grepai-search-basics

Perform semantic code search with GrepAI using natural language queries.

Updated May 12, 2026
One-click install
npx skills add https://github.com/eduardtomasek/codex-dev-stack --skill grepai-search-basics-eduardtomasek
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: grepai-search-basics
Source: https://github.com/eduardtomasek/codex-dev-stack/tree/main/skills/.agents/grepai-search-basics
Command: npx skills add https://github.com/eduardtomasek/codex-dev-stack --skill grepai-search-basics-eduardtomasek

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Traditional text search tools like grep only match exact strings, so finding code by intent or behavior requires knowing exact identifiers. This Skill teaches how to use GrepAI's semantic search to locate code by meaning, interpret relevance scores, and write effective queries. ## Core Features & Use Cases - Semantic Search Fundamentals: Explains how meaning-based search differs from exact text matching and when to use each. - Query Crafting Guidance: Provides patterns for effective natural language queries, including finding entry points, error handling, data access, and business logic. - Result Interpretation: Covers score thresholds, result formatting, limiting output with --limit, and verifying index status. - Use Case: You join an unfamiliar codebase and need to find where user authentication happens. Run a semantic search for "user authentication flow" to get ranked code chunks across auth middleware, JWT validation, and login handlers. ## Quick Start Ask the assistant to search the codebase semantically for "user authentication flow" using GrepAI and explain the top results.

Frequently Asked Questions about grepai-search-basics

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

FAQPage Schema
How do I search code by meaning instead of exact text?▼

Use GrepAI's semantic search with the command grepai search followed by a natural language query describing the code's behavior. It returns ranked code chunks by meaning similarity rather than exact string matches.

What is the difference between semantic search and grep?▼

Grep matches exact strings and returns unranked results, while semantic search finds code by meaning similarity. Searching for "authenticate user" semantically also finds login, signin, and auth code that grep would miss.

What makes a good semantic code search query?▼

Good queries describe intent or behavior in 3-7 English words, such as "validate user credentials" or "connect to the database". Avoid single vague words or exact function names, which are better handled by grep.

Why does GrepAI search return no results?▼

No results usually means the index is missing or empty. Run grepai status to verify the index, start grepai watch to build it, and confirm an embedding provider like Ollama is running.

How do I interpret GrepAI search scores?▼

Scores range from 0.0 to 1.0, with higher values indicating stronger relevance. Scores above 0.90 are excellent matches, 0.80-0.89 are good, and anything below 0.60 is a weak match.