AGENT SKILL

probe

Use this skill when the user wants to search code semantically, extract code blocks with AST context, find patterns using tree-sitter, or analyze code structure.

License 许可
MIT
所属仓库
javimosch/supercli
最近更新
2026-11-02
Stars
★ 57
Forks
5

01 前言 Front Matter

name
probe
description
Use this skill when the user wants to search code semantically, extract code blocks with AST context, find patterns using tree-sitter, or analyze code structure.

03 SKILL.md 正文

Probe Plugin

AI-friendly semantic code search engine for large codebases. Combines ripgrep speed with tree-sitter AST parsing.

Commands

Search

  • probe search query — Semantic code search with Elasticsearch-style queries

Extract

  • probe extract code — Extract code blocks with full AST context

Symbols

  • probe symbols list — List all symbols in files with line numbers

Query

  • probe query ast — AST-based structural pattern matching

Usage Examples

Basic search:

probe search "authentication" ./src
probe search "error AND handling" ./
probe search "login OR auth" ./src

Extract code:

probe extract src/main.rs:42
probe extract src/main.rs#authenticate
probe extract src/main.rs:10-50

List symbols:

probe symbols src/main.rs
probe symbols src/main.rs --format json

AST pattern matching:

probe query "async fn $NAME($$$)" --language rust
probe query "function $NAME($$$)" --language javascript

Installation

npm install -g @probelabs/probe

Or via curl:

curl -fsSL https://raw.githubusercontent.com/probelabs/probe/main/install.sh | bash

Key Features

  • AST-aware code understanding (not just text)
  • Elasticsearch-style query syntax (AND, OR, NOT, phrases)
  • Complete code blocks (not fragmented chunks)
  • Zero indexing, instant results
  • BM25/TF-IDF hybrid ranking
  • Multi-language support (Rust, Python, JS, Go, C++, etc.)
  • Fully local - code never leaves your machine
  • MCP integration for AI coding assistants

04 同仓库更多 Skills