geomaster

Pass

Audited by Gen Agent Trust Hub on Oct 24, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: A thorough review of the skill's instructions, metadata, and the bundled scripts/raster_workflows.py script revealed no malicious patterns, obfuscation, or unauthorized access to sensitive data.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process third-party geospatial data from various external APIs and catalogs, which represents a potential attack surface for indirect injection.
  • Ingestion points: Geospatial data and STAC metadata retrieved from providers described in SKILL.md and references/data-sources.md.
  • Boundary markers: Data is processed primarily as numeric arrays or structured GeoJSON/GeoPackage formats, which naturally separates data content from the agent's control flow.
  • Capability inventory: The skill has the capability to write local files via scripts/raster_workflows.py and perform network operations to access Earth observation catalogs.
  • Sanitization: Relies on established GIS libraries (Rasterio, Fiona, GeoPandas) that perform rigorous schema validation and format-specific parsing.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the download of configuration and imagery assets from well-known technology services and trusted organizations, including Microsoft (Planetary Computer), AWS (Earth Search), and the European Space Agency (Copernicus).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 24, 2026, 09:46 PM