Search markdown collections using BM25, vector retrieval, and LLM reranking.
npx skills add https://github.com/escotilha/claude-public --skill qmd-escotilha
Please help me install this Agent Skill. Skill: qmd Source: https://github.com/escotilha/claude-public/tree/main/skills/qmd Command: npx skills add https://github.com/escotilha/claude-public --skill qmd-escotilha
You can’t quickly locate the exact skill, pattern, decision, or prior note you need across a growing markdown knowledge base, especially when your phrasing doesn’t match the stored wording.
Use the qmd skill to search your knowledge base for the skill or rule related to your question, by asking: " /qmd pattern de tratamento de erros ".
High-intent search queries and answers about installing and using this skill.
Hybrid semantic search combines BM25 keyword search, vector retrieval, and LLM reranking to find relevant markdown documents even when your query phrasing differs from the stored text, returning accurate matches across collections.
Cross-collection discovery searches skills, agents, rules, memory, and project docs to locate documented patterns and decisions, reducing lookup time by applying hybrid retrieval and LLM reranking across all markdown collections.
Vector semantic search can miss exact entity, tool, or error-string terms, but a low-confidence fallback keyword search recovers these cases by automatically executing a BM25 query when vector results are empty or low confidence.
Yes, QMD hybrid query execution supports optional collection scoping, allowing you to restrict searches to specific markdown collections such as rules, memory, or project docs to retrieve more targeted results.
Yes, the search supports both user-facing result formatting and agent-spawned structured responses, enabling automated workflows to consume retrieval outputs directly from markdown collections.
zeronx798Index and retrieve local markdown files using hybrid BM25, vector, and LLM reranking search.
whiteboardevRetrieve memory summaries by ID, project, or recency via the BrAIny API.
gabrielkoerichIndex local Markdown notes with BM25 keyword search, vector semantic search, and LLM reranking.
Yangel-hideSearch local knowledge stores with hybrid BM25, vector, and LLM reranking.
