twilio-messaging-webhooks
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill handles inbound message content, which is a common attack surface for indirect prompt injection if the data is subsequently processed by an LLM.
- Ingestion points: The
Bodyparameter from inbound HTTP POST requests is captured in the code snippets withinSKILL.md(e.g.,request.form.get("Body")in Python andreq.body.Bodyin Node.js). - Boundary markers: The skill proactively includes a "Security" warning callout that explicitly instructs developers to isolate the message content as user input and avoid direct concatenation into system prompts.
- Capability inventory: The code provided is limited to standard web framework response handling and Twilio SDK usage; no dangerous capabilities like
eval,subprocess, or arbitrary file access are present. - Sanitization: The skill provides detailed implementation patterns for verifying the
X-Twilio-Signatureusing the Twilio SDK to ensure that requests originate from a trusted source, effectively mitigating request forgery risks.
Audit Metadata