agentphone
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from voice call transcripts and SMS messages, providing a surface for instructions that could potentially manipulate the agent's behavior.\n
- Ingestion points: Webhook payload processing of 'transcript' and message 'body' content as shown in SKILL.md.\n
- Boundary markers: The provided code examples use minimal boundary markers or delimiters when interpolating external data into prompts.\n
- Capability inventory: The skill has network access (via curl and SDKs) and telephony capabilities (making calls and sending SMS).\n
- Sanitization: While the documentation mentions platform-level filtering, the example webhook implementations do not demonstrate explicit input sanitization or validation of the external content.
Audit Metadata