kb-retriever

Locate and answer questions from local knowledge bases using layered index navigation.

Updated Feb 17, 2026

Search Tags:#pdf #indexing #excel #grep #question-answering #local-knowledge-base #kb-retrieval

Ready to run

npx skills add https://github.com/liuning0820/skills --skill kb-retriever-liuning0820

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: kb-retriever
Source: https://github.com/liuning0820/skills/tree/main/skills/kb-retriever
Command: npx skills add https://github.com/liuning0820/skills --skill kb-retriever-liuning0820

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pdf2image, and includes scripts (resource) and references (resource) components.

What problem does it solve?

本技能帮助从本地知识库中有效检索信息并给出回答，结合分层索引导航减少不相关信息干扰。

Core Features & Use Cases

分层索引导航：从根目录到子领域逐步定位相关内容，并给出可检索的结果。
多类型文件支持：支持 Markdown、文本、PDF、Excel 等格式，必要时先用 references 提供的方法处理。
渐进式检索场景：对每个候选文档局部读取，避免一次性加载大文件，适用于问答、资料检索与资料查找场景。
使用场景示例：在大型本地知识库中快速定位“销售策略”相关文档并汇总要点。

Quick Start

直接向本 Skill 提出问题，Skill 将按分层索引逐步检索并返回最相关的答案。

Frequently Asked Questions about kb-retriever

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I search a local knowledge base of PDF and Excel files without loading entire documents?▼

Layered index navigation enables local knowledge base retrieval by progressively narrowing from root directories to sub-areas. It reads candidate documents locally via predefined references rules, minimizing full-file loading while preserving context for accurate question-answering.

How do I retrieve answers from local Markdown and PDF documents using grep?▼

You can locate answers across Markdown, text, PDF, and Excel files using progressive grep-based retrieval. The system navigates a layered index, partially reading candidate documents to enforce minimal file loading and maintain context for accurate results.

Can I use this local knowledge base retrieval approach for large directories of mixed file types?▼

Yes, local knowledge base retrieval supports Markdown, text, PDF, and Excel files across large directories. It uses layered index navigation to progressively locate relevant sub-areas, preventing full-file loading to manage scale and reduce irrelevant information.

Do I need pdf2image to extract and answer questions from local PDF files?▼

Yes, the pdf2image dependency is required to process PDF files within the local knowledge base. The system uses predefined references rules to handle PDF, Excel, and Markdown formats before executing progressive grep-based retrieval for question-answering.

What is the best way to locate specific information in a large local knowledge base?▼

The best way to locate specific information in a local knowledge base is layered index navigation. It progressively narrows from root directories to sub-domains, applying grep-based retrieval to pinpoint relevant content locally while minimizing full-file loading.

Why does progressive retrieval minimize full-file loading when searching local documents?▼

Progressive retrieval minimizes full-file loading to reduce irrelevant information and preserve context. By navigating layered indexes and locally reading only relevant document sections, it achieves accurate question-answering without overwhelming system resources.