跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
Use-case shortlist

Best Agent Skills for RAG and Knowledge Workflows

Find skills for document ingestion, retrieval, embeddings, source-grounded answers, and agent workflows that need reliable private knowledge.

RAG skills should reduce answer risk

The point of a RAG skill is not only retrieval. It should help an agent ingest clean material, retrieve relevant context, and keep answers grounded in sources.

  • Look for clear ingestion and retrieval behavior.
  • Prefer skills that make source provenance visible.
  • Validate chunking and retrieval quality against real documents.

Pair RAG with upstream preparation

Many failures happen before retrieval. Web scraping, PDF parsing, OCR, and document cleanup skills often matter as much as the RAG layer.

  • Use document processing for messy PDFs and tables.
  • Use web extraction for public-source knowledge bases.
  • Use data analysis skills when retrieved context becomes structured metrics.

操作步骤

  1. Choose a source corpus

    Pick a small but representative set of documents before scaling ingestion.

  2. Measure retrieval quality

    Ask known-answer questions and inspect whether the right source material appears.

  3. Add citations or proof points

    Require the agent to show the evidence behind each answer before shipping.

Do I need a separate scraping skill for RAG?

If the source material lives on the web, a scraping skill can be the upstream ingestion layer. For private files, document processing may be more important.

What is the first RAG skill to install?

Start with a skill that can ingest and retrieve from your actual source format, then add companions for parsing, crawling, or evaluation.