hands-free

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). Outsider free text from the phone caller (user’s spoken reply) is captured by scripts/call_user.py from the Vapi call artifact/transcript (extract_user_answer reads artifact.messages / artifact.transcript) and then printed to stdout, which the agent treats as the user’s reply verbatim and feeds back into the LLM context.