What problem does it solve?
Traditional RAG systems re-discover and re-process raw knowledge from scratch for every query, leading to redundant work, inconsistent cross-references, and missed synthesis of accumulated insights. This Skill eliminates that overhead by maintaining a persistent, pre-compiled knowledge base that grows smarter with every source you ingest.
Core Features & Use Cases
- Persistent Compounding Knowledge: Ingest sources (web articles, PDFs, transcripts, pasted text) once, then query synthesized, cross-referenced answers indefinitely without re-reading raw material.
- Automated Consistency Maintenance: Lint for orphan pages, broken wikilinks, stale content, and tag taxonomy violations to keep your knowledge base clean and navigable.
- Seamless Tool Integration: Works out of the box with Obsidian, VS Code, and any markdown editor, with optional headless sync for server-based agents and desktop vaults.
- Use Case: A researcher studying AI safety can ingest 50 arxiv papers and blog posts once, then instantly query for comparisons between alignment techniques, with all cross-references and contradictions already flagged.
Quick Start
Use the llm-wiki skill to ingest the provided article on large language model scaling laws and update the relevant wiki pages with new findings and cross-references.