Skip to content
Back to skills

Web Demo

ASecurity

"Routes DragGAN browser-demo tasks such as launching the Gradio UI,

  • 247 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 8, 2026
developmentpythonbackend

Works with

  • cli

Security analysis

A100/100

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

Scanned September 8, 2026

npx -y skills add VectorSpaceLab/AREX-Skill --skill web-demo --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Web Demo?

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

Security grade badge for Web Demo
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/vectorspacelab-web-demo-8f70752d/badge)](https://www.skillsdirectory.com/skills/vectorspacelab-web-demo-8f70752d)

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: web-demo
description: "Routes DragGAN browser-demo tasks such as launching the Gradio UI,
  selecting checkpoints, dragging points, and saving edited outputs."
disable-model-invocation: true
metadata:
  disco-role: operating
license: NO_LICENSE
---

# web-demo

Use this sub-skill when the task is about the browser experience: launch the demo, open the web UI, choose a checkpoint, click handle and target points, save an image or video, or troubleshoot the demo runtime.

## Start here

1. Run the bundled preflight:
   `../../scripts/check_install.py --mode web`
2. Launch the demo with the bundled wrapper:
   `../../scripts/launch_web_demo.py --device cuda --ip 0.0.0.0 --port 7860`
3. Read `references/workflows.md` for the point-drag sequence and UI controls.
4. Read `../../references/checkpoints.md` if you need model names or cache behavior.
5. Read `references/troubleshooting.md` if the UI opens but the interaction fails.

## This sub-skill covers

- Launching the browser demo with the supported launch flags.
- Switching between checkpoints in the model dropdown.
- Creating a seeded image, placing handle and target points, and running `Drag it`.
- Resetting, undoing, and saving edited outputs.
- Understanding the saved `draggan_tmp/` files and checkpoint cache usage.

## This sub-skill does not cover

- Writing a custom Python script around `draggan.draggan` functions.
- Low-level tensor shapes, point encoding, or latent manipulation details.
- Claims that CPU or MPS is a verified drag backend.

## Important cautions

- The verified drag path in this snapshot requires CUDA.
- The `Draw a Mask` tab is present in the UI, but the current optimizer does not enforce the mask.
- Uploaded-image inversion is experimental; treat it as a best-effort path rather than the default workflow.
- If you only need the install check or environment diagnosis, use the root scripts and troubleshooting reference first.

## Helpful references

- `references/workflows.md` for the user-facing browser steps.
- `references/troubleshooting.md` for launch, model-download, and UI-specific failures.
- `../../references/deployment.md` for install and launch modes shared with the root skill.
- `../../references/checkpoints.md` for checkpoint names and the cache root.

Files in this skill

  • SKILL.md2.2 KB
  • references/troubleshooting.md1.5 KB
  • references/workflows.md1.9 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.