Security fixes are handled on the latest public version.

Please do not open a public issue for vulnerabilities that could expose private local data.

Instead, contact the maintainer privately through the contact method listed on the GitHub profile. Include:

• the affected version or commit,
• steps to reproduce,
• impact,
• any relevant logs with private prompts and model outputs removed.

Local Model Bench is designed to run on your own machine and talk to your own LM Studio server. Benchmark results are stored locally in the runs folder. Do not publish that folder unless you intentionally want to share your model names, outputs, timings, and scores.