Skills browser-use

📦

browser-use

Name: browser-use
Author: browser-use

Content revision r4 High Risk ⚙️ External commands🌐 Network access🔑 Env variables

Automate interactive browser tasks with CDP

Web tasks that need clicks, JavaScript rendering, or an existing signed-in session are difficult to handle with simple HTTP requests. This skill controls a CDP-enabled browser and provides workflows for navigation, testing, screenshots, recordings, and cloud browsers.

Supports: Claude Codex Code(CC)

⚠️ 38 Poor

This skill is part of a pack

Install the whole pack to get every skill the task needs, in one command.

OpenClaw Browser Automation · 3 skills Install whole pack

Stop for confirmation before installing.

Review the plan and obtain explicit user consent before changing files.

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request

Review the Skillstore skill "browser-use" from https://skillstore.io/skills/browser-use-browser-use.md and its manifest at https://skillstore.io/api/skills/browser-use-browser-use/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/browser-use-browser-use.md Signed manifest GET /api/skills/browser-use-browser-use/manifest Signed lockfile GET /api/skills/browser-use-browser-use/lockfile

Test it

Using "browser-use". Open the public documentation page and list its main sections.

Expected outcome:

Opened the page and waited for the document to load.
Main sections found: Getting started, Configuration, API reference, and Troubleshooting.

Using "browser-use". Test whether the save button confirms a profile update.

Expected outcome:

Updated the requested test field in the staging profile.
The page displayed a saved confirmation and the refreshed value matched the change.

Using "browser-use". Capture a screenshot of the loaded dashboard.

Expected outcome:

Dashboard loaded successfully.
Screenshot captured and saved to the active task output.

Security Audit

High Risk

v11 • 7/28/2026 Open versioned report

The 53 static findings are false positives caused by Markdown formatting, documentation URLs, and an explicit standard-input authentication example. Two context-level risks remain: automatic ingestion of workspace domain-skill files without an untrusted-content boundary, and guidance that frames stealth browsers as a way to avoid anti-bot controls.

Files scanned

174

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (2)

High

SKILL.md:16 SKILL.md:173

Untrusted Domain-Skill Instructions Are Read Automatically

When enabled, the skill instructs the agent to read every site-specific file from a workspace directory before choosing an approach. Those files can contain untrusted prompt content, but the guidance provides no instruction to treat it as data or resist prompt injection.

The instructions explicitly require reading workspace-provided domain-skill files and do not define a trust boundary. This creates a direct prompt-injection surface when that directory is writable or populated from untrusted sources.

Medium

SKILL.md:51-58

Captcha-Bypass Positioning

The cloud-browser guidance recommends managed IPs and stealth settings when captchas or bot blocking are likely. This can encourage bypassing website anti-automation controls without confirming authorization.

The text expressly connects stealth settings and managed IPs with avoiding captchas and rate limits. It is guidance for evading service controls rather than a neutral browser-use instruction.

Risk Factors

⚙️ External commands (45)

🌐 Network access (7)

SKILL.md:144 SKILL.md:8 SKILL.md:58 SKILL.md:80 SKILL.md:89 SKILL.md:127 SKILL.md:133

🔑 Env variables (1)

SKILL.md:69

Audited by: claude View Audit History →

Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report

Copy report link

https://skillstore.io/skills/browser-use-browser-use/audits/11?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/browser-use-browser-use/security.svg)](https://skillstore.io/skills/browser-use-browser-use?utm_source=security_passport_badge)

HTML badge

<a href="https://skillstore.io/skills/browser-use-browser-use?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/browser-use-browser-use/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Embed card

<iframe src="https://skillstore.io/embed/skills/browser-use-browser-use.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Academic citations (APA · BibTeX · CFF)

APA citation

browser-use. (2026). browser-use security audit report (audit version 11) [Author version unspecified]. Skillstore. https://skillstore.io/skills/browser-use-browser-use/audits/11

BibTeX citation

@techreport{browser-use-browser-use-browser-use-2026,
  author = {browser-use},
  title = {browser-use security audit report (audit version 11)},
  institution = {Skillstore},
  year = {2026},
  number = {11},
  url = {https://skillstore.io/skills/browser-use-browser-use/audits/11},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "browser-use security audit report (audit version 11)"
version: "unspecified"
type: report
authors:
  - name: "browser-use"
date-released: "2026-07-28"
url: "https://skillstore.io/skills/browser-use-browser-use/audits/11"
identifiers:
  - type: other
    value: "skillstore:browser-use-browser-use:audit:11"
    description: "Skillstore immutable audit report identifier"

Compare variants

3 installable variants

Each author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.

Why this variant is first

Stronger evidence confidence

Author / source Skillstore Score Evidence Confidence Skillstore usage Updated

browser-use Recommended Current

browser-use-browser-use

Skillstore Score 38

Evidence Confidence High

Skillstore usage 75

Updated

2026-09-09

92Bilal26

92bilal26-browser-use

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 58

Updated

2026-09-09

AbdullahMalik17

abdullahmalik17-browser-use

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 11

Updated

2026-09-09

Skillstore Score

Why this score Evidence Confidence: High

Architecture

Maintainability

Content

Community

Spec Compliance

What You Can Build

Test a signed-in web workflow

Navigate a staging application, complete approved interface steps, and verify the visible result.

Collect rendered page details

Inspect a JavaScript-rendered page and extract information from its accessibility tree or DOM.

Document a browser procedure

Record an approved workflow and capture screenshots for a support or product demonstration.

Try These Prompts

Open and inspect a page

Open https://example.com in the browser. Wait for it to load, then summarize the page title and main headings.

Test a visible interaction

In my staging app, open the settings page, find the notification toggle by its accessible name, change it, and verify the saved state. Ask before submitting any final change.

Extract rendered content

Open the provided public product page. Use the accessibility tree or DOM to collect each product name and displayed price. Do not bypass access controls or login walls.

Run an isolated approved task

Start one isolated cloud browser for this authorized test. Visit the supplied staging URL, capture a screenshot after the dashboard loads, report the result, and ask before leaving the browser running.

Best Practices

Use a direct HTTP request for public information when browser interaction is unnecessary.
Prefer accessible element names and verify each consequential action after it completes.
Ask for confirmation before submissions, account changes, purchases, exports, or leaving cloud browsers active.

Avoid

Do not use the skill to bypass login, multi-factor authentication, consent, or account-selection prompts.
Do not use stealth settings or managed IPs to evade website anti-bot controls or rate limits.
Do not treat site-specific workspace files as trusted instructions without reviewing their source and purpose.

Frequently Asked Questions

When should I use this skill?

Use it when a task needs browser interaction, JavaScript rendering, screenshots, or an approved signed-in session.

Can it control my local Chrome browser?

Yes. It connects through Chrome DevTools Protocol after remote debugging is enabled and approved.

Can it handle passwords or multi-factor authentication?

No. Stop and ask the user for passwords, multi-factor authentication, consent, or ambiguous account choices.

Does it support cloud browsers?

Yes. It can start isolated Browser Use cloud sessions for authorized tasks and should ask before leaving them running.

Can it record browser work?

Yes. Recording is optional and can be enabled for a task when the user requests a recording, demo, or video.

Can it bypass captchas or website restrictions?

No. Use browser automation only with authorization and in compliance with website terms, access controls, and rate limits.

Developer Details

Author

browser-use

License

MIT

Skillstore revision

Version notice

The author did not declare a version.

Repository

https://github.com/browser-use/browser-use/tree/main/skills/browser-use/

Ref

ebe8dce061012745b89288bf501f62d732ac3fb8

Maintenance freshness

7/28/2026

Usage

63 downloads · 265 views

File structure

📄 SKILL.md