# Skill Creator

> Create, improve, or test an Agent Skill (a SKILL.md playbook). Use when the user wants to capture a repeatable workflow as a skill, says "turn this into a skill", or asks how skills are written.

- Skill: `standardharness/skill-creator` (Agent Skill)
- Install (CLI): `npx skillmds@latest add standardharness/skill-creator`
- Raw SKILL.md: https://api.skillmd.com/api/skills/standardharness/skill-creator/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: standardharness (https://skillmd.com/u/standardharness)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/standardharness/skill-creator

---


# Creating a skill

A skill is a folder with a `SKILL.md`. Only `name` and `description` sit in the
catalog; the body loads on demand. Write the description as the trigger.

## Capture first

Pull the workflow from the conversation before asking. Corrections the user
made along the way are the most valuable lines — they are what a fresh session
would get wrong.

Confirm in one or two sentences, then write.

## Where it lives

| Location | Scope |
|---|---|
| `.harness/skills/<name>/SKILL.md` | this project |
| `~/.harness/skills/<name>/SKILL.md` | this user, every project |

A single `<name>.md` file is also valid. Project shadows personal.

## File

```markdown
---
name: run-migrations
description: Apply or roll back database migrations in this repo, and when to reach for it.
---

# Run migrations
1. ...
```

`description` must say when to use it and when not to. Unknown frontmatter is
ignored.

## Test

Ask the user two prompts: one that should load the skill, one that should not.
If the catalog line would fire on the second, tighten the description.