playwright-mcp

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (LOW): The skill configures an MCP server using npx -y @playwright/mcp. This command automatically downloads and executes the package from the npm registry. While this is a remote execution pattern, the @playwright scope is maintained by Microsoft, which is a trusted organization, justifying a LOW severity level.
  • INDIRECT_PROMPT_INJECTION (LOW): The skill is designed to ingest and process data from external websites, creating a vulnerability to indirect prompt injection.
  • Ingestion points: browser_snapshot (reads accessibility tree), browser_console_messages (reads JS logs), and browser_wait_for (reads page text).
  • Boundary markers: Absent. There are no instructions for the agent to ignore or delimit content found on the pages.
  • Capability inventory: The skill can perform active web interactions including browser_navigate, browser_click, and browser_fill_form.
  • Sanitization: Absent. Data from the browser is passed directly to the agent context without filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 11:27 PM