What is P-adic ultrametric distance?

It is a non-Archimedean distance metric used in this skill for hierarchical clustering, offering a more structured way to view relationships between skills than standard Euclidean distance.

Can I search for skills using natural language?

Yes, the skill includes an embed_query function that allows you to find relevant skills based on arbitrary descriptive text.

How does DuckDB fit into this skill?

DuckDB is used as the storage and compute engine, providing the HNSW (Hierarchical Navigable Small World) indexing layer for fast approximate nearest neighbor searches.

Does this skill require specific hardware?

Yes, it is specifically optimized for Apple Silicon (M-series chips) using the MLX framework for efficient local embedding generation.

What is the primary purpose of Skill Embedding VSS?

It is designed to compare and search agent skills by converting their content into high-dimensional vectors and measuring their relational distance using advanced embedding models.

Skill Embedding VSS

Name: Skill Embedding VSS
Author: plurigrid

byplurigrid

•

데이터 과학 및 ML

Performs vector similarity search and hierarchical clustering for agent skills using P-adic ultrametrics and MLX-powered embeddings.

This skill enables Claude to analyze, search, and organize agent skills through high-dimensional vector embeddings and advanced mathematical structures. By leveraging Snowflake Arctic models via MLX on Apple Silicon and DuckDB's HNSW indexing, it provides efficient similarity searches, novelty detection, and relational mapping between disparate skills. It is particularly useful for managing large skill repositories, identifying functional overlaps, and discovering unique architectural patterns through P-adic ultrametric distances and GF(3) triad conservation logic.

주요 기능

01Snowflake Arctic 1024-bit embeddings optimized for Apple Silicon via MLX

02Trifurcated random walks for exploring skill keyspaces with GF(3) conservation

03P-adic ultrametric distance for non-Archimedean hierarchical clustering

04High-speed approximate nearest neighbor search using DuckDB HNSW indexing

052 GitHub stars

06Novelty detection to identify unique skill contributions within a repository

사용 사례

01Analyzing large skill directories to find functionally redundant or similar capabilities

02Identifying 'novel' skills that provide unique functionality not yet represented in the ecosystem

03Searching skill repositories using natural language queries to find the best tool for a specific task

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi skill-embedding-vss

For use in Claude.ai and ChatGPT

Download Skill