Search local markdown notes and docs using BM25, vector, and hybrid retrieval modes.
npx skills add https://github.com/surfingalien/FinSurfing --skill qmd-surfingalien
Please help me install this Agent Skill. Skill: qmd Source: https://github.com/surfingalien/FinSurfing/tree/main/.claude/skills/qmd Command: npx skills add https://github.com/surfingalien/FinSurfing --skill qmd-surfingalien
High-intent search queries and answers about installing and using this skill.
Use qmd search followed by your query to run a BM25 keyword search across indexed markdown collections. Add -c to restrict to one collection, -n to limit results, or --json for machine-readable output.
qmd search uses BM25 keyword matching and is typically instant. qmd vsearch adds semantic vector similarity but can take about a minute on cold start, while qmd query adds LLM reranking on top and is the slowest option.
Install qmd with bun install -g from its GitHub repo, then run qmd collection add with your notes path and a **/*.md mask. Run qmd embed once to enable vector and hybrid search, and verify with qmd status.
New or modified files are missing because the index has not been refreshed. Run qmd status to check index health, then qmd update to re-index changed files, or remove and re-add the collection for a full rebuild.
qmd query and vsearch are slow on large vaults because they load a local LLM for expansion and reranking. Use qmd search instead, limit results with -n, or scope the search to a single collection with -c.
sandalsoftIndex and search local Markdown collections with BM25, semantic, or hybrid search.
skyllwtRetrieve and synthesize wiki pages into cited answers for research questions.
DavidROliverBASearch a knowledge base for notes related to a given topic.
moistalgiaSearch local markdown notes and documents with hybrid keyword, vector, and reranked retrieval.
