Load large documents into a Python REPL for chunked analysis and aggregation.
npx skills add https://github.com/danielsimonjr/claude-skills --skill rlm-skill
Please help me install this Agent Skill. Skill: rlm-skill Source: https://github.com/danielsimonjr/claude-skills/tree/main/rlm-skill Command: npx skills add https://github.com/danielsimonjr/claude-skills --skill rlm-skill
This skill enables processing content that exceeds traditional language model context windows by loading data into a Python runtime, chunking intelligently, and synthesizing results across chunks.
Run the RLM processor on a large file, for example: python rlm_processor.py path/to/large_document.txt "What are the key findings?"
High-intent search queries and answers about installing and using this skill.
To analyze large PDF, DOCX, and HTML files that exceed context windows, use a Python runtime to load data, apply intelligent chunking, and hierarchically aggregate results. This method synthesizes coherent summaries across multi-format documents without truncation.
Yes, directory processing enables cross-file analysis and coherent summaries across many files. It supports multi-format ingestion including PDF, DOCX, HTML, JSON, CSV, XML, archives, and 30+ code or text extensions for combined hierarchical aggregation.
The best way to chunk large codebases and JSON files is using auto-chunking that detects document structure like markdown headers, code blocks, and JSON hierarchy. It selects an optimal chunking strategy and applies pre-filtering with keyword or regex to reduce LLM calls.
Yes, pdfplumber and python-docx are required dependencies for extracting text from PDF and DOCX files. Additional dependencies like beautifulsoup4 and PyPDF2 support multi-format ingestion and HTML parsing within the Python runtime environment.
Hierarchical aggregation recursively combines results from individual chunks to enable analysis of very large contexts. This allows the system to synthesize directory-wide summaries and structured reports across codebases and multi-file collections without losing context.
HokageZParse and extract text, tables, and metadata from PDF, DOCX, HTML, and Markdown files.
ActiveInferenceInstituteGenerate HTML, Markdown, and JSON pipeline analysis reports from GNN outputs.
nirajrathod19Fetch and analyze Repl deployment logs with regex filtering.
rukikotooQuery the econometrics RAG knowledge base to retrieve relevant document chunks.
