pandas
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from various formats.
- Ingestion points:
pd.read_csv(),pd.read_excel(),pd.read_sql(),pd.read_parquet(), andpd.read_json()inSKILL.md. - Boundary markers: No explicit prompt boundaries are defined for the ingested data content.
- Capability inventory: Capabilities are limited to standard tabular data manipulation, transformation, and aggregation using the pandas library.
- Sanitization: Instructions focus on data cleaning (handling missing values, deduplication) but do not explicitly address sanitizing data content against prompt injection, which is standard for data processing libraries.
- [NO_CODE]: The skill package consists solely of documentation and configuration metadata. No executable scripts, binaries, or dynamic commands were found in the provided files.
Audit Metadata