lazy-explorer

name	lazy-explorer
description	Deep-dive codebase exploration producing a TL;DR summary and detailed architectural analysis. Use when asked to explore a codebase, understand a project structure, analyze architecture, explain how a codebase works, map out a repository, or answer "what does this project do". Covers technology stack, core workflows, data flow, and key components with file references.

Lazy Explorer

Explore the codebase to understand architecture, workflows, and key functionality.

Exploration Process

1. Discover structure

   • Read package manifests (package.json, pyproject.toml, Cargo.toml, etc.)
   • Identify languages, frameworks, monorepo configs
   • Map directory structure

2. Parallel deep dives — Investigate:

   • Core modules and entry points
   • Services/APIs and handlers
   • Data layer (models, schemas, migrations)
   • UI/Frontend (components, state, routing)
   • Configuration and environment
   • Infrastructure (build, CI/CD, Docker)
   • Tests

3. Synthesize findings into cohesive analysis

Output Format

TL;DR (3–5 sentences)

• What the project does
• Key technologies
• Target users/use cases
• Current state (production-ready, WIP, etc.)

Detailed Analysis

• Project Overview — Purpose and goals
• Technology Stack — Languages, frameworks, database, infrastructure
• Architecture — Entry points, service boundaries, design patterns
• Directory Structure — Annotated tree of important directories
• Core Workflows — Step-by-step breakdowns with file references
• Key Components — Important modules with path/to/file references
• Data Flow — Input sources, processing, output/storage
• Configuration — Env vars, config files, feature flags
• External Dependencies — APIs, services, notable libraries
• Areas of Interest — Complex or notable parts worth investigation

Usage

• No arguments: explore entire codebase
• Path argument (e.g. src/services/): focus on specific area
• Question (e.g. "How does authentication work?"): question-driven exploration

Clarifying Questions

Use ask_questions tool when:

• Scope is unclear (whole codebase vs specific area)
• Multiple areas could be relevant
• User's underlying goal isn't clear