This repository is a synthetic benchmark and local Python package. It does not call external LLMs, make network requests, or require API keys.

Please report security issues privately to the repository maintainer. Do not open public issues for vulnerabilities involving command execution, data exposure, malicious generated artifacts, or package supply-chain risk.

If you need to report a security issue, contact: security@example.com

• Benchmark generation should remain offline and deterministic.
• Tests may spawn local Python subprocesses, but runtime benchmark code should not execute untrusted shell commands.
• Do not commit credentials, real customer data, or proprietary model outputs.