Skip to content
Back to skills

References

ASecurity

A plugin teaches agents how to use its engine with a skill: a Markdown file in the plugin's folder, declared in `skills` with a one-line `summary`. Agents see the summary in their brief and read the file on demand with `<id>__skill` (for Sketch, `sketch__skill {"name": "sketch"}`), so the summary says when to read it and the file says what to do. - Scope it: `facts` names the kinds of project it is for (none: every project), and `tools` the plugin tools it explains, so it reaches a session on...

  • 798 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 5, 2026
ai-agentsgodocumentation

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned November 5, 2026

npx -y skills add genex-games/genex-desktop --skill references --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of References?

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

Security grade badge for References
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/genex-games-references/badge)](https://www.skillsdirectory.com/skills/genex-games-references)

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
# The plugin's own skill

A plugin teaches agents how to use its engine with a skill: a Markdown file in the plugin's folder,
declared in `skills` with a one-line `summary`. Agents see the summary in their brief and read the
file on demand with `<id>__skill` (for Sketch, `sketch__skill {"name": "sketch"}`), so the summary
says when to read it and the file says what to do.

- Scope it: `facts` names the kinds of project it is for (none: every project), and `tools` the
  plugin tools it explains, so it reaches a session only while one of them does.
- Write what an agent can't find in the engine's own documentation: the project's layout, which
  tool to call for what, what each tool answers, what a failure means and how to fix it, and what
  never to do (files the engine writes itself, folders it locks while running).
- Keep it short and concrete. A long topic goes in a reference file the skill lists in `references`
  (read with `file`), each at most 128 KiB.
- A skill holds no secrets and gives no instructions to the person; it is read by agents.

Write `genex-plugins/sketch/skills/sketch.md`:

```markdown
# Sketch projects

A Sketch project is a folder with one `*.sketchproj` file at its top, its scenes in `scenes/`
(`*.skscene`) and Sketch's scratch in `Cache/`, which Genex keeps out of history.

- Build with `sketch__build`, optionally naming the `scene` to build first. It answers
  `{ok: true, output}` or `{ok: false, problem}`; a problem quotes the editor's own error, with the
  scene and line it names: fix that file and build again.
- `sketch__build` waits while the person works in the Sketch editor; never ask them to close it.
- A problem saying the Sketch editor isn't answering means it isn't open on this project: ask the
  person to open it, or build without it by running `sketchc build` as a job (it asks as the chat's
  permission mode says). Never try to install Sketch.
- Never edit files in `Cache/` or `Exports/`: sketchc writes them.
- Each scene is one file: a worker of the `scene` type edits one scene at a time, in its own copy.
```

Files in this skill

  • backend.md6.9 KB
  • manifest.md9.2 KB
  • skill.md2 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.

View all in ai-agents →

Comments

Loading comments…

Newsletter

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