Combine BM25 keyword and vector semantic search across indexed reference collections.
npx skills add https://github.com/ramonclaudio/skills --skill search-ramonclaudio
Please help me install this Agent Skill. Skill: search Source: https://github.com/ramonclaudio/skills/tree/main/plugins/qmd/skills/search Command: npx skills add https://github.com/ramonclaudio/skills --skill search-ramonclaudio
This skill helps developers quickly locate relevant information across large indexed reference collections, reducing time spent on manual searching.
Use qmd_query to search for a topic, then refine results with qmd_search or qmd_vsearch. Retrieve a document with qmd_get using a path or docid.
High-intent search queries and answers about installing and using this skill.
Hybrid search combines BM25 keyword search with vector semantic scoring to balance precision and recall across indexed reference collections. It targets exact keyword hits and semantic understanding for codebase exploration and design reasoning.
Semantic search across multiple repos works by selecting and ranking results using vector scoring. It surfaces context-rich results from indexed repositories and embedded documentation to support architectural queries.
The best way to find exact keyword hits is using BM25 keyword search within a guided workflow. It targets precise retrieval across indexed reference collections, ranking exact matches to reduce manual searching time.
Yes, you can retrieve specific documents from indexed reference collections using a path or docid. The structured workflow uses qmd_get to retrieve exact files after locating them through query, search, or vsearch.
Yes, hybrid search works effectively for exploring large codebases and embedded documentation. It combines BM25 keyword search with vector semantics to locate relevant information across indexed repositories, reducing manual search time.
qmd_query initiates a topic search, qmd_search refines via BM25 keyword scoring, and qmd_vsearch applies vector semantic ranking. The guided workflow selects and ranks results across these commands before retrieval via qmd_get.
steipeteIndex local files and search them with BM25, vector, and hybrid queries.
WILLOSCARExtract plain text from PDF documents and generate index files.
ASRagabSearch past coding sessions across multiple agents with lexical, semantic, or hybrid modes.
ccfCombine vector similarity with keyword matching for hybrid retrieval.
