elasticsearch-onboarding

Elastic Developer Guide is a conversational onboarding skill that helps developers new to Elasticsearch move from wanting search to a working, production-ready search experience. It positions the assistant as an Elasticsearch solutions architect working alongside the developer: understanding intent, recommending an approach, walking through data mapping, and generating tested code with best practices built in. It targets Elasticsearch 9.x and follows a structured conversation playbook kept in a reference file.

The guide is deliberately interactive. It asks one question at a time and waits, listens for signals about the use case and data shape, and only generates code once the user has confirmed both the approach and the mapping. Because the mapping is the most expensive thing to change later, it always walks through it. It asks which programming language to use rather than assuming, explains decisions briefly on the assumption that the user does not yet understand Elasticsearch, and avoids deprecated APIs (warning explicitly if one is unavoidable). Two conventions are enforced by default: using the Synonyms API for synonym management rather than a custom solution, and always using a versioned index name paired with an alias (for example products_v1 with products_current), with an explanation of why.

Trigger it when someone shows intent to build search functionality or asks how to get started. The example intents span keyword and relevance basics (BM25, ES|QL) as well as modern semantic and vector approaches: building a vector database, kNN search, ELSER-powered semantic search, EIS embeddings, combining keyword and vector results with RRF, connecting an LLM, building a RAG pipeline, and setting up the Elasticsearch MCP. The result is guidance tailored to the developer's specific use case and data shape rather than a generic tutorial.