@exiavaganza@AliceInDisarray I think the safeguards are triggered by the user prompt, not by the LLM activity. It warned me about security issues several times while scanning my nginx and docker config files without issues, because it came across it while doing non-security, mundane tasks.
@AliceInDisarray And doing that is dangerous? Shouldn't every developer in the world be scanning their codebase for vulnerabilities? Or the issue is that only some people should be able to do that?
@rez0__ Ban smithing because knifes and swords are dangerous. Only the nobility should have access to metals. Let China build steam engines and nuclear reactors.