Retrieve and synthesize lorekit corpus content using exact and semantic routing.
npx skills add https://github.com/GYF0311/lorekit --skill wiki-query-gyf0311
Please help me install this Agent Skill. Skill: wiki-query Source: https://github.com/GYF0311/lorekit/tree/main/skills/wiki-query Command: npx skills add https://github.com/GYF0311/lorekit --skill wiki-query-gyf0311
wiki-query helps you find existing knowledge in your lorekit corpus and generate a synthesized answer with truthful, traceable citations, instead of guessing or forcing web research.
知识库/摘要/<slug>.md for core claims.lorekit vector status, lorekit search, and lorekit vector query --hybrid to balance precision and semantic coverage.corpus/log.md, and proposes fileback to store valuable syntheses.Ask your AI: 查一下 “XXX”，并用 corpus 里的现有资料回答（如果 corpus 没有就诚实说没有），同时给出每条结论的页面来源引用并在末尾询问是否要把答案存回 corpus。
High-intent search queries and answers about installing and using this skill.
To retrieve corpus answers with citations, use wiki-query to synthesize existing content using exact and semantic routing across text and vector modes. It enforces traceable citations grounded in real wiki links and summary pages, avoiding web research guesses.
Multi-hop reasoning across wiki links is handled by wiki-query's three-layer retrieval routing. It selects the right lookup strategy based on query intent, using wikilinks for limited exploration instead of forcing external web research to connect disparate concepts.
Combine BM25 and vector search by running lorekit vector status and using lorekit vector query --hybrid. This balances precision and semantic coverage, routing between exact text lookups and fuzzy concept queries based on corpus mode.
wiki-query requires running vector status to utilize hybrid vector queries effectively. If no vector index is available, it falls back to exact and fuzzy text routing using ripgrep search, though semantic coverage will be limited.
Answers lack source provenance if the synthesis is not grounded in real wiki links. wiki-query enforces source-guaranteed synthesis by requiring evidence from summary pages, outputting citations per claim, and logging queries to ensure traceability.
You can store synthesized answers back into the knowledge base through wiki-query's fileback UX loop. After generating a grounded synthesis with citations, it proposes saving the valuable conclusions back to the corpus for future retrieval.
GYF0311Report Lorekit central corpus health with doctor, vector index, and ingest queue status.
servitolaBuild and maintain a persistent Markdown wiki knowledge base with interlinked pages.
GYF0311Query the Lorekit GBrain index for multi-hop relationships and map candidates to canonical corpus pages.
pvillegaGather evidence-based findings from multi-source web searches with citations.