PR", "audit security", etc.

- Do NOT use for: (cases that seem similar but aren't)

argument-hint: "Describe what the user should provide (e.g., file path)"

requires:

- name-of-external-cli-or-tool (e.g., kubectl, psql, gh)

---

Readable Skill Name

Overview

One or two sentences explaining the purpose of this skill and the problem it solves.

When to use

• Trigger 1 (e.g.: user asks to "review code")
• Trigger 2
• Do NOT use when: (edge cases that seem similar but don't apply)

Instructions

Objective step-by-step of what the agent should do. Keep it direct and actionable.

1. First step (e.g.: read the relevant file, run command X)
2. Second step
3. Third step — if scripts are needed, point to scripts/ here
4. If extensive context is needed, point to references/ instead of pasting everything here (keep the body of this file under ~500 words)

Output format

How the final response should be structured (e.g.: list, table, commit message, JSON, etc.)

Examples

Input: example of a user request Expected output: example of the result the skill should produce

Notes / Limitations

• Any restrictions, external dependencies, or special care needed
• E.g.: "requires prior authentication via gh auth login"