shiny-spatial
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from SQLite database files, specifically via feature properties that can contain arbitrary text. Maliciously crafted data within these databases could attempt to influence the agent's behavior when retrieved and presented in the conversation context.
- Ingestion points: Loading database files through the
SpatialDatabaseconstructor and retrieving features usingSpatialTable.GetByIdorSpatialTable.Querymethods inSKILL.md. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the documentation for handling feature property strings.
- Capability inventory: The skill allows reading from the local file system (database files) and accessing real-time location data (GPS) for geofencing purposes as described in the
Shiny.Spatial.Geofencingsection. - Sanitization: The library does not specify sanitization or validation protocols for text-based property values stored in the WKB-encoded geometry features.
Audit Metadata