israeli-healthcare-payroll

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides comprehensive instructions for understanding complex Israeli public-sector collective agreements for healthcare workers. It correctly identifies the 'Combined Salary' (sachar meshulav) as a base index and provides structural guidance on how various supplements (tosafot) are applied. The inclusion of historical (2008) data is clearly labeled as structural only, with explicit warnings against using it for current calculations, which demonstrates a high standard for data accuracy and safety. The skill maintains a clear separation between its informational role and the professional advice required for actual tax filing.
  • [COMMAND_EXECUTION]: The skill includes a local Python script scripts/healthcare_gross.py used to perform arithmetic calculations for gross salary. The script uses standard libraries, performs basic math on user-provided inputs, and contains no dangerous command execution or remote calls.
  • [EXTERNAL_DOWNLOADS]: All external links point to authoritative Israeli government sites (gov.il, btl.gov.il), official healthcare unions (ima.org.il), and well-known payroll service providers (malam-payroll.com). These resources are used strictly for reference and do not involve executable code downloads. In accordance with safety rules, these references to well-known and professional organizations are documented neutrally.
  • [DATA_EXFILTRATION]: There is no evidence of data exfiltration. The skill does not access sensitive local files (such as SSH keys or AWS credentials) and does not perform any unauthorized network operations. Its scope is limited to processing public salary structures.
  • [PROMPT_INJECTION]: The skill contains clear instructions for the agent to follow a specific four-step process. There are no attempts to bypass safety filters or override the agent's core behavioral constraints. The language is purely instructional and professional.
  • [CREDENTIALS_SAFE]: No hardcoded API keys, secrets, or credentials were found. The skill correctly instructs users to manage their own sensitive data and consult professional advisers.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 02:21 PM
Security Audit — agent-trust-hub — israeli-healthcare-payroll