smart-explore

Parse code with tree-sitter ASTs to map codebase structure and symbols.

Updated Apr 15, 2026
One-click install
npx skills add https://github.com/sheehe/coase --skill smart-explore-sheehe
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: smart-explore
Source: https://github.com/sheehe/coase/tree/main/resources/plugins/coase-builtin/skills/smart-explore
Command: npx skills add https://github.com/sheehe/coase --skill smart-explore-sheehe

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Token-optimized structural code search using tree-sitter AST parsing. It helps you understand code structure and navigate large codebases without loading every file.

Core Features & Use Cases

  • Structural exploration with AST parsing to map code organization before drilling into details.
  • Index-first offline map creation with on-demand fetch to speed up understanding.
  • Use Case: quickly locate a function or class across a repo and then unfold its implementation as needed.

Quick Start

Run smart_search to map the codebase, then use smart_outline or smart_unfold to inspect a symbol or file.

Frequently Asked Questions about smart-explore

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

FAQPage Schema
How do I navigate a large codebase without loading every file?

You can navigate a large codebase by generating a structural map using tree-sitter AST parsing. This technique indexes files and symbols first, allowing on-demand unfolding of specific implementations without loading full sources.

What is AST-powered code search and how does it work?

AST-powered code search is a technique that parses code into an abstract syntax tree to map structural relationships. It works by applying tree-sitter parsing across repositories to discover symbols and organization before fetching full file contents.

How do I locate a specific function or class across a repository?

To locate a specific function or class across a repository, run a structural search to map the codebase organization. Once mapped, you can unfold the specific symbol's implementation details on demand.

Can I use tree-sitter AST parsing for scalable code exploration?

Yes, tree-sitter AST parsing supports scalable code exploration by creating deterministic structural maps. It safely queries large repositories to discover files and relationships without requiring full source loading.

Does structural code search require loading full source files?

No, structural code search does not require loading full source files. It uses tree-sitter ASTs to build an index-first offline map, allowing you to safely query and unfold specific symbols only when needed.