What infrastructure is required for this skill?

This skill is designed for the CodeCompass platform, which requires a vector database (Weaviate) and an embedding service (Ollama) to index and retrieve code context.

What makes semantic search different from traditional grep?

While grep finds exact character matches, semantic search uses vector embeddings to find code based on meaning and intent, allowing you to find concepts like 'payment processing' even if those exact words aren't in the code.

When should I choose grep over semantic search?

Use grep for exact matches like specific function names, class definitions, or imports where you know the precise syntax you are looking for.

How can I improve my semantic search results?

Use high-context queries following the '[Action] for [Entity] with [Constraints]' formula to provide the vectorizer with enough information to find specific implementations.

Semantic Code Search

Name: Semantic Code Search
Author: PearlThoughts

byPearlThoughts

開発者ツール

Enables intent-based code discovery by providing a decision framework for using vector-based semantic search versus traditional grep.

概要

The Semantic Search skill provides a sophisticated methodology for navigating complex codebases using natural language and conceptual queries. By leveraging the CodeCompass platform's integration with Weaviate and Ollama embeddings, it helps developers move beyond keyword matching to find business logic, architectural patterns, and cross-module implementations. This skill guides the user through formulating high-context queries, interpreting relevance scores, and choosing the most efficient discovery method—whether semantic vectors for exploratory research or grep for exact structural matches.

主な機能

Iterative refinement strategies for complex architectural exploration
Decision tree for selecting between semantic search and grep/glob
Relevance score interpretation to filter noise from large codebases
Context-aware query formulation framework for higher accuracy
0 GitHub stars
Integration guide for Weaviate vector databases and Ollama embeddings

ユースケース

Onboarding onto massive codebases by searching for intent-based concepts
Discovering business logic implementations across unfamiliar, polyglot repositories
Locating architectural patterns like error handling or auth flows without knowing specific class names

概要

主な機能

Iterative refinement strategies for complex architectural exploration
Decision tree for selecting between semantic search and grep/glob
Relevance score interpretation to filter noise from large codebases
Context-aware query formulation framework for higher accuracy
0 GitHub stars
Integration guide for Weaviate vector databases and Ollama embeddings

ユースケース

Onboarding onto massive codebases by searching for intent-based concepts
Discovering business logic implementations across unfamiliar, polyglot repositories
Locating architectural patterns like error handling or auth flows without knowing specific class names