"Operate LaVague Gradio demos, Chrome extension server workflows,
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill server-extension-gradio --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Server Extension Gradio?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-server-extension-gradio)More formats (shields.io, HTML) on the badges page.
---
name: server-extension-gradio
description: "Operate LaVague Gradio demos, Chrome extension server workflows,
lavague-serve, AgentServer, and DriverServer safely."
disable-model-invocation: true
metadata:
disco-role: operating
root-skill: la-vague
sub-skill: server-extension-gradio
license: Apache 2.0
---
# LaVague server, extension, and Gradio workflows
Use this sub-skill when the user wants to run or debug LaVague's interactive surfaces:
- `WebAgent.demo()` and the optional `lavague-gradio` package.
- The Chrome extension server started by `lavague-serve`.
- Custom extension agents using `AgentServer`, `AgentSession`, and `DriverServer`.
- Extension connection, port, and logs troubleshooting.
Do **not** start a persistent UI, websocket server, browser, provider API call, or public Gradio share link unless the user explicitly asks for a live run. Start with the safe probe:
```bash
python sub-skills/server-extension-gradio/scripts/lavague_ui_server_probe.py --check both --port 8000
```
## Route by user intent
| User intent | Use this path |
| --- | --- |
| "Check whether Gradio/server pieces are installed" | Run the safe probe script with `--check gradio`, `--check server`, or `--check both`. |
| "Show me how to launch the Gradio demo" | Use [references/ui-and-server-workflows.md](references/ui-and-server-workflows.md#gradio-demo-workflow) and optionally print the template with `--print-gradio-template`. |
| "Run the Chrome extension backend" | Confirm the target port, model credentials, and that the browser extension will connect; then use [references/cli-reference.md](references/cli-reference.md). |
| "Customize the extension agent" | Use the `AgentServer(create_agent)` factory recipe in [references/ui-and-server-workflows.md](references/ui-and-server-workflows.md#custom-agentserver-factory). Route provider selection to `../contexts-and-retrievers/`. |
| "The extension cannot connect" | Follow [references/troubleshooting.md](references/troubleshooting.md#extension-cannot-connect). |
| "The browser driver itself fails" | Route to `../browser-drivers/`. |
| "The WebAgent construction or run loop fails" | Route to `../core-web-agent/`. |
## Safety defaults
1. Treat Gradio and Chrome extension flows as **interactive/live** workflows.
2. Probe imports, signatures, and port availability before live launch.
3. Prefer `LAVAGUE_TELEMETRY=NONE` for checks and reproducible debugging.
4. Use `lavague-serve`, not the stale `lavague-server` command spelling.
5. Use `from lavague.server.driver import DriverServer`; do not use stale `lavague.drivers.driverserver` snippets.
6. Keep extension port values synchronized between the server and the extension Connection tab.
7. For custom model contexts, construct the context through `../contexts-and-retrievers/` and pass it into `WorldModel.from_context(context)` and `ActionEngine.from_context(context, driver)`.
## Deep references
- [UI and server workflows](references/ui-and-server-workflows.md)
- [CLI reference](references/cli-reference.md)
- [Troubleshooting](references/troubleshooting.md)
- [Safe probe script](scripts/lavague_ui_server_probe.py)
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!