🔍 Research

Deep Dive

Research any topic within a GitHub repo and generate a polished HTML deep-dive document. Onboard to any codebase in minutes.

Saves 2–4 hours

What It Does

Targets a topic inside any GitHub repository—public or your local clone—and produces a self-contained HTML reference you can share or archive. The output explains architecture, surfaces the files that matter, traces data flows, links recent issues and PRs for living context, and calls out configuration and extension points. It’s built for PMs and engineers who need credible onboarding docs without spending a day spelunking.

Example Prompts

Deep dive into the router in remix-run/remix
Teach me about the plugin system in grafana/grafana
Generate a reference doc for authentication in our repo

How It Works

  1. Discovery — Identifies the relevant directory, package boundaries, and primary entry points for your topic.
  2. Code architecture — Maps structure, important types, and how data moves through the system.
  3. Documentation — Collects READMEs, design docs, ADRs, and high-signal inline comments.
  4. Issues & PRs — Pulls recent GitHub activity so you see what’s changing and why.
  5. Synthesis — Merges everything into one polished HTML page with consistent headings and navigation-friendly sections.

What You Get

A single, self-contained HTML file organized for scanning and deep reading, typically including:

  • Overview — Scope, audience, and why the area exists.
  • Architecture — Layers, modules, and dependencies.
  • Key Files — Files to open first when you need to change behavior.
  • How It Works — Runtime and request/flow narratives.
  • Key Concepts — Terms and patterns the codebase assumes you know.
  • Configuration — Env vars, flags, and config files.
  • Related Issues — Recent threads worth following.
  • Further Reading — Links and docs to go deeper.

Try It Now

Clone the repo and start using this skill in under 60 seconds.