sinch-rcs
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages communication over the RCS channel, which requires processing inbound message payloads and suggested-reply postbacks containing content generated by end-users.
- Ingestion points: Inbound RCS message content and webhook delivery reports (as identified in the Security section of
SKILL.md). - Boundary markers: The agent is explicitly instructed to treat inbound content as untrusted data and is forbidden from executing, evaluating, or interpolating this content into prompts.
- Capability inventory: The skill provides capabilities to perform network operations (HTTPS POST) to send various message types (text, cards, media) to the Sinch API using the provided scripts.
- Sanitization: The documentation requires the verification of webhook signatures (HMAC) and the sanitization of all inbound content prior to handling.
Audit Metadata