Search local knowledge bases with hybrid keyword, vector, and reranked retrieval.
npx skills add https://github.com/JKhyro/HERMES-AGENT --skill qmd-jkhyro
Please help me install this Agent Skill. Skill: qmd Source: https://github.com/JKhyro/HERMES-AGENT/tree/main/optional-skills/research/qmd Command: npx skills add https://github.com/JKhyro/HERMES-AGENT --skill qmd-jkhyro
qmd removes the friction of searching large local knowledge bases by combining keyword search, semantic retrieval, and reranking so you can find the right note, transcript, or document quickly.
Ask qmd to search your local notes for the topic, then review the best result and open the source document if needed.
High-intent search queries and answers about installing and using this skill.
To search local markdown notes by concept, use hybrid retrieval that combines BM25 keyword search with vector embeddings and LLM reranking to find relevant documents when exact wording is unknown.
Hybrid retrieval combining BM25 search with vector embeddings is a mechanism that matches exact keywords while performing semantic discovery to find conceptually relevant notes in local knowledge archives.
Yes, you can use semantic retrieval for project documentation through MCP integration, connecting the local search index directly to Hermes Agent for seamless agentic retrieval workflows.
Yes, local indexing is required to run reranked searches on personal notes, because the system must generate vector embeddings and build BM25 indexes on-device before executing fast retrieval.
The best way to retrieve specific terms from large local text files is using a hybrid search approach that runs BM25 keyword matching alongside vector search and applies LLM reranking for precise results.
cdeistopenedIndex and search local Markdown collections with BM25 and vector embeddings.
eddieluengSearch local markdown and text collections with hybrid retrieval and reranking.
yusufcmgImplement Retrieval-Augmented Generation systems with vector databases and semantic search.
JuanMS20Index Markdown and text files for local hybrid search with BM25, vectors, and reranking.
