debugging-strategies

debugging-strategies is a skill for Claude Code, Codex from orlando-japan/claude-code-setting. It costs 37 tokens per session (704 once invoked), scanned A, original, MIT.

A step-by-step method for finding software bugs through reproduction, narrowing the search, forming hypotheses, and testing them with evidence.

In plain words
What is it for?
Use it to create minimal reproductions, use Git or code bisection, isolate bad inputs or outputs, and test competing explanations.
Why use it?
It reduces guesswork when a bug is intermittent, difficult to reproduce, or keeps returning after attempted fixes.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to create minimal reproductions, use Git or code bisection, isolate…

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/orlando-japan/claude-code-setting/debugging-strategies
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add orlando-japan/claude-code-setting --skill debugging-strategies
Clone the repo
git clone --depth 1 https://github.com/orlando-japan/claude-code-setting

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for debugging-strategies

README.md
[![agentmods](https://agentmods.dev/badge/skills/orlando-japan/claude-code-setting/debugging-strategies.svg)](https://agentmods.dev/skills/orlando-japan/claude-code-setting/debugging-strategies)
Your own site
<a href="https://agentmods.dev/skills/orlando-japan/claude-code-setting/debugging-strategies"><img src="https://agentmods.dev/badge/skills/orlando-japan/claude-code-setting/debugging-strategies.svg" alt="Measured on agentmods" height="20"></a>
Per session 37 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 704 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00037 $0.00704
Opus 5 $0.00018 $0.00352
Sonnet 5 $0.00007 $0.00141
Haiku 4.5 $0.00004 $0.00070

Measured 6d ago against content hash 312088e74ec6, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

debugging-strategies scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

templates/extra/skills/debugging-strategies/SKILL.md · 71 lines

How it starts

The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Debugging strategies

The fastest debugger is one that doesn't guess. Every step should be driven by evidence, not hunches.

Phase 1 — Reproduce

If you can't reproduce, you can't debug. Before any hypothesis:

  • Minimal reproduction: smallest input / sequence / env that triggers the bug.
  • Deterministic: same inputs, same bug, every time. If it's intermittent, find the hidden non-determinism first (timing, order, state).
  • Isolated: no other noise. Strip the scenario down.

If you can't make it reproduce, spend effort on reproduction tooling, not on fixing.

Phase 2 — Bisect

Once reproducible, narrow the search:

  • Git bisect if the bug is new — what commit introduced it?
  • Code bisect — comment out half the code path. Does the bug remain? Which half?
  • Data bisect — half the input. Does the bug remain with the smaller input?
  • Binary search in output — where does good data become bad data? Log at midpoints.

Each step halves the search space. Don't skip halves out of laziness.

Phase 3 — Hypothesis & test

State the hypothesis out loud: "I think X is happening because of Y."

Then ask: "What experiment distinguishes this from alternatives?"

  • A test that passes if X, fails if not-X.
  • A log line that proves which branch was taken.
  • A breakpoint with a specific inspection.

Run the experiment. Update your belief. Repeat.

Phase 4 — Fix at the root, not the symptom

When you find a misbehaving line, ask: "why is this line wrong?" Trace upward:

  • Is the input wrong? Why?
  • Is the caller wrong?
  • Is the caller's caller wrong?

Stop at the first cause that, if fixed, would prevent this class of bug, not just this instance.

Symptom patches (null-check the crash site, catch the error, retry the request) silence the bug without fixing it. Reserve them for "ship now, fix next week" — and actually do the fix next week.

Common traps

  • Confirmation bias. You form a hypothesis and only look at data confirming it. Ask: "what would prove me wrong?"
  • Changing two things at once. When multiple variables change simultaneously, you can't attribute the outcome. Change one, observe, then the next.
  • Fixing what isn't broken. The bug is somewhere; it's probably not where you first looked. Prove the bug is where you think before patching.
  • Stopping at "it works now." If you don't know why it works, you don't know why it'll break again.

Read the full file on GitHub · 71 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 6d ago First seen · 71 lines · 37 tokens per session scan A 312088e74ec6

Subscribe to this mod's changes

debugging-strategies is a skill published in the GitHub repository orlando-japan/claude-code-setting (2 stars, last pushed 3mo ago), licensed MIT. It adds 37 tokens to every session and 704 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.