geopandas
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents capabilities to ingest untrusted data from external sources and perform file/database writes, representing a theoretical attack surface for indirect prompt injection.\n
- Ingestion points:
read_file(handles local files, URLs, ZIP archives, and cloud storage) andread_postgis(database queries) documented inSKILL.mdandreferences/data-io.md.\n - Boundary markers: The documentation does not specify the use of delimiters or 'ignore' instructions for external content.\n
- Capability inventory: The skill documents file writing (
to_file), database writing (to_postgis), and interactive map generation (explore().save()) across multiple files.\n - Sanitization: No explicit sanitization or validation of external data is mentioned in the library usage examples.\n- [SAFE]: All documented workflows and code snippets are standard for the GeoPandas library and align with its intended purpose. Dependencies are official packages from the Python Package Index (PyPI). No attempts at prompt injection, obfuscation, privilege escalation, or persistence were found.
Audit Metadata