root-cause-tracing

Root-cause tracing involves systematically working backward through the call chain from where an error occurs, asking 'what called this function?' at each level until you identify where invalid data originated. Add validation and fixes at the source rather than treating symptoms at the failure point.

Use bisection methodology to narrow down which test creates unwanted state. Trace backward through 5+ levels of function calls to locate the test accessing uninitialized variables or missing setup, then add guards to prevent the pollution at its source.

When an error deep in execution masks its true origin—such as a `.git` repository appearing unexpectedly—backward tracing with instrumentation across multiple layers reveals the actual trigger. This prevents recurring bugs from incomplete fixes applied only at symptom locations.

Multi-layer validation catches invalid data before it propagates through the call stack. Defensive guards at each layer stop problems at their source rather than allowing them to travel deeper, where they become harder to diagnose and fix.

Yes. Backward tracing works by collecting stack traces and instrumenting functions to log execution paths. This reveals which function in a deep chain introduced bad state, allowing you to fix the origin rather than patching intermediate failures.