polars
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents methods for the agent to ingest and process untrusted data from various external sources, creating a potential surface for indirect prompt injection attacks.
- Ingestion points: SKILL.md and references/io_guide.md describe functions like pl.read_csv, pl.scan_csv, pl.read_parquet, pl.read_json, pl.read_excel, and pl.read_database for importing data.
- Boundary markers: No specific delimiters or instructions to ignore embedded prompts are provided in the documentation.
- Capability inventory: The documented library provides extensive file system access (reading and writing CSV, Parquet, JSON, Arrow) and network capabilities (connecting to PostgreSQL, MySQL, SQLite, AWS S3, Azure Blob Storage, and Google Cloud Storage).
- Sanitization: The skill does not discuss data sanitization or validation of input content before processing.
- [EXTERNAL_DOWNLOADS]: The documentation includes instructions for installing the Polars library and its optional dependencies from the standard Python package registry.
- Evidence: uv pip install "polars==1.41.2" and uv pip install "polars[excel,database,fsspec,pandas,numpy]==1.41.2" in SKILL.md.
Audit Metadata