Code PluginExecutes codesource-linked

Browser Use

Browser automation plugin for OpenClaw via browser-use CLI

Community code plugin. Review compatibility and verification before install.

browser-use-plugin · runtime id browser-use

Install

openclaw plugins install clawhub:browser-use-plugin

Latest Release

Version 1.0.0

Download zip

Compatibility

{
  "builtWithOpenClawVersion": "1.0.0",
  "pluginApiRange": ">=2026.3.23"
}

Capabilities

{
  "bundledSkills": [],
  "capabilityTags": [
    "executes-code"
  ],
  "channels": [],
  "commandNames": [],
  "configSchema": true,
  "configUiHints": false,
  "executesCode": true,
  "hooks": [],
  "httpRouteCount": 0,
  "materializesDependencies": false,
  "providers": [],
  "runtimeId": "browser-use",
  "serviceNames": [],
  "setupEntry": false,
  "toolNames": []
}

Security Scan

VirusTotal

Pending

View report →

OpenClaw

Suspicious

medium confidence

✓

Purpose & Capability

Name/description align with the implementation: the plugin simply invokes a separate 'browser-use' CLI to perform browser automation. Requiring the external CLI and mentioning Python 3.11+ is coherent for a CLI that may be Python-based.

Instruction Scope

SKILL.md instructs the operator to install the CLI by piping a remote script (curl -fsSL https://browser-use.com/cli/install.sh | bash). The tool also exposes actions like 'eval' that run arbitrary JavaScript in page contexts — expected for a browser automation tool but high-privilege. The instructions do not request unrelated environment variables or file reads, but they do direct you to execute remote code to install the CLI.

Install Mechanism

No built-in install spec in the top-level registry; installation guidance points to running a shell script hosted at https://browser-use.com/cli/install.sh via curl|bash. Download-and-run from an unverified domain is a high-risk install mechanism. The skill's own metadata also references a 'uv' installer in its SKILL.md, but the visible, explicit instructions recommend the remote script.

✓

Credentials

The plugin does not request environment variables, credentials, or config paths. It only spawns the external 'browser-use' process and returns its JSON output; this is proportionate to the stated function.

✓

Persistence & Privilege

always is false and the plugin registers an optional tool. It does not request permanent/always-on privileges or attempt to modify other skills. Autonomous invocation is allowed by default (normal for skills) but not elevated here.

What to consider before installing

This plugin is coherent with its stated purpose — it wraps an external 'browser-use' CLI to automate browsers. However, the authors instruct you to install the CLI by piping a script from https://browser-use.com into bash, which is high risk because that runs remote code with your privileges. Before installing, review the install script (open the URL in your browser) or obtain the CLI from a trusted/reviewed source (GitHub Releases, your distro/package manager, or the plugin registry 'uv' installer if available). Be aware that the tool can run arbitrary JavaScript on pages (the 'eval' action) and navigate or interact with sites, so do not point it at sensitive internal sites or reuse credentials without caution. If you lack confidence in the CLI source, test in an isolated environment (sandbox/VM) or decline installation.

✗

dist/index.js:16

Shell command execution detected (child_process).

✗

src/index.ts:23

Shell command execution detected (child_process).

Patterns worth reviewing

These patterns may indicate risky behavior. Check the VirusTotal and OpenClaw results above for context-aware analysis before installing.

Verification

{
  "hasProvenance": false,
  "scanStatus": "pending",
  "scope": "artifact-only",
  "sourceCommit": "ea475df45fb44a84cfc1221b7a2673d2cbb9d9a6",
  "sourceRepo": "ShawnPana/browser-use-openclaw-plugin",
  "sourceTag": "main",
  "summary": "Validated package structure and linked the release to source metadata.",
  "tier": "source-linked"
}

browser-use plugin for OpenClaw

Browser automation plugin that gives your OpenClaw agent the ability to navigate websites, click elements, fill forms, take screenshots, and extract data.

Prerequisites

Install the browser-use CLI:

# macOS / Linux
curl -fsSL https://browser-use.com/cli/install.sh | bash

# Windows (PowerShell)
& "C:\Program Files\Git\bin\bash.exe" -c 'curl -fsSL https://browser-use.com/cli/install.sh | bash'

Validate your installation:

browser-use doctor

Requires Python 3.11+.

Install

openclaw plugins install browser-use

Enable the plugin and allow the tool:

openclaw plugins enable browser-use
openclaw config set tools.allow '["browser_use"]'

Usage

Once installed, your agent can use the browser_use tool to automate browser interactions. Ask it to:

"Open https://example.com and tell me the page title"
"Go to Hacker News and summarize the top 3 stories"
"Fill out the contact form on example.com"
"Take a screenshot of the current page"

Configuration

Optional settings in your OpenClaw config:

Setting	Type	Default	Description
`headed`	boolean	false	Show browser window
`timeoutMs`	number	30000	Command timeout (ms)

License

MIT