Can I manually load specific IR sub-skills?

Yes, you can manually load individual skills like 'ir-vector-search' or 'ir-ranking-reranking' using the cat command provided in the skill documentation.

What is progressive loading in this context?

Progressive loading means the skill only pulls in the specific search or retrieval logic you need as you need it, preventing the AI from becoming bogged down with unnecessary information.

How does the discover-ir skill activate?

The skill activates automatically when Claude Code detects keywords related to intermediate representation or search systems, such as 'vector search', 'ranking', or 'LLVM IR'.

Does this skill help with LLVM or compiler development?

Yes, it is designed to activate for compiler-related tasks (SSA, LLVM IR) as well as Information Retrieval (search/data science) tasks, acting as a bridge for both domains.

Discover Information Retrieval (IR) Skills

Name: Discover Information Retrieval (IR) Skills
Author: rand

byrand

•

データサイエンスとML

Automates the discovery and activation of specialized Information Retrieval and search engine optimization skills for advanced data tasks.

The discover-ir skill acts as an intelligent gateway for Claude Code, automatically detecting when you are working on search systems, query processing, or data retrieval architectures. It utilizes a progressive loading system to provide context-efficient access to deep domain expertise in vector search, ranking algorithms, and recommendation systems without overwhelming the conversation window. This skill ensures that Claude applies the most relevant best practices and implementation patterns whenever search-related development or compiler-level IR tasks are identified.

主な機能

01Quick-reference indexing for search fundamentals and ranking logic

02Automatic activation based on task context and terminology

03Integrated access to vector search and recommendation system workflows

04Seamless switching between query understanding and re-ranking skills

0557 GitHub stars

06Progressive loading architecture for maximum context efficiency

ユースケース

01Developing and optimizing vector-based search pipelines

02Building recommendation systems and ranking algorithms

03Refining search query understanding and retrieval accuracy

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rand/cc-polymath discover-ir

For use in Claude.ai and ChatGPT

Download Skill