Online Shopping

Security checks across malware telemetry and agentic risk

Overview

This skill openly uses stealth browser automation to bypass Cloudflare and other bot protections while persisting shopping sessions and automating checkout flows.

Avoid installing this skill unless you intentionally accept stealth automation that bypasses store bot protections. If you do use it, restrict it to sites that permit automation, run setup in an isolated environment, review every checkout step manually, and clear the persistent browser profile after use.

VirusTotal

64/64 vendors flagged this skill as clean.

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.

#
ASI02: Tool Misuse and Exploitation
High
What this means

Using this skill could cause the agent to interact with online stores in ways that bypass site protections and may violate site rules or trigger account restrictions.

Why it was flagged

The skill explicitly frames its browser automation around evading Cloudflare/bot-protection controls, not merely normal browsing.

Skill content
Uses a stealth browser (Patchright) to bypass bot detection.
Recommendation

Do not install or use this skill for sites where automation is not explicitly allowed; prefer normal browser workflows and require explicit user review for every store interaction.

#
ASI03: Identity and Privilege Abuse
Medium
What this means

Mistakes or overbroad use could affect carts, accounts, delivery choices, payment methods, or purchases on the user's behalf.

Why it was flagged

The skill is designed to act inside checkout flows and handle personal, delivery, and payment-selection steps for online purchases.

Skill content
Checkout — fill shipping/contact details, select delivery and payment
Recommendation

Only allow this type of automation with narrow site and item scope, and require explicit confirmation before filling personal data, selecting payment, placing orders, or modifying account state.

#
ASI06: Memory and Context Poisoning
Medium
What this means

Shopping account sessions and cart state may remain available to later runs of the skill, increasing the chance of unintended account actions or exposure of private shopping data.

Why it was flagged

Persistent browser state can preserve logged-in shopping sessions, carts, and site identifiers across separate agent tasks without a clear retention or isolation policy.

Skill content
Use persistent context (`launchPersistentContext`) — retains cookies, localStorage, and session state between runs.
Recommendation

Use separate per-site profiles, clear the persistent context after each task, and avoid storing account sessions or sensitive checkout data unless the user explicitly approves.

#
ASI04: Agentic Supply Chain Vulnerabilities
Low
What this means

Running setup changes the local system and global OpenClaw environment, and it trusts current upstream package/browser downloads.

Why it was flagged

The setup script installs system packages and unpinned external dependencies, including browser binaries, as a disclosed user-run setup step.

Skill content
sudo apt-get update && sudo apt-get install -y xvfb ... npm install patchright --legacy-peer-deps ... npx patchright install chrome
Recommendation

Review the setup script before running it, run it in an isolated environment if possible, and prefer pinned dependency versions.

#
ASI05: Unexpected Code Execution
Low
What this means

The script performs local command execution as part of launching the browser helper.

Why it was flagged

The helper script executes a local shell command to discover the npm global root; the shown command is fixed and purpose-aligned, but it is still shell execution.

Skill content
execSync('npm root -g 2>/dev/null').toString().trim()
Recommendation

Avoid expanding this pattern to user-controlled shell arguments, and inspect future updates for additional command execution.