analyze

Root cause analysis for runtime bugs requires cross-file reasoning and evidence falsification. You investigate competing hypotheses, rank evidence strength, and run discriminating probes to identify the true source rather than jumping to immediate fixes.

Evidence-driven diagnostics for performance anomalies involve structuring observations, ranking competing hypotheses, and gathering file:line evidence. You attempt to falsify leading explanations and execute discriminating probes to isolate the root cause of latency or throughput issues.

Dependency impact assessment requires cross-file reasoning because failures often stem from orchestration behavior or configuration mismatches across multiple files. Structured investigation maps evidence for and against hypotheses to specific file:line references to pinpoint the root cause.

Debugging architecture failures with competing hypotheses involves defining observations, listing explanations, and weighing evidence for and against each. You identify critical unknowns and execute discriminating probes to falsify incorrect assumptions and confirm the root cause.

Team-mode orchestration parallelizes root cause investigation by spawning tracer lanes for code-path, config/environment, and measurement mismatches. Lanes investigate concurrently and reconcile findings into a ranked synthesis of evidence and hypotheses.

Premature certainty during debugging limits investigation by skipping falsification of leading hypotheses. Without ranking evidence strength and identifying critical unknowns, investigators risk fixing symptoms instead of the root cause, leading to recurring runtime failures.