swarm-debug — systematic debugging

No fix without a reproduced cause. Shotgun patching is how the same bug costs twice.

The loop

1. Reproduce — a command/test that shows the bug, or record explicitly why reproduction is impossible. This is the entry gate for everything below.
2. Evidence — read the FULL error, stack, and surrounding logs (not the first line); note what changed recently (git log, new deps, config).
3. Locate — binary-search the causal chain: which layer last had correct data? Instrument with targeted prints/asserts if needed; remove them after.
4. Hypothesize — ranked list, cheapest-to-test first. State each as a falsifiable claim ("the cache returns stale X when Y").
5. Verify — one targeted experiment per hypothesis. Evidence kills or confirms; opinion does neither.
6. Fix the root cause — not the symptom, when the root is reachable. If you must ship a symptom patch, say so in the code-note and open a ticket for the root.
7. Regression test — must fail on the pre-fix code, pass on the fix. No test, no fix.
8. Record (J1) — non-trivial diagnosis → memory note (the footgun, compact, ≤ 15 lines) + code-note link on the affected file, so no future session pays for this twice. Check the vault FIRST next time: swarmvault.py query --search "<symptom>" — someone may already have paid.

Hard rules

• Three dead hypotheses → stop, step back, re-examine assumptions — including the spec and design ("is the requirement itself contradictory?").
• A diagnosis that reveals a spec/design gap → route it to the question queue / SRS change management (swarm-spec), don't silently patch around it.
• Performance bugs: measure before and after; a fix without numbers is a guess.

Influences: superpowers systematic-debugging; Pocock's diagnosing-bugs; Jeffallan's debugging-wizard — see CREDITS.md.