sap-hana-ml

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were identified. The skill focuses on providing documentation and templates for the official hana-ml library.
  • [DATA_EXPOSURE]: Connection templates correctly utilize placeholders for sensitive credentials such as hostnames and passwords, adhering to security best practices.
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of the legitimate hana-ml package from PyPI and mentions standard system dependencies like graphviz through official package managers.
  • [INDIRECT_PROMPT_INJECTION]: The skill includes functionality for reading data from external sources (HANA tables, CSVs) and performing database writes. These capabilities are essential for its machine learning purpose. * Ingestion points: references/DATAFRAME_REFERENCE.md (import_csv_from), SKILL.md (conn.table). * Boundary markers: Absent. * Capability inventory: references/DATAFRAME_REFERENCE.md (save, execute_sql), references/VISUALIZERS.md (to_file). * Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 04:50 AM