twilio-webhook-architecture
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill architecture is designed to ingest and process untrusted data from external messaging and voice sources, creating a surface for indirect prompt injection.
- Ingestion points: The logic processes external input from the
Bodyparameter of inbound HTTP POST requests withinSKILL.md(e.g., incoming SMS messages). - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when echoing input back in the TwiML responses.
- Capability inventory: The skill demonstrates capabilities for sending messages, placing voice calls, and configuring event sinks via the Twilio API.
- Sanitization: While the skill correctly promotes transport security via HMAC signature validation, it does not demonstrate content-level sanitization for the reflected message body in the basic quickstart examples.
Audit Metadata