grepai-search-tips_skill

Overview

This skill provides concise tips and best practices for writing effective semantic search queries with GrepAI. It helps you shift from literal text matching to meaning-driven queries so you find the right code faster. Use these guidelines to improve precision, speed up discovery, and troubleshoot poor results.

How this skill works

The skill explains how GrepAI interprets natural-language queries and what kinds of phrasing produce high-quality semantic matches. It shows patterns for describing intent, choosing context, and iteratively refining queries. It also explains score ranges so you can interpret match confidence and decide when to refine a query or switch to text search.

When to use it

• When search results are too broad or irrelevant
• When you know desired behavior but not exact names
• When learning how to write effective semantic queries
• When troubleshooting low-quality or missing results
• When exploring domain-specific code across a repo

Best practices

• Describe intent, not implementation (e.g., 'fetch user by ID' vs 'getUserById')
• Use 3–7 descriptive English words focused on behavior and context
• Be specific about inputs/outputs or the scenario (e.g., 'validate email format')
• Iteratively refine: start broad, add context, then be precise
• Use domain vocabulary and synonyms to surface related implementations

Example use cases

• Find authentication logic: 'user authentication middleware' rather than 'auth'
• Locate error handling: 'catch and log API call errors'
• Discover data operations: 'save user to database' or 'transform data before storage'
• Track configuration: 'load configuration from environment' or 'database connection settings'
• Search security code: 'password hashing and verification' or 'input sanitization'