The Agent Skills Directory

Indirect Prompt Injection (LOW): The skill's core functionality involves ingesting untrusted data from external sources, which constitutes a surface for indirect prompt injection.\n
Ingestion points: Data is read via gpd.read_file (supporting local files, ZIPs, and remote URLs), gpd.read_postgis, gpd.read_parquet, and gpd.read_feather in references/data-io.md.\n
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are provided in the examples.\n
Capability inventory: The skill enables file system writing (to_file), network requests (downloading data), and database execution (to_postgis).\n
Sanitization: No explicit sanitization or validation of external input data is included in the instructions.\n- External Downloads (SAFE): The skill instructs the agent to install standard and reputable libraries such as geopandas, folium, and pyarrow. These dependencies are trusted in the context of geospatial analysis.\n- Credentials Safety (SAFE): Database connection examples in references/data-io.md use generic placeholders (user, password) and do not expose hardcoded secrets.

geopandas