Skip to content
Back to skills

Skill Creator

ASecurity

Use when creating, adapting, evaluating, or improving an Agent Skill and its trigger description.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 13, 2026
ai-agentsgo

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned October 13, 2026

npx -y skills add z-Zihan/Evir --skill skill-creator --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Skill Creator?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Skill Creator
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/z-zihan-skill-creator/badge)](https://www.skillsdirectory.com/skills/z-zihan-skill-creator)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: skill-creator
description: Use when creating, adapting, evaluating, or improving an Agent Skill and its trigger description.
---

# Skill Creator

Adapted for Evir from Superpowers `writing-skills`.

## Define the behavior first

Capture the user task, triggering phrases and contexts, expected output, non-goals, tools,
platform restrictions, risk, and objective success evidence. A Skill teaches judgment or a
reusable method; deterministic constraints belong in code, schemas, or tests.

## Authoring rules

- Use a lowercase hyphenated identifier and a concrete name.
- Make the description answer only "when should this load?" with real symptoms and synonyms.
- Keep the main file focused; every referenced resource must be bundled and loadable by the
  target runtime.
- Use imperative steps, decision points, recovery paths, and a verifiable output contract.
- State required and optional capabilities accurately. A Skill never grants tools or bypasses
  mode, workspace, network, or approval policy.
- Remove platform-specific paths, commands, sub-agent assumptions, and dependencies that the
  target runtime does not provide.
- Record upstream author, repository, license, path, immutable revision, and whether content was
  modified.

## Evaluate

Create representative positive, negative, and ambiguous prompts. Verify that the Skill triggers
when useful, stays inactive when irrelevant, improves the result over baseline, respects safety
boundaries, and does not require missing files. Review token cost and trim repeated instructions.

Before bundling third-party content, verify its license and retain required notices. Default new
Skills to disabled so users opt into context and behavior changes.

Files in this skill

  • SKILL.md1.7 KB
  • manifest.json1.1 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.