Verified:GitHub

corpus-query

Search a Lorekit corpus and return cited, synthesized answers.

5|1|Updated Apr 15, 2026

Search Tags:#knowledge retrieval #semantic search #citations #hybrid search #wiki synthesis #vector query #information recall

Ready to run

npx skills add https://github.com/GYF0311/lorekit --skill corpus-query

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: corpus-query
Source: https://github.com/GYF0311/lorekit/tree/main/skills/corpus-query
Command: npx skills add https://github.com/GYF0311/lorekit --skill corpus-query

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you find and synthesize accurate answers from a persistent Lorekit knowledge base without re-reading scattered sources every time.

Core Features & Use Cases

Cross-project global querying: Routes a question to a canonical central corpus and returns a synthesized response grounded in canonical pages.
Hybrid retrieval workflow: Combines exact term search, semantic/hybrid vector recall, and optional GBrain-assisted candidate recall while keeping canonical pages authoritative.
Safe read-only operation with clear boundaries: Ensures query-time reads from the corpus only, and supports confidence notes plus citations to real corpus pages.

Quick Start

Ask your AI agent: “Search the default Lorekit corpus for the answer to ‘What is lorekit’s incremental compilation workflow and how should evidence be promoted into the raw layer?’ and cite the exact canonical pages.”

Frequently Asked Questions about corpus-query

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I perform semantic search and synthesize answers from a knowledge corpus?▼

Semantic search synthesizes answers from a knowledge corpus by routing questions to a central Lorekit database, applying hybrid vector recall, and returning cited results grounded in canonical pages. It combines exact-term lookup with optional GBrain-assisted candidate recall.

What is hybrid retrieval and how does it work for cross-project knowledge queries?▼

Hybrid retrieval for cross-project knowledge queries works by combining exact term search, semantic vector recall, and optional GBrain-assisted candidate recall. It routes questions to a persistent central Lorekit corpus and synthesizes responses grounded in authoritative canonical documentation.

How do I get cited answers with confidence notes from a documentation knowledge base?▼

To get cited answers with confidence notes from a documentation knowledge base, you query the central corpus using hybrid retrieval. The system returns synthesized responses that include citations pointing to real corpus pages and provide confidence notes indicating the reliability of the recalled information.

Can I use this knowledge retrieval approach for read-only documentation workflows?▼

Yes, you can use this knowledge retrieval approach for read-only documentation workflows. It ensures query-time reads from the Lorekit corpus only, keeping the knowledge base safe and unmodified while applying per-corpus rules and utilizing configured corpus-local binaries.

Does this semantic search tool require reading per-corpus rules before querying?▼

Yes, this semantic search tool requires reading per-corpus rules and canonical pages before querying. This ensures the synthesized responses accurately respect the specific boundaries and configurations of the central Lorekit knowledge base you are searching.

What are the limitations of using a central knowledge corpus for information recall?▼

Limitations of using a central knowledge corpus for information recall include being restricted to read-only operations and depending on the configured corpus-local binaries. You cannot write to the corpus, and recall accuracy relies on the existing canonical pages within the persistent knowledge base.