Skip to content
Get packs

Parallel Debugging

Debug tricky bugs by testing competing hypotheses in parallel and weighing the evidence

Source Download skill
Repository
wshobson/agents
Stars
42,095
Licence
MIT
Files
2 files

What it does

When installed, this skill tells the AI agent to approach a hard bug the way a careful detective would: list several possible causes instead of jumping to the first guess, then check each one against evidence from the code. It organises the work around six failure categories (such as logic errors, bad data, race conditions, integration problems, resource leaks and environment issues), requires file and line citations for any claim, and rates each hypothesis as confirmed, plausible, falsified or inconclusive before naming a root cause. It also gives a checklist to confirm a fix really works before declaring the bug closed.

Great for

  • Developers stuck on a bug with several plausible causes
  • Engineers who want a structured root cause analysis with written evidence
  • Teams splitting an investigation across several agents or people

Skip it if

  • Simple bugs with an obvious cause, where the framework adds overhead without value
  • Anyone expecting the skill to run tests or read logs automatically; it is a method and checklist, not a debugging tool

Try asking

  • “Our checkout service intermittently double-charges customers. Use parallel debugging to generate and test hypotheses.”
  • “Run a root cause analysis on why the nightly export job's memory usage keeps climbing.”
  • “Help me organise a parallel investigation into the login failures across three modules, citing evidence for each hypothesis.”
Step-by-step guide: how to use Parallel DebuggingInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the MIT licence and an attribution note.

    parallel-debugging.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip parallel-debugging.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip parallel-debugging.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent the reproduction steps, error messages and relevant code paths up front so hypotheses can be tested against real evidence.
  • Ask for the evidence table and confidence levels in the output so you can review the reasoning, not just the conclusion.
  • Treat the 'parallel' part as optional: the method works with one agent working through hypotheses one at a time.

What's inside

  • references/hypothesis-testing.md
  • SKILL.md

Redistributed unmodified from wshobson/agents under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.