skills.rest
GYF0311
GYF0311COMMUNITY
Explore all skills by this author →
GitHub

corpus-gbrain-query

Query the Lorekit GBrain index for multi-hop relationships and map candidates to canonical corpus pages.

5|1|Updated Apr 15, 2026
Search Tags:#knowledge graph#gbrain#fact grounding#multi-hop retrieval#corpus query#lorekit wiki#candidate mapping
One-click install
npx skills add https://github.com/GYF0311/lorekit --skill corpus-gbrain-query
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: corpus-gbrain-query
Source: https://github.com/GYF0311/lorekit/tree/main/skills/corpus-gbrain-query
Command: npx skills add https://github.com/GYF0311/lorekit --skill corpus-gbrain-query

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you discover graph-style, multi-hop relationships in a Lorekit central corpus when keyword-based retrieval returns weak or incomplete candidates.

Core Features & Use Cases

  • Multi-hop GBrain candidate recall: Expands beyond direct matches by using GBrain to retrieve relationship candidates for a given query.
  • Canonical mapping to knowledge base pages: Translates GBrain candidates back to the corresponding authoritative Lorekit 知识库/ pages.
  • Fact-source alignment and safety boundaries: Ensures the skill treats Lorekit 知识库/ as the only source of truth and avoids any mutations of 知识库/ or 原料/.

Quick Start

Ask your agent to run the global GBrain query against the central Lorekit corpus and then answer using the mapped canonical 知识库/ pages as the factual basis.

Frequently Asked Questions about corpus-gbrain-query

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

FAQPage Schema
How does multi-hop retrieval in a knowledge graph work for finding indirect relationships?▼

Multi-hop graph retrieval finds indirect relationships by querying the Lorekit GBrain index to discover adjacent concept candidates. It then maps these candidates back to canonical knowledge base pages for accurate fact grounding.

What is the best way to query multi-hop relationships when keyword search returns incomplete results?▼

The best way is running a global GBrain query against your central corpus to retrieve multi-hop relationship candidates. It maps these candidates back to canonical knowledge base pages, overcoming keyword search limitations.

How do I query the Lorekit GBrain index from the default corpus directory?▼

You query the Lorekit GBrain index by having your agent read the global corpus configuration and run the configured query from the default corpus directory. The system then maps returned candidates to canonical knowledge base pages.

Does the GBrain query skill modify or write to the knowledge base and source material directories?▼

No, the GBrain query skill is strictly read-only. It ensures no writes occur to the knowledge base, raw material, or output directories, treating the canonical knowledge base as the sole source of truth.

Why does my GBrain query return freshness warnings during multi-hop retrieval?▼

Freshness warnings are returned by the GBrain query when the central index data may be outdated relative to the canonical corpus pages. This alerts you to verify fact-source alignment before relying on the mapped candidates.

Related Skills

Explore More Education & Research Skills→
GYF0311GYF0311
5

corpus-capture

Capture ad-hoc notes into a Lorekit corpus workbench inbox as timestamped Markdown fragments.

Community
Basic
latestaiagentslatestaiagents
5

graphrag-patterns

Combine knowledge graphs with hybrid retrieval for relationship-aware querying.

Official
Advanced
ArogyaReddyArogyaReddy

hybrid-search-implementation

Combine vector similarity and keyword matching with RRF, Linear, Cross-encoder, or Cascade fusion.

Community
Advanced
JeffallanJeffallan
10.9k

rag-architect

Design and implement retrieval-augmented generation pipelines with vector stores and hybrid search.

Community
Advanced