Queries markdown knowledge bases with the qmd CLI and retrieves high-scoring chunks.
npx skills add https://github.com/paulirish/dotfiles --skill qmd-expert
Please help me install this Agent Skill. Skill: qmd-expert Source: https://github.com/paulirish/dotfiles/tree/main/agents/skills/qmd-expert Command: npx skills add https://github.com/paulirish/dotfiles --skill qmd-expert
Searching a large markdown knowledge base often returns fragmented results, and naive queries lose the pointers to the most relevant content. This Skill provides a structured workflow for running multi-query searches with the qmd CLI, extracting high-scoring chunk IDs, and fetching exactly the context needed without overwhelming the context window.
Use the qmd skill to search the knowledge base for information about Interaction to Next Paint and retrieve the most relevant chunks.
High-intent search queries and answers about installing and using this skill.
Run qmd query for hybrid auto-expanding search, qmd search for strict BM25 keyword matching, or qmd vsearch for vector similarity on natural language questions. Combining multiple query types improves recall across the knowledge base.
Use the --all flag with the --files option to bypass the default 20-result limit. Add --min-score 0.7 to enforce a quality threshold, and redirect output to a file for later parsing.
qmd query performs hybrid search with automatic expansion, qmd search runs strict BM25 text matching without LLM involvement, and qmd vsearch uses vector similarity for natural language questions where exact vocabulary is unknown.
Chunk hash IDs point to the specific high-scoring fragment within a document, while URLs only reference the whole file. Fetching by hash with qmd get retrieves exactly the relevant snippet, keeping context usage lean.
Use the -l flag with qmd get to control how many lines of surrounding context are returned, for example qmd get "#888e5a" -l 20. You can also rank chunks by score and fetch only the top N results.
tan-yong-shengIndex and query Markdown content locally with keyword and semantic search.
gagan114662Search the web, evaluate sources, and synthesize findings with cited URLs.
AvinavaOrganize Salesforce Workforce Engagement guidance into a structured knowledge map.
markus41Manage vector embeddings for semantic search across Pinecone, Weaviate, Chroma, and pgvector.
