back
loading skill details...
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user…
Diagnose
A discipline for hard bugs. Skip phases only when explicitly justified.
When exploring the codebase, use the project's domain glossary to get a clear mental model of the relevant modules, and check ADRs in the area you're touching.
Phase 1 — Build a feedback loop
This is the skill. Everything else is mechanical. If you have a fast, deterministic, agent-runnable pass/fail signal for the bug, you will find the cause — bisection, hypothesis-testing, and instrumentation all just consume that signal. If you don't have one, no amount of staring at code will save you.
Spend disproportionate effort here. Be aggressive. Be creative. Refuse to give up.
Ways to construct one — try them in roughly this order
Failing test at whatever seam reaches the bug — unit, integration, e2e.
Curl / HTTP script against a running dev server.
CLI invocation with a fixture input, diffing stdout against a known-good snapshot.
Headless browser script (Playwright / Puppeteer) — drives the UI, asserts on DOM/console/network.
Replay a captured trace. Save a real network request / payload / event log to disk; replay it through the code path in isolation.
Throwaway harness. Spin up a minimal subset of the system (one service, mocked deps) that exercises the bug code path with a single function call.
Property / fuzz loop. If the bug is "sometimes wrong output", run 1000 random inputs and look for the failure mode.
Bisection harness. If the bug appeared between two known states (commit, dataset, version), automate "boot at state X, check, repeat" so you can git bisect run it.
Differential loop. Run the same input through old-version vs new-version (or two configs) and diff outputs.
HITL bash script. Last resort. If a human must click, drive them with scripts/hitl-loop.template.sh so the loop is still structured. Captured output feeds back to you.
Build the right feedback loop, and the bug is 90% fixed.
Iterate on the loop itself
Treat the loop as a product. Once you have a loop, ask:
Can I make it faster? (Cache setup, skip unrelated init, narrow the test scope.)
Can I make the signal sharper? (Assert on the specific symptom, not "didn't crash".)
Can I make it more deterministic? (Pin time, seed RNG, isolate filesystem, freeze network.)
A 30-second flaky loop is barely better than no loop. A 2-second deterministic loop is a debugging superpower.
Non-deterministic bugs
The goal is not a clean repro but a higher reproduction rate. Loop the trigger 100×, parallelise, add stress, narrow timing windows, inject sleeps. A 50%-flake bug is debuggable; 1% is not — keep raising the rate until it's debuggable.
When you genuinely cannot build a loop
Stop and say so explicitly. List what you tried. Ask the user for: (a) access to whatever environment reproduces it, (b) a captured artifact (HAR file, log dump, core dump, screen recording with timestamps), or (c) permission to add temporary production instrumentation. Do not proceed to hypothesise without a loop.
Do not proceed to Phase 2 until you have a loop you believe in.
Phase 2 — Reproduce
Run the loop. Watch the bug appear.
Confirm:
The loop produces the failure mode the user described — not a different failure that happens to be nearby. Wrong bug = wrong fix.
The failure is reproducible across multiple runs (or, for non-deterministic bugs, reproducible at a high enough rate to debug against).
You have captured the exact symptom (error message, wrong output, slow timing) so later phases can verify the fix actually addresses it.
Do not proceed until you reproduce the bug.
Phase 3 — Hypothesise
Generate 3–5 ranked hypotheses before testing any of them. Single-hypothesis generation anchors on the first plausible idea.
Each hypothesis must be falsifiable: state the prediction it makes.
Format: "If is the cause, then will make the bug disappear / will make it worse."
If you cannot state the prediction, the hypothesis is a vibe — discard or sharpen it.
Show the ranked list to the user before testing. They often have domain knowledge that re-ranks instantly ("we just deployed a change to #3"), or know hypotheses they've already ruled out. Cheap checkpoint, big time saver. Don't block on it — proceed with your ranking if the user is AFK.
Phase 4 — Instrument
Each probe must map to a specific prediction from Phase 3. Change one variable at a time.
Tool preference:
Debugger / REPL inspection if the env supports it. One breakpoint beats ten logs.
Targeted logs at the boundaries that distinguish hypotheses.
Never "log everything and grep".
Tag every debug log with a unique prefix, e.g. [DEBUG-a4f2]. Cleanup at the end becomes a single grep. Untagged logs survive; tagged logs die.
Perf branch. For performance regressions, logs are usually wrong. Instead: establish a baseline measurement (timing harness, performance.now(), profiler, query plan), then bisect. Measure first, fix second.
Phase 5 — Fix + regression test
Write the regression test before the fix — but only if there is a correct seam for it.
A correct seam is one where the test exercises the real bug pattern as it occurs at the call site. If the only available seam is too shallow (single-caller test when the bug needs multiple callers, unit test that can't replicate the chain that triggered the bug), a regression test there gives false confidence.
If no correct seam exists, that itself is the finding. Note it. The codebase architecture is preventing the bug from being locked down. Flag this for the next phase.
If a correct seam exists:
Turn the minimised repro into a failing test at that seam.
Watch it fail.
Apply the fix.
Watch it pass.
Re-run the Phase 1 feedback loop against the original (un-minimised) scenario.
Phase 6 — Cleanup + post-mortem
Required before declaring done:
Original repro no longer reproduces (re-run the Phase 1 loop)
Regression test passes (or absence of seam is documented)
All [DEBUG-...] instrumentation removed (grep the prefix)
Throwaway prototypes deleted (or moved to a clearly-marked debug location)
The hypothesis that turned out correct is stated in the commit / PR message — so the next debugger learns
Then ask: what would have prevented this bug? If the answer involves architectural change (no good test seam, tangled callers, hidden coupling) hand off to the /improve-codebase-architecture skill with the specifics. Make the recommendation after the fix is in, not before — you have more information now than when you started.don't have the plugin yet? install it then click "run inline in claude" again.
restructured the original into implexa's 6-component format, added explicit decision points for non-deterministic bugs and seam selection, documented edge cases (flaky loops, missing artifacts, performance vs correctness debugging), embedded all ranked methods as sub-steps with io contracts, clarified when to halt without a loop, and kept mattpocock's voice and discipline intact.
diagnose is a structured framework for tracking down root causes in hard bugs and performance regressions. use it when you hit a bug that doesn't yield to casual inspection, or when a performance regression appeared but the cause isn't obvious. the core insight: build a fast, deterministic feedback loop first (reproduce the bug reliably in under 2-3 seconds), then hypothesis-test against it. skip phases only when explicitly justified. the discipline prevents thrashing, anchoring bias, and the trap of "let me just stare at the code."
phase 1: build a feedback loop
choose a feedback loop construction method from the ranked list below. pick the fastest deterministic option first; fall back to slower methods only if earlier ones fail.
ranked methods (try in order):
iterate on the loop itself. ask:
sub-steps:
for non-deterministic bugs, raise the reproduction rate until debuggable:
sub-steps:
if you cannot build a loop: stop, state explicitly what you tried, ask the user for (a) access to the reproducing environment, (b) a captured artifact (har file, log dump, core dump, screen recording with timestamps), or (c) permission to add temporary production instrumentation. do not proceed to phase 2 without a loop.
phase 2: reproduce
run the loop. watch the bug appear.
confirm all three:
phase 3: hypothesise
generate 3-5 ranked hypotheses before testing any of them. do not anchor on the first plausible idea.
sub-steps:
phase 4: instrument
probe each hypothesis methodically. change one variable at a time.
tool preference (in order of power):
phase 5: fix + regression test
write the regression test before the fix, but only if a correct seam exists.
a correct seam is one where the test exercises the real bug pattern as it occurs at the call site. if the only available seam is too shallow (single-caller test when the bug needs multiple callers, unit test that can't replicate the chain that triggered the bug), a regression test there gives false confidence. if no correct seam exists, that itself is the finding: document it. the codebase architecture is preventing the bug from being locked down; flag this for the /improve-codebase-architecture skill.
sub-steps (if a correct seam exists):
phase 6: cleanup + post-mortem
complete the required cleanup before declaring done.
required:
ask: what would have prevented this bug? if the answer involves architectural change (no good test seam, tangled callers, hidden coupling), hand off to the /improve-codebase-architecture skill with the specifics. make this recommendation after the fix is in, not before; you have more information now than when you started.
phase 1 breakpoint: if you cannot build a feedback loop after trying all ranked methods, halt. do not proceed to phase 2. explicitly list what you tried and ask the user for environment access, a captured artifact, or permission to add production instrumentation. proceeding without a loop is debugging in the dark.
non-deterministic bug fork: if the bug is intermittent (race condition, timing-dependent, flaky), do not try to build a deterministic repro. instead, optimize for reproduction rate (50%+ triggers the bug reliably). once the rate is high enough, proceed to phase 2. if it stays below 10%, loop back to phase 1 and request environment access or captured traces.
phase 3 checkpoint: before probing hypotheses in phase 4, show the ranked list to the user if they are available. if they are afk, proceed with your ranking. this is optional but saves time; domain knowledge re-ranks instantly.
phase 5 seam decision: before writing a regression test, ask: does a correct seam exist? a seam is correct if it exercises the bug at the call site where it manifests. if the only available seam is too shallow (mocked dependencies, single caller, isolated unit test), skip the regression test and document why. note this finding in the pr message so the team knows the architecture prevents proper test coverage.
phase 5 performance regression fork: if the bug is a performance regression, do not rely on logs to debug it. establish a baseline measurement first (profiler, query plan, timing harness). measure before hypothesizing. apply fixes only after a baseline is locked in.
on successful diagnosis:
data format:
file location:
the skill worked if:
the skill failed if: