What happens if a PDF is an image and not text-based?

The Reference Indexer suggests using OCR tools like 'ocrmypdf' to convert scanned images into searchable text before indexing.

How does this skill improve Claude's performance?

It provides Claude with a Retrieval-Augmented Generation (RAG) corpus, allowing it to search through your specific reference documents for facts instead of relying solely on general training data.

How do I add a new document to the index?

You can use the '/ref-add {path}' command, which automatically validates the file, extracts the text, and updates the RAG corpus.

What file formats are supported by the Reference Indexer?

The skill supports PDF, Word (.docx), Markdown (.md), HTML, and plain text (.txt) files.

Can I organize my references into categories?

Yes, it supports a structured hierarchy including Legal, Technical, Business, and Internal categories, all tracked via a central index file.

Reference Indexer

Name: Reference Indexer
Author: arbgjr

byarbgjr

•

データサイエンスとML

Indexes and manages external reference documents to power Retrieval-Augmented Generation (RAG) for domain-specific AI workflows.

概要

The Reference Indexer skill provides a robust framework for ingesting, processing, and querying external documentation within the Claude Code environment. By automating text extraction from various formats—including PDFs, Word documents, and Markdown—it builds a structured local RAG corpus. This allows Claude to leverage specialized domain knowledge, such as legal regulations, technical specifications, and internal business rules, ensuring that AI-generated responses are grounded in your specific project references and compliance standards.

主な機能

Categorized reference management for legal, technical, and business docs
Automated indexing workflow with validation and chunking
Automated document summarization and metadata generation
Semantic search capabilities with relevance scoring
Multi-format text extraction supporting PDF, DOCX, MD, HTML, and TXT
1 GitHub stars

ユースケース

Managing internal business manuals for instant retrieval of project rules
Processing legal and regulatory documents for compliance-driven development
Indexing technical RFCs and standards to guide software architecture

概要

主な機能

Categorized reference management for legal, technical, and business docs
Automated indexing workflow with validation and chunking
Automated document summarization and metadata generation
Semantic search capabilities with relevance scoring
Multi-format text extraction supporting PDF, DOCX, MD, HTML, and TXT
1 GitHub stars

ユースケース

Managing internal business manuals for instant retrieval of project rules
Processing legal and regulatory documents for compliance-driven development
Indexing technical RFCs and standards to guide software architecture