computer-use-browser-desktop

computer-use-browser-desktop is a skill for Claude Code, Codex from ariana-dot-dev/cloud-sandbox-vm-skills. It costs 80 tokens per session (438 once invoked), scanned C, original, MIT.

A cloud Linux desktop with Chrome that an agent can control by clicking, typing, and logging in. It is for tasks that need a visible graphical interface instead of only a command line.

In plain words
What is it for?
Visual UI checks, logging into web apps, testing Electron apps, controlling Chrome, and recording GUI demonstrations.
Why use it?
It lets you check and operate real web or desktop interfaces when shell commands or headless browser tools are not enough. Sessions are recorded as demos.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Not installable: its command points at a path on the author’s own machine, so it runs nowhere else. The line is /home/user/lux-demos/).

Good fit Visual UI checks, logging into web apps, testing Electron apps, controlling Chrome, and recording GUI demonstrations.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

There is no command for this one: it runs only inside a plugin, and the catalogue could not identify which plugin ships it. The source is linked below.

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for computer-use-browser-desktop

README.md
[![agentmods](https://agentmods.dev/badge/skills/ariana-dot-dev/cloud-sandbox-vm-skills/computer-use-browser-desktop/github.svg)](https://agentmods.dev/skills/ariana-dot-dev/cloud-sandbox-vm-skills/computer-use-browser-desktop)
Your own site
<a href="https://agentmods.dev/skills/ariana-dot-dev/cloud-sandbox-vm-skills/computer-use-browser-desktop"><img src="https://agentmods.dev/badge/skills/ariana-dot-dev/cloud-sandbox-vm-skills/computer-use-browser-desktop/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for computer-use-browser-desktop

Your own site · 80×15
<a href="https://agentmods.dev/skills/ariana-dot-dev/cloud-sandbox-vm-skills/computer-use-browser-desktop"><img src="https://agentmods.dev/badge/skills/ariana-dot-dev/cloud-sandbox-vm-skills/computer-use-browser-desktop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 80 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 438 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 2 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00080 $0.00438
Opus 5 $0.00040 $0.00219
Sonnet 5 $0.00016 $0.00088
Haiku 4.5 $0.00008 $0.00044

Measured 9d ago against content hash 8a4d5743e390, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade C, and why

computer-use-browser-desktop scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Downloads and executes remote codehighSupply chain

curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.

`curl -fsSL https://box.ascii.dev/install | sh`, then `box login "$BOX_API_KEY" --json`.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

`curl -fsSL https://box.ascii.dev/install | sh`, then `box login "$BOX_API_KEY" --json`.
skills/computer-use-browser-desktop/SKILL.md · 35 lines

What it actually says

Computer use: browser + desktop (Box)

Use this when a task needs a real GUI, not just a shell: logging into a web app, validating a rendered UI, controlling Chrome, testing an Electron app, or recording a demo. Box runs a full Linux desktop with Chrome; the Lux agent automates desktop/browser interactions and auto-records the session.

Recipe

box_id="$(box new --json | jq -r 'select(.event == "ready") | .id')"
# Get a live desktop stream URL (60fps WebRTC, or VNC over HTTPS on bad networks)
box desktop "$box_id" --json
# Automate a GUI goal with Lux (records to /home/user/lux-demos/)
box ssh "$box_id" -- "lux start 'log in to example.com and screenshot the dashboard'"
box ssh "$box_id" -- "lux run --max-steps 30"
box ssh "$box_id" -- "lux status"

Why Box

  • Real desktop + Chrome — does GUI work headless setups can't.
  • Two stream modes — 60fps WebRTC (Moonlight) for quality, VNC-over-HTTPS for reliability on restrictive networks.
  • Auto-recorded demos — every Lux run produces a video artifact.

Prereqs

curl -fsSL https://box.ascii.dev/install | sh, then box login "$BOX_API_KEY" --json.

Limits

EU-only; 100 active VMs soft cap. For massive parallel headless browser fleets, dedicated browser providers (Browserbase) scale further; Box is best when you need a full machine behind the browser.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 9d ago First seen · 35 lines · 80 tokens per session scan C 8a4d5743e390

Subscribe to this mod's changes

computer-use-browser-desktop is a skill published in the GitHub repository ariana-dot-dev/cloud-sandbox-vm-skills (2 stars, last pushed 2mo ago), licensed MIT. It adds 80 tokens to every session and 438 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

creating-bookmarklets

Creates browser-executable JavaScript bookmarklets with strict formatting requirements. Use when users mention bookmarklets, browser utilities, dragging code to bookmarks bar, or need JavaScript that runs when clicked in the browser toolbar.

oaustegard/claude-skills · 47 tokens

using-webctl

Browser automation via webctl CLI in Claude.ai containers with authenticated proxy support. Use when users mention webctl, browser automation, Playwright browsing, web scraping, or headless Chrome in container environments.

oaustegard/claude-skills · 44 tokens

fetching-blocked-urls

Retrieve clean markdown from URLs when webfetch fails. Converts pages via Jina AI reader service with automatic retry. Use when webfetch or curl returns 403, blocked, paywall, timeout, JavaScript-rendering errors, or empty content or user explicitly suggests using jina.

oaustegard/claude-skills · 62 tokens

download-xhs-videos

A process for downloading a Xiaohongshu creator’s videos to a local folder. Xiaohongshu, also called RedNote, is a Chinese social-media platform.

YijiaDuan/download-xhs-videos · 181 tokens

webclaw

Web extraction for AI agents. Scrape, crawl, map, batch, extract, summarize, diff, brand, search, and 28 site-specific extractors turn any URL into clean Markdown, text, or JSON. Runs locally with no key via the webclaw MCP server; set an optional WEBCLAWAPIKEY to handle bot-protected and JavaScript-rendered pages.…

0xMassi/webclaw-skill · 101 tokens

add-tavily-tool

Add Tavily Search and Extract as keyless remote MCP tools for selected NanoClaw agent groups. Use when installing Tavily web search or URL extraction without an API key.

nanocoai/nanoclaw · 41 tokens