smart-grep

Search codebases with token-efficient ripgrep JSON output.

Updated Nov 23, 2025
One-click install
npx skills add https://github.com/ak-eyther/model-project --skill smart-grep
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: smart-grep
Source: https://github.com/ak-eyther/model-project/tree/main/.claude/plugins/shared-superpowers/skills/smart-grep
Command: npx skills add https://github.com/ak-eyther/model-project --skill smart-grep

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Smart Grep helps you search large codebases while keeping token usage low, so you can find the right code fast without flooding your context window with irrelevant matches.

Core Features & Use Cases

  • Token-efficient search: Uses ripgrep with JSON output to stream only the most relevant matches.
  • Controlled result volume: Truncates long lines and stops when the token budget is reached.
  • Common use cases: Locating authentication logic in Python, finding React state usage in TypeScript, or scanning a repository for a specific symbol, config key, or implementation detail.

Quick Start

Use smart-grep to search the codebase for your target pattern and return the most relevant matches within a strict token budget.

Frequently Asked Questions about smart-grep

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

FAQPage Schema
How do I search a large codebase without flooding my context window?

Search a large codebase without flooding your context window by using token-efficient ripgrep output to stream only the most relevant matches. This approach truncates long lines and stops when the token budget is reached, preserving context window space.

How do I find specific functions or configuration values in a Python or TypeScript repository?

Find specific functions or configuration values in a Python or TypeScript repository by applying repository-wide lookups for symbols, configuration keys, and implementation details. The search targets source files directly to locate the relevant code fast.

Can I control the volume of search results returned when scanning source code?

Control the volume of search results returned when scanning source code through line truncation and token budgeting. The search stops automatically when the token budget is reached, ensuring only the most relevant matches are returned.

What is the best way to locate authentication logic or React state usage in a codebase?

Locate authentication logic or React state usage in a codebase by running targeted repository-wide lookups for specific symbols and implementation details. This method uses JSON match streaming to return only the most relevant code snippets.

Do I need ripgrep to search source files efficiently?

You need ripgrep to search source files efficiently with JSON match streaming. The search mechanism relies on ripgrep output to truncate long lines and apply token budgeting, which preserves context window space during repository-wide lookups.