“RAG”

Use this skill when you need evidence-bounded query variants, chunking, filters, recall/precision proxies, ranking, freshness, and retrieval evidence…

Perform exhaustive, multi-pass web research with citations.

Turn resources shared in Telegram chat (attachments, URLs, pasted text) into searchable vector memory, and answer later questions from it.

Build or modify the Legal Tech Debt local-first legal discovery, instrumentation, and retrieval workflow for Sandbox 002.

High-performance local File RAG suite (Basic Edition).

Multi-agent presuppositional research under TwinGlass geometry constraints.

Perform fast, lean web research with concise citations.

Audits whether website content has clear, bounded, retrieval-friendly structural units for AI retrieval and answer generation.

Retrieval-augmented literature search over local paper library.

Fail-fast gate and domain integrity repair in the RAG graph.

Implement Retrieval-Augmented Generation pipelines — ingestion, chunking, embedding, retrieval, and generation.

Expert in building Retrieval-Augmented Generation systems.

Select compact, step-specific OpenUSD and Omniverse retrieval context for planning, generation, validation, or repair.

Perform exhaustive web research prioritising official primary sources.

Practical anti-hallucination workflow for RAG: deterministic vs retrieval, query_type/workflow_phase routing, top-k/scoring, grounding/freshness/waste…

检索多模态数据处理平台已完成工作流的最新处理结果。适用于网页爬取、文档解析、OCR、Markdown/JSON 转换等已写入数据集的结果问答，并返回可回溯的数据集版本与来源引用。

Perform fast, lean web research prioritising official primary sources.

Implement local Retrieval-Augmented Generation using Room and Vector embeddings.

Augment model context with external factual retrieval through query reformulation, hybrid dense-sparse indexing, re-ranking, and citation grounding.

Knowledge base assistant with document retrieval and citation

deciding whether to stuff documents into a long window.

first end-to-end chain after chunks exist.

Answer a factual question from institutional docs, memory, and ingested history.

Design and evaluate retrieval-augmented generation systems end-to-end.