root-cause-analyzer

Root cause analysis uses evidence-based reproduction to trace state and data flows, identify faulty transitions, and validate assumptions. It produces an evidence-backed report containing reproduction steps, execution traces, and confidence levels without proposing fixes.

Tracing state and data flows captures variable values, call sequences, and transformation points to locate where data becomes incorrect. This process identifies faulty state transitions by comparing expected versus actual behavior during systematic reproduction.

Checking for race conditions involves adding delays, running stress tests, and testing concurrency timing during reproduction. This detects intermittent failures by exposing timing vulnerabilities in state transitions and data access patterns.

Git history inspection links regressions to specific commits by reviewing recent changes and using blame analysis. This identifies when faulty behavior was introduced by tracing modifications to state transitions and data flow logic.

Reliable reproduction creates minimal, repeatable steps with exact inputs, expected versus actual behavior, and frequency notes. This captures intermittent failures by documenting precise conditions and executing systematic stress tests to trigger race conditions.

A root cause report includes reproduction steps, state and data traces, alternative explanations, confidence level, and missing information. It validates assumptions through evidence-backed analysis without proposing fixes or implementation changes.