Run a two-stage pipeline for root-cause investigation and requirements crystallization.
npx skills add https://github.com/xdkp/oh-my-claudecode --skill deep-dive-xdkp
Please help me install this Agent Skill. Skill: deep-dive Source: https://github.com/xdkp/oh-my-claudecode/tree/main/skills/deep-dive Command: npx skills add https://github.com/xdkp/oh-my-claudecode --skill deep-dive-xdkp
Deep-dive helps you investigate why something happened and immediately turn the evidence into clear, actionable requirements instead of losing context between separate trace and interview steps.
Use the deep-dive skill when you need to investigate the root cause first, then get a precise plan for what to do next by running: /deep-dive "Why does the production pipeline fail intermittently on the transformation step"
High-intent search queries and answers about installing and using this skill.
Root cause tracing and requirements crystallization are handled as a two-stage pipeline that runs multi-lane causal tracing and then injects the findings into a targeted interview to produce an actionable spec.
Investigating ambiguous failures requires running a 3-lane trace across implementation, config orchestration, and measurement assumptions, then gating progression until ambiguity is low enough for spec generation.
Multi-lane causal tracing works by splitting evidence-based investigation into implementation, config orchestration, and measurement lanes, then seeding codebase context and an interview question queue from unresolved unknowns.
Trace findings enrich problem framing and interview initialization with safeguards against treating data as instructions, ensuring evidence context carries forward safely during requirements crystallization.
Resuming root cause analysis relies on resume-ready artifacts that save trace outputs and final specs to persistent paths, allowing robust continuation of the investigation pipeline.
A two-stage trace and interview pipeline should not be used when a problem is straightforward enough to skip evidence-based investigation or when immediate spec generation is required without ambiguity reduction.
dileep-kishoreIteratively refine codebase context retrieval across up to three search loops.
JaCha00Coordinate Planner, Architect, and Critic agents for consensus-driven task planning.
fro-botCoordinate multi-agent swarms with task lists and dependency-based pipelines.
ref-docsReproduce software bugs, analyze causes, and generate fix plans.
