Browser Js

WarnAudited by ClawScan on May 10, 2026.

Overview

The skill coherently provides browser automation, but it gives an agent broad control over a logged-in browser, including sensitive clicks, uploads, and page JavaScript.

Install only if you are comfortable letting an agent drive a CDP-enabled browser. Prefer a dedicated automation profile with minimal logins, supervise sensitive sites, and require confirmation before uploads, payments, OAuth/login prompts, public posts, or other account-changing actions.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

An agent using this skill could act as you on sites where your browser is already signed in.

Why it was flagged

The skill is designed to operate through the user's existing browser profile, so actions can be performed under already-authenticated web accounts.

Skill content
Requires a Chrome/Chromium browser running with --remote-debugging-port ... Signed-in sessions carry over automatically.
Recommendation

Use a separate limited browser profile for automation, avoid personal or financial accounts, and require explicit confirmation before account, payment, posting, or data-changing actions.

What this means

The agent may be able to click through sensitive browser flows such as OAuth prompts, payment widgets, or captcha-like controls if instructed or misled.

Why it was flagged

The documented coordinate commands are meant to interact with sensitive cross-origin frames and bot/security-related UI, without clear approval or containment rules.

Skill content
Coordinate commands (cross-origin iframes, captchas, overlays) ... captcha checkbox, payment form, OAuth widget ... These bypass all DOM boundaries.
Recommendation

Limit use to trusted tasks, supervise sensitive flows, and add explicit user-approval rules for payments, login/OAuth, captcha, account changes, and public submissions.

What this means

JavaScript executed in a logged-in page could read or change page state and trigger actions on that site.

Why it was flagged

The skill exposes a raw JavaScript evaluation command in the current page context; this is disclosed and browser-related, but it is an escape-hatch capability.

Skill content
bjs eval <js>               Run JavaScript in page
Recommendation

Use eval only for trusted, task-specific snippets and avoid running code copied from webpages or untrusted sources.

What this means

Installing dependencies adds normal package-supply-chain exposure.

Why it was flagged

The setup requires a manual npm dependency install even though the registry metadata lists no install spec; the provided package files show a small dependency, but users should still review it before running.

Skill content
cd scripts && npm install
Recommendation

Review scripts/package.json and scripts/package-lock.json before installation and run npm install only in the skill's scripts directory.