smart-explore

Provides tree-based structural analysis and targeted extraction of code symbols and files.

11|2|Updated Mar 31, 2026
One-click install
npx skills add https://github.com/ajsai47/holyclaude --skill smart-explore-ajsai47
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: smart-explore
Source: https://github.com/ajsai47/holyclaude/tree/main/skills/memory/smart-explore
Command: npx skills add https://github.com/ajsai47/holyclaude --skill smart-explore-ajsai47

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates the excessive token waste and slow, multi-step discovery process of exploring large codebases with standard tools like Read, Grep, and Glob, which require reading full files or running separate searches to understand code structure.

Core Features & Use Cases

  • AST-Powered Structural Search: Find functions, classes, and symbols across entire directories in a single call, with ranked results and folded file structure previews.
  • Targeted File Outlining: Get the full structural skeleton of any code file (all functions, methods, imports) without loading implementation code, using only 1-2k tokens per file.
  • Precision Symbol Unfolding: Load only the specific function, method, or class implementation you need, avoiding the cost of reading entire large files.
  • Use Case: For example, when debugging a cross-cutting shutdown flow in a 50k-line backend codebase, use this Skill to find all related shutdown symbols in one search, then unfold only the specific implementation you need, saving thousands of tokens compared to reading full files.

Quick Start

Use the smart-explore skill to find all functions related to payment processing in the ./src directory and show the implementation of the processPayment method.

Frequently Asked Questions about smart-explore

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

FAQPage Schema
How do I reduce token usage when searching and reading large source files?

You can reduce token usage during code exploration by using AST parsing to extract file structure outlines and unfold specific symbols, avoiding full file reads and cutting context consumption by 4-8x.

What is the best way to find specific functions or classes across a large codebase?

The best way to find specific functions or classes is using structural AST search to locate symbols across entire directories in a single call, providing ranked results and folded file structure previews.

How do I navigate a 50k-line codebase to debug a cross-cutting feature without exceeding context limits?

You can navigate large codebases by searching for related symbols in one query, then using precision symbol unfolding to load only the specific implementations needed, saving thousands of tokens.

Does tree-sitter AST parsing work for generating code documentation from large files?

Yes, tree-sitter AST parsing works for generating accurate code documentation by extracting the full structural skeleton of any code file, including all functions and imports, using only 1-2k tokens per file.

Why does reading full files with standard tools waste context when exploring code?

Reading full files with standard tools wastes context because it loads unnecessary implementation details when you only need the structure, whereas AST-based outlining and targeted unfolding reduce context usage by 4-8x.