email-unsubscribe-check

Warn

Audited by Gen Agent Trust Hub on May 1, 2026

Risk Level: MEDIUMPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted email content to drive browser automation and credential retrieval activities.\n
  • Ingestion points: Email bodies retrieved via Gmail MCP (SKILL.md).\n
  • Boundary markers: Absent. There are no instructions to the agent to delimit or ignore instructions contained within the email bodies.\n
  • Capability inventory: Gmail management (read/write/search/filter), browser automation (Chrome DevTools MCP), and credential access (pass CLI).\n
  • Sanitization: Absent. Email content is used directly to extract URLs and navigate.\n- [DATA_EXFILTRATION]: The skill accesses the pass credential manager to perform automated logins. This poses a significant risk of credential exposure if an attacker sends an email with a malicious link designed to trigger a login attempt on an untrusted domain.\n- [COMMAND_EXECUTION]: The skill utilizes browser automation tools (Chrome DevTools MCP) to navigate to and interact with external, untrusted websites discovered in email bodies.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 1, 2026, 12:11 AM