Playwright MCP

Security checks across malware telemetry and agentic risk

Overview

This is a coherent browser-automation skill, but it gives the agent broad website-control abilities and depends on an external npm MCP server that users should trust and scope carefully.

Install only if you trust the @playwright/mcp package and want the agent to control a browser. Prefer limiting allowed hosts, avoid unsupervised use on logged-in or high-value sites, and review any action that submits forms, uploads files, purchases items, changes account settings, or runs page JavaScript.

VirusTotal

66/66 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
Medium
What this means

The agent could perform real browser actions on websites, including actions that change account or form state, when a user asks it to automate those pages.

Why it was flagged

These tools are expected for Playwright browser automation, but they can submit forms, interact with accounts, or upload files if used on sensitive sites.

Skill content
`browser_click` | Click element by selector | ... `browser_type` | Type text into input | ... `browser_choose_file` | Upload file |
Recommendation

Use host allow-lists where possible, supervise actions on logged-in or sensitive sites, and require explicit review before form submission, purchases, uploads, or account changes.

#
ASI05: Unexpected Code Execution
Low
What this means

JavaScript evaluation can inspect or modify the current page and may interact with sensitive page state if run on logged-in sites.

Why it was flagged

The skill exposes JavaScript execution inside the browser page context. This is purpose-aligned for automation and extraction, but it is a powerful escape-hatch capability.

Skill content
`browser_evaluate` | Execute JavaScript |
Recommendation

Limit JavaScript evaluation to trusted pages and user-reviewed scripts, especially when authenticated sessions or sensitive data are present.

#
ASI04: Agentic Supply Chain Vulnerabilities
Low
What this means

Installing or running the external package trusts the npm package contents and whatever version is resolved at install time.

Why it was flagged

The installation path relies on an external npm package and npx execution without a pinned version in the artifact instructions.

Skill content
npm install -g @playwright/mcp
# Or
npx @playwright/mcp
Recommendation

Verify the package source, consider pinning a known-good version, and install from a trusted registry before granting the browser automation skill access to sensitive workflows.