Stats
Actions
Tags
Help us improve
Share bugs, ideas, or general feedback.
npx claudepluginhub punt-labs/claude-plugins --plugin quarryHow this command is triggered — by the user, by Claude, or both
Slash command
/quarry-dev:find <query>The summary Claude sees in its command listing — used to decide when to auto-load this command
<!-- markdownlint-disable MD041 --> ## Input Arguments: $ARGUMENTS Use the full arguments text as the search query. ## Task Call `mcp__plugin_quarry_quarry__find` with `query` set to the arguments. If the result says "No results", say so briefly. **Determine the query type:** - **Question** — the query asks for an answer (starts with who/what/how/why/when/where, ends with `?`, or contains phrases like "is the", "how to", "how do"): provide a concise direct answer based on the search results, citing the source document and page number for each claim. Do not dump the raw results. - *...
/searchSearches Glean enterprise knowledge base with a query, showing top relevant results with titles, sources, recency indicators, snippets, and relevance notes.
/wiki-queryQueries the compiled knowledge base wiki to answer a question with cited facts from relevant topics, points to sources if needed, and offers to file novel syntheses.
/assist-searchSearches lm-assist knowledge base across entries, milestones, and architecture via API. Displays ID, title, relevance score, excerpt per result, total count. Supports --limit flag.
/knowledge-searchSearches OpenArche knowledge stores using vector embeddings, prioritizing repository-local over global, with expansion of linked neighbors and scored/tag results.
/searchSearches Obsidian vault using qmd semantic search (BM25 + vectors + reranking), showing top 10 results with relevance scores, titles, paths, and snippets.
Share bugs, ideas, or general feedback.
Arguments: $ARGUMENTS
Use the full arguments text as the search query.
Call mcp__plugin_quarry_quarry__find with query set to the arguments.
If the result says "No results", say so briefly.
Determine the query type:
Question — the query asks for an answer (starts with who/what/how/why/when/where, ends with ?, or contains phrases like "is the", "how to", "how do"): provide a concise direct answer based on the search results, citing the source document and page number for each claim. Do not dump the raw results.
Keywords — the query is exploratory (keywords, phrases, topic names): emit the tool output exactly as returned — character for character, including the leading ▶ unicode character. Do not reformat, add commentary, wrap in code fences, convert to markdown tables, or add boxes around the output.