diagnoselisted

# Diagnose — a disciplined loop Don't guess-and-patch. Prove the cause, then fix the cause. ## The loop 1. **Reproduce.** Get a reliable, minimal repro. If you can't reproduce it, you can't claim to have fixed it. Say so, and gather more info instead of guessing. 2. **Minimise.** Strip the repro to the smallest input/state that still triggers it. 3. **Hypothesise.** State a specific, falsifiable hypothesis about the cause. 4. **Instrument.** Add logging/assertions/measurements that will confirm or refute the hypothesis. Let the data decide. Don't pattern-match to a "usual suspect". 5. **Fix the cause, not the symptom.** Once the data proves the cause, make the smallest change that addresses it. 6. **Regression-test.** Add a test that would have caught this (see the `tdd` skill). Confirm the repro is gone and nothing else broke. ## For performance regressions - Measure first: find the actual hot path / bottleneck with data, not intuition. - Compare against a known-good baseline (a prior commit / tag) where possible. - Fix the dominant cost. Re-measure to confirm the win is real. ## Discipline - Never assume. Prove from returned data / measurements. - Never patch code you don't understand the failure of. - If the evidence contradicts your first hypothesis, drop it. Don't defend it.