voice-agent-prompt

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists entirely of documentation, architectural diagrams, and prompt templates for building voice-driven conversational agents. It contains no executable code, remote downloads, or credential exposure.- [PROMPT_INJECTION]: The skill defines an architecture for an agent that processes untrusted voice input from customers, representing a surface for indirect prompt injection.
  • Ingestion points: Customer audio/text input ingested via Twilio SIP and LiveKit Room as described in the architecture diagram in SKILL.md.
  • Boundary markers: The provided system prompt templates do not include specific delimiters or boundary markers to isolate customer input from instructions.
  • Capability inventory: The agent has access to state-changing tools including submit_order, add_to_order, and book_appointment.
  • Sanitization: No input sanitization or validation logic is specified in the prompt documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 10:32 PM