Query a niche-specific qmd index to rank relevant wiki pages.
npx skills add https://github.com/00PZ/oopz-inc --skill qmd-search-00pz
Please help me install this Agent Skill. Skill: qmd-search Source: https://github.com/00PZ/oopz-inc/tree/main/companies/shoshin/skills/qmd-search Command: npx skills add https://github.com/00PZ/oopz-inc --skill qmd-search-00pz
In the Librarian workflow, users need fast, accurate access to the most relevant wiki pages within a specific niche to ground knowledge synthesis. This skill provides a constrained search mechanism to surface top pages from the compiled wiki, reducing noise and iteration time.
Run a constrained qmd search for this niche to surface the most relevant wiki pages for synthesis.
High-intent search queries and answers about installing and using this skill.
Hybrid search for niche wiki retrieval combines BM25 keyword matching, vector embeddings, and LLM re-ranking to identify and rank the most relevant pages for a specific query.
The best way to search a single niche index without data leakage is to enforce explicit niche binding, which scopes queries strictly to one collection and prevents results from crossing into other domains.
To find relevant wiki pages for knowledge synthesis, query the niche-specific qmd index using the provided search index and embedding model to return highly ranked results suitable for a Librarian workflow.
Yes, you can use this constrained search to supply citations and context for a Librarian workflow by returning the top ranked wiki pages from the isolated niche index for knowledge synthesis.
No, niche search does not support querying across multiple collections; it is strictly scoped to a single niche with enforced explicit binding to ensure complete isolation of search results.
jpjednorskiIndex and search Obsidian vault notes with qmd keyword and semantic queries.
maneeshanifSearch tasks by keyword in titles, descriptions, and tags.
zilliztechDecompose questions, retrieve across documents, and synthesize cited answers.
sheawinklerSearch local markdown and text collections using hybrid BM25 and vector ranking.
