deep-dive
Interview the user relentlessly about a plan, design, strategy, or decision until reaching shared understanding. Walks down each branch of the decision tree, resolving dependencies one-by-one. Use when user wants to stress-test a plan, get grilled on a design, deep-dive into a phase, or clarify a complex decision — works for both technical and non-technical topics.
Best use case
deep-dive is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Interview the user relentlessly about a plan, design, strategy, or decision until reaching shared understanding. Walks down each branch of the decision tree, resolving dependencies one-by-one. Use when user wants to stress-test a plan, get grilled on a design, deep-dive into a phase, or clarify a complex decision — works for both technical and non-technical topics.
Teams using deep-dive should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/deep-dive/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How deep-dive Compares
| Feature / Agent | deep-dive | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Interview the user relentlessly about a plan, design, strategy, or decision until reaching shared understanding. Walks down each branch of the decision tree, resolving dependencies one-by-one. Use when user wants to stress-test a plan, get grilled on a design, deep-dive into a phase, or clarify a complex decision — works for both technical and non-technical topics.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
Interview me about this specific phase/topic until we reach a shared understanding. Walk down each branch of the decision tree, resolving dependencies between decisions one-by-one. If a plan file exists in the project, read it first to understand the full context. If a question can be answered by exploring the codebase, explore the codebase instead of asking. ## Interview checklist For each item, make sure these angles are covered before moving on: - **Implementation details** — What exactly, where in the code, how it integrates - **Edge cases & error handling** — Missing data, failures, fallbacks - **UX decisions** (if applicable) — Visual design, interactions, mobile, loading/error states - **Dependencies** — Does decision A constrain decision B? What must be decided first? - **Scope boundaries** — What is explicitly out of scope or deferred? - **Risks** — What is the biggest risk? What could go wrong? ## Rules - **Always use AskUserQuestion** with 2-4 concrete options ranked by recommendation. - **One question at a time.** Ask the most critical unresolved question, wait, then proceed. - **Be relentless.** Do not accept vague answers. If the user says "something like X", push for specifics. - **Track progress.** Periodically remind the user of remaining open items. - **Respect the user's expertise.** Your job is to extract clarity, not to lecture. - **Match the user's language.** Conduct the interview in whatever language the user communicates in. - **Stop when clear.** Once the core goal, constraints, major decisions, and biggest risks are all resolved — stop. ## When done Produce a short summary of all decisions made, open questions, biggest risk, and recommended next step. Ask the user what to do next. **Do NOT jump into implementation.**
Related Skills
web-design-guidelines
Review UI code for Web Interface Guidelines compliance. Use when asked to "review my UI", "check accessibility", "audit design", "review UX", or "check my site against best practices".
update-deps
Audit all outdated dependencies with detailed research on changelogs, breaking changes, bug fixes, and deprecations. Creates a temporary plan without updating anything. Use when you want to review what changed in your dependencies before upgrading.
react-useeffect
Audit React components for unnecessary useEffect hooks. Use when reviewing code that has useEffect, useState patterns, or when asked to optimize React components. Identifies anti-patterns like derived state, chained effects, event-driven effects and suggests better alternatives (computed values, useMemo, key prop, event handlers, useSuspenseQuery).
deep-research
Multi-source deep research using firecrawl and exa MCPs. Searches the web, synthesizes findings, and delivers cited reports with source attribution. Use when the user wants thorough research on any topic with evidence and citations.
deep-research
Run autonomous research tasks that plan, search, read, and synthesize information into comprehensive reports.
deep-research
Run autonomous research tasks that plan, search, read, and synthesize information into comprehensive reports.
deeptools
NGS analysis toolkit. BAM to bigWig conversion, QC (correlation, PCA, fingerprints), heatmaps/profiles (TSS, peaks), for ChIP-seq, RNA-seq, ATAC-seq visualization.
deepchem
Molecular machine learning toolkit. Property prediction (ADMET, toxicity), GNNs (GCN, MPNN), MoleculeNet benchmarks, pretrained models, featurization, for drug discovery ML.
deepspeed
Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
deep-research-notebooklm
Deep research skill powered by NotebookLM MCP. Conducts structured multi-source research (market analysis, competitive intel, trend analysis, prospect research) using Google NotebookLM as the research engine, then delivers formatted briefs and optional studio artifacts (slides, audio podcasts, videos, infographics, reports, mind maps).
deepinit
Deep codebase initialization with hierarchical AGENTS.md documentation
deep-interview
Socratic deep interview with mathematical ambiguity gating before autonomous execution