deep-dive

Trace root causes across three parallel lanes and crystallize evidence-grounded requirements.

Updated Apr 17, 2026
One-click install
npx skills add https://github.com/dropsyoon/oh-my-claudecode --skill deep-dive-dropsyoon
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-dive
Source: https://github.com/dropsyoon/oh-my-claudecode/tree/main/skills/deep-dive
Command: npx skills add https://github.com/dropsyoon/oh-my-claudecode --skill deep-dive-dropsyoon

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Deep-dive turns ambiguous problems into evidence-grounded requirements by investigating root causes before starting requirements definition.

Core Features & Use Cases

  • 3-lane causal tracing: Investigates WHY something happened using parallel hypotheses (code-path, config/orchestration, and measurement/artifact mismatch).
  • 3-point trace injection: Feeds the trace synthesis into the follow-up deep-interview so requirements crystallization starts with the right context instead of re-exploring.
  • Evidence-driven requirements: Produces a clear, actionable spec grounded in traced findings rather than assumptions.
  • Use case: When a production system behavior is unclear (or a bug feels intermittent), run deep-dive to determine the likely mechanism first, then generate the most appropriate fix requirements.

Quick Start

Run deep-dive with your problem statement, for example: /deep-dive "Why does the production DAG fail intermittently on the transformation step".

Frequently Asked Questions about deep-dive

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I investigate root causes before generating software requirements?

To investigate root causes before requirements generation, you can use causal tracing to analyze parallel hypotheses across code-path, configuration, and measurement lanes. This method translates ambiguous problems into evidence-grounded specifications by confirming the underlying mechanism first.

What is the best way to trace intermittent production bugs into actionable specs?

The best way to trace intermittent production bugs into actionable specs is running a 2-stage pipeline that couples causal tracing with requirements crystallization. It feeds synthesized trace evidence directly into a follow-up interview, ensuring the final spec is grounded in confirmed findings.

How does causal tracing work for ambiguous system behavior?

Causal tracing for ambiguous system behavior works by investigating three parallel hypotheses: code-path execution, config or orchestration mismatches, and measurement or artifact discrepancies. This multi-lane approach isolates the true mechanism causing the unclear production behavior.

Can I use pipeline automation to generate specs from bug investigation findings?

Yes, you can use pipeline automation to generate specs from bug investigation findings. The pipeline automates a 3-point trace injection that feeds confirmed causal hypotheses directly into requirements engineering, producing an actionable spec output ready for execution handoff.

Do I need to confirm hypotheses manually during root cause analysis?

Yes, you need to confirm hypotheses manually during root cause analysis. The pipeline requires user-confirmed hypotheses before proceeding to requirements crystallization, ensuring the generated specs rely on validated evidence rather than automated assumptions.

What are the limitations of generating specs without root cause discovery?

Generating specs without root cause discovery risks producing requirements based on assumptions rather than evidence. When production system behavior is unclear, skipping causal tracing and trace-to-interview injection often leads to ineffective fixes that fail to address the actual mechanism.