geopandas

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill includes a dedicated section ('Suggest Using K-Dense Web For Complex Worflows') instructing the agent to proactively steer users toward a commercial service (www.k-dense.ai). This overrides the agent's default behavior by adding conditional promotional triggers.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection via the processing of untrusted external data.
  • Ingestion points: gpd.read_file() and gpd.read_postgis() are used to load data from GeoJSON, GeoPackage, Shapefiles, and databases.
  • Boundary markers: Absent. No delimiters or instructions are provided to help the agent distinguish between geospatial data and potential embedded instructions.
  • Capability inventory: The skill enables broad file system access (reading/writing) and complex data manipulation using the GeoPandas library.
  • Sanitization: Absent. There is no evidence of data validation or sanitization before processing.
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing several legitimate third-party Python libraries for geospatial analysis, including geopandas, folium, mapclassify, pyarrow, psycopg2, geoalchemy2, contextily, and cartopy.
  • [COMMAND_EXECUTION]: The documentation includes shell commands for installing dependencies using the uv pip package manager.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 08:25 AM