vapi-webhooks

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to ingest and process data from Vapi webhooks, including transcripts of user speech and tool execution requests. This creates an indirect prompt injection surface where a remote caller could attempt to influence the AI agent's behavior by embedding instructions in the conversation that are then passed to the handler.
  • Ingestion points: The webhook endpoint /webhooks/vapi implemented in examples/express/src/index.js, examples/fastapi/main.py, and examples/nextjs/app/webhooks/vapi/route.ts.
  • Boundary markers: The skill does not explicitly use boundary markers or provide instructions to downstream agents to ignore embedded commands within the JSON payload fields.
  • Capability inventory: The skill allows for influencing live telephony sessions by returning JSON responses that control tool execution and assistant selection.
  • Sanitization: The implementation performs structure validation via JSON parsing but does not include sanitization or filtering for the textual content of the webhook messages.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 03:06 PM
Security Audit — agent-trust-hub — vapi-webhooks