StackSleuth

Debugging eats up way too much of a developer's day. You see a failing test, then you spend hours reproducing it, digging through git history, and guessing where the bug came from. StackSleuth does that grunt work for you. Here is how it works. You point StackSleuth at a failing test. It reproduces the failure inside a real IBM Bob IDE session, so you can see exactly what happened. Then it bisects your git history to track down the exact commit that broke things. Once it finds the culprit, it explains the root cause in plain words and proposes a fix. The important part: StackSleuth never touches your code on its own. Every fix waits for a human to approve it first. Only after you say yes does it apply the patch and re-run the tests to prove the failure is gone. We tested this on real code, not a toy example. In Bob IDE 2.2.0, StackSleuth ran 10/10 tests and found two real bugs: an off-by-one error in a bulk discount calculation and a wrong rate on the SAVE10 coupon. Both patches were reviewed and approved by a human before they were applied. Every step is recorded in an append-only evidence trail, so you can audit exactly what the agent did and why. No fake demos. The video shows real IDE sessions with real bugs getting fixed.