sonamu-frontend
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents hooks that ingest untrusted data from URL search parameters and utilizes this data in network mutation calls and request functions. This constitutes an attack surface where maliciously crafted external data could potentially influence the behavior of the agent.
- Ingestion points: The
useListParamshook inreferences/lists.mdreads from the application's search state (URL parameters). - Boundary markers: The documentation does not provide specific boundary markers or instructions for the agent to distinguish ingested data from instructions.
- Capability inventory: The skill includes network mutation hooks (
references/provider.md), service request functions (references/generated-services.md), and file generation CLI commands (references/scaffolded-views.md). - Sanitization: The framework employs Zod schemas to parse and validate search parameters and form inputs, providing a layer of validation for ingested data.
Audit Metadata