seamless-rag

Use Seamless-RAG (the MariaDB-native vector + RAG toolkit) to embed table rows, run vector search with VEC_DISTANCE_COSINE/HNSW, and return query results in TOON tabular format that costs 10–55% fewer LLM tokens than JSON. Trigger this skill whenever the user is working with the seamless-rag CLI or Python package, mentions MariaDB VECTOR columns, asks to embed/index/search rows in MariaDB, wants to convert SQL output into a token-efficient context for an LLM, asks about TOON v3 format, or sets up auto-embedding/watch-mode against a MariaDB table — even if they don't say 'seamless-rag' by name. Also use when the user pairs this with the text-to-sql skill for hybrid (precise filter + semantic) queries.

Translate a natural-language data question into a safe SELECT query against MariaDB and return the result in TOON tabular format (token-efficient context for LLMs/agents). Trigger this skill whenever the user has structured tables in MariaDB and asks a question that maps to precise SQL — revenue/sales reports, top-N rankings, GROUP BY aggregations, exact-value filters, JOINs across tables, counts, averages, percentile-style queries — rather than fuzzy/semantic search. Also trigger when the user explicitly says 'write SQL for…', 'query the DB for…', or hands you a schema and asks a question. Pairs with the seamless-rag skill: this one handles the precise-SQL leg, seamless-rag handles the semantic / vector-search leg, and they combine via `seamless-rag ask … --where` for hybrid queries.

Work with the Seamless-RAG toolkit — MariaDB vector search, TOON encoding, auto-embedding, and RAG queries. Use this skill when the user works in the seamless-rag project, asks about MariaDB vector operations, TOON format encoding, embedding providers, or RAG pipeline tasks. Also trigger when the user wants to query databases via vector search, convert data to TOON format, or manage MariaDB embedding workflows.

Convert natural language questions into SQL, execute against MariaDB, and return results in TOON format. Use when the user asks a data question that needs a precise SQL query — revenue reports, aggregations, filtered lookups, JOINs — rather than semantic/vector search. Works with any MariaDB database the user has access to.