Agent Browser Clawdbot 0

Security checks across static analysis, malware telemetry, and agentic risk

Overview

This is a coherent browser-automation skill, but it gives an agent broad logged-in browser control and session persistence through an external CLI, so it deserves careful review.

Install only if you trust the external agent-browser CLI and need an agent to automate websites. Avoid using it on personal, financial, admin, or production accounts unless you explicitly approve each sensitive action, and treat any saved auth.json or cookie output like a password.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

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

A saved auth state file could let future browser sessions act as the logged-in user on websites, and cookie output may expose sensitive session data if mishandled.

Why it was flagged

These commands persist and expose browser cookies/storage, which can include logged-in session material. The skill does not specify consent, scope, encryption, retention, or restrictions on when the agent may save or reload that authority.

Skill content
agent-browser state save auth.json        # Save cookies/storage
agent-browser state load auth.json        # Load (skip login)
...
agent-browser cookies                     # Get all
Recommendation

Use only with accounts you intend the agent to access, prefer test or isolated sessions, require explicit user approval before saving/loading auth state, and store/delete auth files securely.

What this means

The agent could click buttons, fill forms, change website state, or alter network behavior during browsing tasks.

Why it was flagged

The skill documents commands that can submit UI actions and intercept or mock network requests. This is expected for browser automation, but it can have real effects on websites if used without care.

Skill content
agent-browser click @e2
agent-browser fill @e3 "text"
...
agent-browser network route "**/api/*" --body '{"x":1}'
Recommendation

Set clear task boundaries and require confirmation before purchases, account changes, public posts, administrative actions, or network mocking against sensitive services.

What this means

Installing the CLI runs external code outside the reviewed skill artifacts and may change the local environment.

Why it was flagged

The skill relies on an external, unpinned global npm package and downloaded browser/system components, while the reviewed package itself contains no CLI code or install spec.

Skill content
npm install -g agent-browser
agent-browser install                     # Download Chromium
agent-browser install --with-deps         # Linux: + system deps
Recommendation

Verify the npm package and GitHub project before installing, pin trusted versions where possible, and run the browser automation in a constrained environment.