so I'm guessing now I can finally say "go fix the bug" and the future agents can run blame on old tracked prompts, find the thought process behind a line themselves, and fix the bug?