Skip to main content
skill CX Codex

diagnose-hard-problem — Claude Code skill: install

The diagnose-hard-problem skill guides Claude Code through a structured bug investigation, starting with a repeatable failure signal and narrowing the case before testing possible causes. Use it for regressions and intermittent bugs.

Apache-2.0 ★ 173 updated 2026-10-24

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add sammcj/agentic-coding --skill diagnose-hard-problem -a claude-code
Open on GitHub

What diagnose-hard-problem does

The first task is to build a command or test that reproduces the user's exact symptom and returns a useful pass/fail signal. The skill suggests routes ranging from a failing test or CLI run to replaying a captured request or using a human-driven script. Secret values must be redacted from displayed output.

After reproducing the bug, the agent strips the case down to the smallest scenario that still fails. It then ranks testable explanations and checks them one at a time, with each probe tied to a prediction. The skill explicitly blocks moving on to hypotheses before a reproducible loop exists.

Install it for Claude Code with npx skills add sammcj/agentic-coding --skill diagnose-hard-problem -a claude-code; the skill file is at ~/.claude/skills/diagnose-hard-problem/SKILL.md. For flaky bugs, the loop should raise the reproduction rate before investigation continues.

What it helps with

  • reproduce a bug with a command or test
  • reduce a failing case to its smallest form
  • rank competing explanations for a regression
  • raise the reproduction rate of an intermittent failure
  • verify that a fix removes the reported symptom

Example prompts for the agent

  • Use diagnose-hard-problem to investigate this regression; first build a command that reproduces the reported symptom.
  • This failure is intermittent. Follow diagnose-hard-problem and make the reproduction loop catch it more often.
  • Reduce this bug to a minimal failing case with diagnose-hard-problem, then propose testable causes.

How to install diagnose-hard-problem

1Skills CLI (Vercel)

npx skills add sammcj/agentic-coding --skill diagnose-hard-problem -a claude-code

2Manually: put the files in place

~/.claude/skills/diagnose-hard-problem/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/diagnose-hard-problem/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about diagnose-hard-problem

How do I install diagnose-hard-problem in Claude Code?

Run npx skills add sammcj/agentic-coding --skill diagnose-hard-problem -a claude-code. The skill file is located at ~/.claude/skills/diagnose-hard-problem/SKILL.md.

What does diagnose-hard-problem do?

It guides a bug investigation through reproduction, minimisation, ranked hypotheses and targeted probes. The goal is to test the exact failure reported by the user.

How does the skill handle intermittent bugs?

It recommends repeating and stressing the trigger, adjusting timing, and increasing the reproduction rate so the failure can be investigated.

Source text

SKILL.md · Apache-2.0
---
name: diagnose-hard-problem
description: Disciplined diagnosis loop for hard problems, diagnosing bugs and regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test.
disable-model-invocation: true
metadata:
  source: https://github.com/mattpocock/skills (adapted)
---

# 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.

Note: In addition to this skill, you may consider activating the `systematic-debugging` skill when diagnosing complex, persistent issues.

## Redact

This skill has you show commands, outputs and captured artifacts. **Redact every secret first** - write `<REDACTED>` in its place. Build loops against env vars, so the credential stays in the environment rather than in what you show. Captured artifacts carry auth headers: quote only the lines that carry the signal.

If the redacted output is not enough to diagnose the bug, say so and ask the user.

## 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

1. **Failing test** at whatever seam reaches the bug - unit, integration, e2e.
2. **Curl / HTTP script** against a running dev server.
3. **CLI invocation** with a fixture input, diffing stdout against a known-good snapshot.
4. **Headless browser script** (Playwright / Puppeteer) - drives the UI, asserts on DOM/console/network.
5. **Replay a captured trace.** Save a real network request / payload / event log to disk; replay it through the code path in isolation.
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.
7. **Property / fuzz loop.** If the bug is "sometimes wrong output", run 1000 random inputs and look for the failure mode.
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.
9. **Differential loop.** Run the same input through old-version vs new-version (or two configs) and diff outputs.
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.

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 redacted 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.

### Completion criterion - a red-capable command

Phase 1 is done when you can name **one command** - a script path, a test invocation, a curl - that you have **already run at least once** (show the invocation and its output, redacted), and that is:

- [ ] **Red-capable** - it drives the actual bug code path and asserts the **user's exact symptom**, so it goes red on this bug and green once fixed. Not "runs without erroring" - it must be able to catch this specific bug.
- [ ] **Deterministic** - same verdict every run (flaky bugs: a pinned, high reproduction rate, per above).
- [ ] **Fast** - seconds, not minutes.
- [ ] **Agent-runnable** - you can run it unattended; a human in the loop only via `scripts/hitl-loop.template.sh`.

If you catch yourself reading code to build a theory before this command exists, **stop** - jumping straight to a hypothesis is the exact failure this skill prevents. Build the command first. No red-capable command, no Phase 2.

## Phase 2 - Reproduce + minimise

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.

### Minimise

Once it's reproduced, shrink the repro to the **smallest scenario that still goes red**. Cut inputs, callers, config, data, and steps **one at a time**, re-running the loop after each cut - keep only what's load-bearing for the failure.

Why bother: a minimal repro shrinks the hypothesis space in Phase 3 (fewer moving parts left to suspect) and becomes the clean regression test in Phase 5.

Done when **every remaining element is load-bearing** - removing any one of them makes the loop go green.

Do not proceed until you have reproduced **and** minimised 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 <X> is the cause, then <changing Y> will make the bug disappear / <changing Z> 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:

1. **Debugger / REPL inspection** if the env supports it. One breakpoint beats ten logs.
2. **Targeted logs** at the boundaries that distinguish hypotheses.
3. 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:

1. Turn the minimised repro into a failing test at that seam.
2. Watch it fail.
3. Apply the fix.
4. Watch it pass.
5. 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.

© sammcj · Apache-2.0 · sammcj/agentic-coding

Similar skills

All skills for Testing and QA →

From the same set73

The whole set →
ai-changelog ai-changelog sets up an AI-maintained CHANGELOG.md workflow across your project. Configures CalVer, SemVer, or simple date-based logging for non-code repositories. CX Codex +1 ★ 173repo stars rust-engineer The rust-engineer skill turns Claude Code into a Rust-focused coding assistant for ownership, async code, safety checks, testing, and performance work. CX Codex +1 ★ 173repo stars software-research-assistant software-research-assistant — a subagent that digs into how to actually implement a library, framework, SDK, or API: pulls code examples, version facts, compatibility, and current best practices from live sources. CX Codex ★ 173repo stars guided-demo guided-demo adds interactive, self-running guided walkthroughs and typed narration overlays to web apps and HTML pages. CX Codex +1 ★ 173repo stars yue2-music yue2-music — a Claude skill that carries a YuE2 song from request to finished audio: three generation modes, covers transcribed with SheetSage2, and a local player for A/B listening. CX Codex +1 ★ 173repo stars swift-development swift-development enforces modern Swift 6 guidelines, region-based concurrency checks, and clean API conventions for macOS and iOS codebases. CX Codex +1 ★ 173repo stars github-triage The github-triage skill sorts GitHub issues through labels and state changes, then guides maintainers through a decision. Use it for bugs, feature requests, and agent handoff. CX Codex +1 ★ 173repo stars authoring-claude-md authoring-claude-md is a Claude Code skill for creating and maintaining CLAUDE.md and .claude/rules/ files that record project context, architectural decisions, and non-obvious conventions. CX Codex +1 ★ 173repo stars

Project sponsors

Companies that keep this analytics open