[teiferer]

All my sibling comments are missing the message here which is that if Claude can find security issues then it can avoid them right when writing the code, so it could just never commit anything containing a security issue.

[joquarky]

You're assigning human capabilities to fancy linear algebra.

[teiferer]

It's not a matter of its own capabilities but of how you query it.