debug-detective

A structured software-debugging assistant that investigates bugs by reproducing, isolating, testing hypotheses, and verifying the cause.

In plain words
What is it for?
Use it for crashes, failing or flaky tests, regressions, inconsistent environments, deadlocks, memory leaks, race conditions, and other hard-to-reproduce problems.
Why use it?
It replaces guesswork with evidence, helping distinguish a real fix from a change that only hides the symptom.

Skill for Claude CodeCodex

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.

agentmods
npx agentmods add skills/jayrha/agentskills/debug-detective
Any agent
npx skills add JayRHa/AgentSkills --skill debug-detective
Clone the repo
git clone --depth 1 https://github.com/JayRHa/AgentSkills

Made for: Claude Code, Codex.

Per session 153 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,814 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00153 $0.01814
Opus 5 $0.00077 $0.00907
Sonnet 5 $0.00031 $0.00363
Haiku 4.5 $0.00015 $0.00181

Measured yesterday against content hash d6d3e806d94d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

debug-detective 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 yesterday.

The scan reads SKILL.md. This mod also ships 1 executable file (scripts/git_bisect_helper.sh), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

debug-detective/SKILL.md · 95 lines

How it starts

The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Debug Detective

Overview

Debugging fails when it becomes random poking — changing lines, re-running, hoping. This skill enforces a scientific, evidence-first process: you never change code to "see what happens" until you have a falsifiable hypothesis backed by an observation. The goal is to find the root cause, not to make the symptom disappear.

Keywords: debug, root cause, bug, regression, crash, stack trace, flaky test, intermittent, heisenbug, deadlock, memory leak, race condition, bisect, repro, reproduce, isolate, hypothesis, works on my machine.

Core principle: A bug you cannot reproduce is a bug you cannot confirm fixed. Spend your effort earning a reliable reproduction first; everything else gets faster after that.

Workflow

Follow these phases in order. Do not skip ahead — a fix found by skipping is usually a coincidence, not a cure.

1. Capture & Reproduce

  • Collect the exact symptom: error message, stack trace, wrong vs. expected output, exit code, screenshot, logs.
  • Record the environment: OS, runtime/language version, dependency versions, config, env vars, data, concurrency, time/locale.
  • Build a minimal reliable reproduction — the smallest command/input that triggers the bug. Note how often it fails (100%? 1 in 20?).
  • If it is intermittent, quantify the rate and try to raise it (loop it, add load, shrink timeouts) before proceeding. See references/intermittent-bugs.md.
  • Gate: Do not continue until you can trigger the bug on demand (or have a measured failure rate). Write the repro down.

2. Isolate

  • Shrink the surface area. Remove inputs, disable features, stub dependencies until the bug still reproduces with the least machinery. (Delta-debugging mindset.)
  • Establish the boundary: which layer/module/function is the last place behavior is correct, and the first place it is wrong?
  • Add observation points (logging, breakpoints, assertions, print of invariants) — observe, don't fix.

3. Hypothesize

  • State a single, falsifiable hypothesis: "The crash happens because users is None when the cache misses." A good hypothesis predicts an observation you haven't made yet.
  • Rank candidate hypotheses by likelihood × cheapness-to-test. Test the cheapest discriminating one first.
  • For each, write the prediction before you run the experiment. See references/hypothesis-log.md for the format and use templates/investigation-log.md to track them.

Read the full file on GitHub · 95 lines

Files

What ships with it

7 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. yesterday First seen · 95 lines · 153 tokens per session scan A d6d3e806d94d

Subscribe to this mod's changes

debug-detective is a skill published in the GitHub repository JayRHa/AgentSkills (4 stars, last pushed 1mo ago), licensed MIT. It adds 153 tokens to every session and 1,814 once invoked, about $0.0008 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.

Related

Other skills, from other repositories

reverse-engineering-tools

Guide for reverse engineering protected games and anti-cheat components across user mode, kernel mode, and hypervisor-aware environments. Use this skill when analyzing drivers, IOCTL protocols, callback registration, injected-code artifacts, integrity checks, protected binaries, or debugging security-sensitive game…

gmh5225/awesome-game-security · 60 tokens

unity-agent-workflows

Use for AI-assisted Unity work that needs live repo discovery, project-derived routing, runtime-owner proof, runtime-visible output hard stops, runtime numeric proof for repeated visible-output failures, state-step guards, multi-agent scope ownership, modular C#/asmdef safety, UI/scene/visual asset gates, data-first…

hashgraph-online/awesome-codex-plugins · 146 tokens

interpreting-mod-author-instructions

Use when deciding how to correctly download/install a Bethesda mod per the author's instructions — triggers "how do I install", "FOMOD choices", "which file to download", "author说明", "install instructions", "which variant", "按作者说明安装", "这个mod怎么装". NOT for deciding whether to include the mod (use evaluating-bgs-mods)…

hashgraph-online/awesome-codex-plugins · 133 tokens

using-bgs-archive

Use when the user wants to inspect, list, extract, unpack, or repack Bethesda BA2/BSA archives; determine archive format/version/compression; or build archive assets for an MO2 overlay. Triggers - "unpack BA2", "extract BSA", "pack archive", "inspect archive", "bgs-archive".

hashgraph-online/awesome-codex-plugins · 75 tokens

orchestrate

Pipeline orchestration: dispatch the highest-priority ready tasks/work units to agents, manage capacity, and coordinate the Todo to Done flow. Invoked as /agiflow:orchestrate. Uses listtasks, listactivetasksbyorg, listmembers, updatetask, getworkunitprogress.

hashgraph-online/awesome-codex-plugins · 64 tokens

design-director

Используй только внутри активного Codex Project Autopilot-проекта, когда уже выбрана проектная фаза и нужен design direction по плану автопилота.

hashgraph-online/awesome-codex-plugins · 41 tokens