playwright-mcp

Warn

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates the execution of arbitrary JavaScript logic through the CLI tool's --script flag, providing a direct interface for running unvalidated code strings.
  • [REMOTE_CODE_EXECUTION]: The integration of page.evaluate() and similar browser automation functions allows for the dynamic execution of code within a browser session, which can be exploited if malicious payloads are introduced from untrusted web pages or inputs.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process content from external web pages (ingestion point: URLs accessed via the MCP server) without utilizing boundary markers, content sanitization, or explicit instructions to ignore embedded commands (Capability Inventory: browser navigation, element interaction, and dynamic script execution).
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 7, 2026, 05:44 PM