agentsclimarketplace

030101 lancedb search

Skill natuleadan/skills/skills/030101-lancedb-search

Multi-domain agent skills collection for AI coding agents (Claude, Cursor, Copilot, OpenCode, and more). Covers programming, biology, cooking, and future domains. Installable via npx skills add.

Install
npx -y skills add natuleadan/skills --skill 030101-lancedb-search

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

LanceDB vector search fundamentals — distance metrics, ANN indexing, embeddings, similarity search patterns, and performance tuning.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

1.5 KB, as published. Nobody here has run it

Vector Search Core

Overview

Core concepts for vector search with LanceDB: distance metrics, ANN indexing, embedding models, and search patterns.

Quick Reference

Distance Metrics

MetricUse CaseBehavior
l2General purposeSmaller = more similar
cosineText/document similaritySmaller = more similar
dotNormalized vectorsLarger = more similar
hammingBinary vectorsSmaller = more similar

Search Modes

  • Exact search: Brute force, 100% recall, no index needed
  • ANN search: Approximate, fast, requires index, configurable recall
  • Hybrid: Vector + FTS combined with reranking

Embedding Functions

  • OpenAI: text-embedding-ada-002, text-embedding-3-small, text-embedding-3-large
  • Sentence Transformers: local models via sentence-transformers
  • Custom: create your own embedding function

References

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.