jimeng-login

Warn

Audited by Gen Agent Trust Hub on Apr 25, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill includes instructions to extract sensitive session data using document.cookie via JavaScript execution.
  • [CREDENTIALS_UNSAFE]: The skill directs the agent to store these extracted authentication cookies in a local file at memory/jimeng-cookies-backup.json, which exposes authentication tokens to anything that can read the skill's memory context.
  • [COMMAND_EXECUTION]: The skill uses the evaluate functionality to execute arbitrary JavaScript code within the browser context to interact with page elements and retrieve session data.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 25, 2026, 05:49 PM