investigate

Root-cause debugging for failing tests requires reproducing the symptom in a read-only phase, analyzing stack traces, and verifying fixes with a regression test. This enforces evidence-driven diagnosis before applying code changes.

The best way to debug intermittent runtime errors is through stepwise diagnosis and evidence collection. This approach enforces a reproduce, diagnose, and hypothesize workflow to identify the root cause, requiring a test-first regression verification before implementing the actual fix.

Yes, you can use git bisect to support root-cause debugging for unexplained defects and failing tests. It integrates into the evidence-driven diagnosis phase to help locate the specific commit causing the regression before you design a specific fix and regression test.

To prevent shotgun debugging when fixing runtime errors, enforce an investigation-first workflow with guardrails against guesswork. This requires read-only evidence collection, stepwise diagnosis, and test-first regression verification to ensure fixes target the actual root cause.

Yes, root-cause debugging works for TDD workflows by aligning with test-first principles. It requires implementing a regression test for every change, ensuring that fixes for failing tests and runtime errors preserve test coverage and minimize blast radius.

You should avoid symptom-only fixes during code-level debugging when you cannot reproduce the symptom or collect read-only evidence. Without stepwise diagnosis and a regression test, symptom-only fixes risk unresolved root causes and unverified changes to the code flow.