Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add leeguooooo/chrome-use --skill openaigit clone --depth 1 https://github.com/leeguooooo/chrome-useWrote 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.
[](https://agentmods.dev/skills/leeguooooo/chrome-use/openai)<a href="https://agentmods.dev/skills/leeguooooo/chrome-use/openai"><img src="https://agentmods.dev/badge/skills/leeguooooo/chrome-use/openai/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.
<a href="https://agentmods.dev/skills/leeguooooo/chrome-use/openai"><img src="https://agentmods.dev/badge/skills/leeguooooo/chrome-use/openai.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00096 | $0.00668 |
| Opus 5.5 | $0.00038 | $0.00267 |
| Sonnet 5.5 | $0.00019 | $0.00134 |
| Haiku 4.5 | $0.00010 | $0.00067 |
Grade A, and why
chrome-use scanned grade A with 0 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 13d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
chrome-use
Drives the open-source chrome-use CLI, which controls the Chrome browser on the user's computer. It reads pages through Chrome's accessibility tree and gives each control a short ref (@e3), so actions target real elements instead of screen coordinates.
Before the first call
- Run
chrome-use --version. If the command is not found, stop and tell the user chrome-use must be installed on their computer first. Point them to https://github.com/leeguooooo/chrome-use#install; do not download or run an installer yourself. - Run
chrome-use skills get coreonce per conversation. It prints the usage guide that matches the installed version; follow it.
The action loop
chrome-use open https://example.com # open a page in a new or current tab
chrome-use read # the page as readable text
chrome-use snapshot -i # interactive controls with @refs
chrome-use click @e3 --observe # act, and see what changed
chrome-use fill @e2 "hello" --observe
chrome-use screenshot page.png
- Take a fresh
snapshot -ibefore acting on a new page; refs reset after navigation. - Pair actions with
--observeand read the result before acting again. "No change" does not prove failure; read thewhynote first. - Never repeat an action that may already have taken effect, such as a send, submit, purchase or delete. Check the page state instead.
Rules
- Act only on sites and tasks the user asked for, using their existing access. Do not create accounts, scrape at volume, or work around a site's rate limits or access controls.
- Before submitting a form, sending a message, buying something, or changing account settings, show the user what will happen and get their approval.
- Sensitive field values (card numbers, security codes, passwords, one-time codes) are masked in
snapshotandget valueoutput. Do not reveal them in the conversation. - If a page shows a CAPTCHA or a "verify you are human" check (chrome-use reports known ones as
blocked_by_human_check), or a sign-in page refuses the browser, stop and ask the user to complete that step in their own browser; do not try to get past it. - Identity checks, live selfies, and password-manager unlocks are always the user's to do.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 13d ago First seen · 42 lines · 96 tokens per session scan A 6c201005f328
chrome-use is a skill published in the GitHub repository leeguooooo/chrome-use (202 stars, last pushed today), licensed Apache-2.0. It adds 96 tokens to every session and 668 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. It comes from a forked repository.
Other skills, from other repositories
surf
Control Chrome browser via CLI for testing, automation, and debugging. Use when the user needs browser automation, screenshots, form filling, page inspection, network/CPU emulation, DevTools streaming, or AI queries via ChatGPT/Gemini/Perplexity/Grok/AI Studio.
Agent Browser Automation
Fast Rust-based headless browser automation CLI with Node.js fallback for AI agents, featuring navigation, clicking, typing, snapshots, and structured commands optimized for agent workflows.
Browser-Use Automation
CLI tool for persistent browser automation with multi-session support, featuring Chromium/Real/Remote browser modes, cookie management, JavaScript execution, and long-running automation workflows.
cursor-browser
Drive Cursor IDE's built-in Browser Tab via cursor-browser-cli. Navigate, ref snapshots, click/type/fill/scroll/select-option by ref, wait-for, lock, screenshot, console/network/inspect. For Grok Build, Claude Code, Codex in Cursor terminal. Triggers: Cursor browser, Browser Tab, /cursor-browser, snapshot, under the…
wbrowser
Use when a task needs a browser that is already signed in — reading a dashboard, checking a web app, filling a form, taking a screenshot of a real page, or debugging a site's console. Drives a real Chrome window the user signed into by hand, so no re-login and no API keys. Do not guess selectors; run wb read first and…
ultimate-playwright
Browser automation via ultimate-playwright-mcp. Use when the agent needs to control a shared Chrome browser — open tabs, navigate, click, type, fill forms, take snapshots. Handles multi-user tab group isolation so multiple sessions can share one browser without interference. Use for web browsing, scraping, form…