diagnose

Name: diagnose
Author: stevesolun

560stars68forks

How to install diagnose?

stevesolun/ctx/diagnose

$npx -y skills add stevesolun/ctx --skill diagnose

Installs into the current project.

Summary

Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance r

Files · 1

View on GitHub

SKILL.md

1	# Diagnose
2
3	A discipline for hard bugs. Skip phases only when explicitly justified.
4
5	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.
6
7	## Phase 1 — Build a feedback loop
8
9	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.
10
11	Spend disproportionate effort here. Be aggressive. Be creative. Refuse to give up.
12
13	### Ways to construct one — try them in roughly this order
14
15	1. Failing test at whatever seam reaches the bug — unit, integration, e2e.
16	2. Curl / HTTP script against a running dev server.
17	3. CLI invocation with a fixture input, diffing stdout against a known-good snapshot.
18	4. Headless browser script (Playwright / Puppeteer) — drives the UI, asserts on DOM/console/network.
19	5. Replay a captured trace. Save a real network request / payload / event log to disk; replay it through the code path in isolation.
20	6. 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.
21	7. Property / fuzz loop. If the bug is "sometimes wrong output", run 1000 random inputs and look for the failure mode.
22	8. 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.
23	9. Differential loop. Run the same input through old-version vs new-version (or two configs) and diff outputs.
24	10. 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.
25
26	Build the right feedback loop, and the bug is 90% fixed.
27
28	### Iterate on the loop itself
29
30	Treat the loop as a product. Once you have _a_ loop, ask:
31
32	- Can I make it faster? (Cache setup, skip unrelated init, narrow the test scope.)
33	- Can I make the signal sharper? (Assert on the specific symptom, not "didn't crash".)
34	- Can I make it more deterministic? (Pin time, seed RNG, isolate filesystem, freeze network.)
35
36	A 30-second flaky loop is barely better than no loop. A 2-second deterministic loop is a debugging superpower.
37
38	### Non-deterministic bugs
39
40	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.
41
42	### When you genuinely cannot build a loop
43
44	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.
45
46	Do not proceed to Phase 2 until you have a loop you believe in.
47
48	## Phase 2 — Reproduce
49
50	Run the loop. Watch the bug appear.
51
52	Confirm:
53
54	- [ ] The loop produces the failure mode the user described — not a different failure that happens to be nearby. Wrong bug = wrong fix.
55	- [ ] The failure is reproducible across multiple runs (or, for non-deterministic bugs, reproducible at a high enough rate to debug against).
56	- [ ] You have captured the exact symptom (error message, wrong output, slow timing) so later phases can verify the fix actually addresses it.
57
58	Do not proceed until you reproduce the bug.
59
60	## Phase 3 — Hypothesise
61
62	Generate 3–5 ranked hypotheses before testing any of them. Single-hypothesis generation anchors on the first plausible idea.
63
64	Each hypothesis must be falsifiable: state the prediction it makes.
65
66	> Format: "If <X> is the cause, then <changing Y> will make the bug disappear / <changing Z> will make it worse."
67
68	If you cannot state the prediction, the hypothesis is a vibe — discard or sharpen it.
69
70	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.
71
72	## Phase 4 — Instrument
73
74	Each probe must map to a specific prediction from Phase 3. Change one variable at a time.
75
76	Tool preference:
77
78	1. Debugger / REPL inspection if the env supports it. One breakpoint beats ten logs.

People who liked this also viewed

6 picks

Type

diagnose

560stars68forks

View on GitHub

How to install diagnose?

stevesolun/ctx/diagnose

$npx -y skills add stevesolun/ctx --skill diagnose

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Diagnose
2
3	A discipline for hard bugs. Skip phases only when explicitly justified.
4
5	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.
6
7	## Phase 1 — Build a feedback loop
8
9	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.
10
11	Spend disproportionate effort here. Be aggressive. Be creative. Refuse to give up.
12
13	### Ways to construct one — try them in roughly this order
14
15	1. Failing test at whatever seam reaches the bug — unit, integration, e2e.
16	2. Curl / HTTP script against a running dev server.
17	3. CLI invocation with a fixture input, diffing stdout against a known-good snapshot.
18	4. Headless browser script (Playwright / Puppeteer) — drives the UI, asserts on DOM/console/network.
19	5. Replay a captured trace. Save a real network request / payload / event log to disk; replay it through the code path in isolation.
20	6. 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.
21	7. Property / fuzz loop. If the bug is "sometimes wrong output", run 1000 random inputs and look for the failure mode.
22	8. 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.
23	9. Differential loop. Run the same input through old-version vs new-version (or two configs) and diff outputs.
24	10. 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.
25
26	Build the right feedback loop, and the bug is 90% fixed.
27
28	### Iterate on the loop itself
29
30	Treat the loop as a product. Once you have _a_ loop, ask:
31
32	- Can I make it faster? (Cache setup, skip unrelated init, narrow the test scope.)
33	- Can I make the signal sharper? (Assert on the specific symptom, not "didn't crash".)
34	- Can I make it more deterministic? (Pin time, seed RNG, isolate filesystem, freeze network.)
35
36	A 30-second flaky loop is barely better than no loop. A 2-second deterministic loop is a debugging superpower.
37
38	### Non-deterministic bugs
39
40	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.
41
42	### When you genuinely cannot build a loop
43
44	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.
45
46	Do not proceed to Phase 2 until you have a loop you believe in.
47
48	## Phase 2 — Reproduce
49
50	Run the loop. Watch the bug appear.
51
52	Confirm:
53
54	- [ ] The loop produces the failure mode the user described — not a different failure that happens to be nearby. Wrong bug = wrong fix.
55	- [ ] The failure is reproducible across multiple runs (or, for non-deterministic bugs, reproducible at a high enough rate to debug against).
56	- [ ] You have captured the exact symptom (error message, wrong output, slow timing) so later phases can verify the fix actually addresses it.
57
58	Do not proceed until you reproduce the bug.
59
60	## Phase 3 — Hypothesise
61
62	Generate 3–5 ranked hypotheses before testing any of them. Single-hypothesis generation anchors on the first plausible idea.
63
64	Each hypothesis must be falsifiable: state the prediction it makes.
65
66	> Format: "If <X> is the cause, then <changing Y> will make the bug disappear / <changing Z> will make it worse."
67
68	If you cannot state the prediction, the hypothesis is a vibe — discard or sharpen it.
69
70	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.
71
72	## Phase 4 — Instrument
73
74	Each probe must map to a specific prediction from Phase 3. Change one variable at a time.
75
76	Tool preference:
77
78	1. Debugger / REPL inspection if the env supports it. One breakpoint beats ten logs.

People who liked this also viewed

6 picks

Type