hr-business-partner
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external workforce data from CSV files through its included scripts.
- Ingestion points: The scripts
org_health_scorer.py,compensation_analyzer.py, andworkforce_dashboard.pyingest data via the--filecommand-line argument. - Boundary markers: Data is processed as structured CSV rows; however, there are no specific prompt-level delimiters since the analysis is performed by standalone Python code.
- Capability inventory: The scripts are limited to reading CSV files and outputting results to the standard output. They do not possess network access, file-writing permissions, or dynamic code execution capabilities (no
evalorexec). - Sanitization: The scripts implement basic data sanitization using helper functions like
safe_floatand structured date parsing to handle malformed input. - [COMMAND_EXECUTION]: The skill provides standard command-line instructions for executing its internal Python scripts. These commands are used for legitimate HR analysis tasks, such as pay equity reviews and organizational health scoring, and do not involve privilege escalation or remote command injection risks.
Audit Metadata