expense-report
Warn
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: MEDIUMPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [PERSISTENCE]: The skill can establish recurring scheduled tasks to perform automated processing and reporting. This allows the skill's logic to persist and run in the background across different sessions.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided receipt files (images, PDFs, screenshots), creating an attack surface for indirect prompt injection. \n
- Ingestion points: Files found in folders such as receipts/, expenses/, or user-specified cloud-sync directories. \n
- Boundary markers: No explicit markers or instructions are provided to the model to ignore potential commands hidden within the extracted receipt text. \n
- Capability inventory: The skill can write files (.xlsx, .html, .md) and schedule system tasks. \n
- Sanitization: There is no evidence of sanitization or filtering of the text extracted from receipts before it is used to populate reports or configuration files.
Audit Metadata