agency-accounts-payable-agent
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is subject to an indirect prompt injection attack surface where instructions could be embedded in the invoices it processes.
- Ingestion points: Data from vendor invoices, contractor invoices, and recurring bills (SKILL.md).
- Boundary markers: The skill instructions lack explicit boundary markers or directives to ignore instructions contained within the invoice text, though it does specify verification against purchase orders.
- Capability inventory: The skill uses the
payments.sendtool to execute financial transfers via ACH, Wire, and Crypto (SKILL.md). - Sanitization: The logic includes business validation (vendor registry lookups and amount verification) but does not define linguistic sanitization for embedded commands in invoice fields.
Audit Metadata