investigate

Systematic debugging enforces root cause investigation before applying fixes, eliminating whack-a-mole debugging and preventing recurring incidents. It requires reading code paths, checking recent changes, and reproducing bugs deterministically.

To find the root cause of a 500 error or stack trace, follow a four-phase workflow: investigate, analyze, hypothesize, and implement. This requires reading code paths and validating hypotheses before shipping any fixes.

Debug production incidents without guessing by enforcing an Iron Law: no fixes ship without identifying the true root cause. This involves reproducing bugs deterministically and validating hypotheses through systematic error analysis.

The best way to troubleshoot regressions where code previously worked is checking recent code changes and reading execution paths. This structured approach ensures you identify the true root cause rather than applying superficial patches.

Root cause investigation supports long debugging sessions through context recovery. It restores recent project artifacts, checkpoints, and session history after context compaction, ensuring debugging continuity survives throughout extended troubleshooting.