This is common in the hype era of AI models that can look for security bugs. An open-source project I used to work on that is basically "distributed code execution as a service" keeps getting reports that the job submission function is a vulnerability. The reporters don't even understand what the project does.