charitable-giving
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted transaction data which creates an attack surface for indirect prompt injection.
- Ingestion points: Merchant names and transaction descriptions retrieved from the transaction_search tool.
- Boundary markers: None; the instructions do not use delimiters to isolate external data from the prompt.
- Capability inventory: The skill utilizes the transaction_search tool to access financial records.
- Sanitization: None; transaction content is processed as raw text without validation.
Audit Metadata