llm-wiki

To build a cross-referenced Markdown wiki, you ingest raw sources into a structured directory layout featuring schema, raw files, and wiki pages. An agent then summarizes, cross-references, and files materials to prevent information loss and maintain consistency.

Organizing ingested materials into an interlinked knowledge base requires a three-layer structure: raw sources, wiki pages, and schema. This layout ensures robust provenance, enables navigable cross-page synthesis, and maintains consistency across sources.

Frontmatter metadata improves wiki cross-referencing and linting by providing structured, agent-readable context for each page. This supports automated curation, highlights contradictions, and maintains consistency across the knowledge base.

Yes, you can use an automated agent to audit and query a Markdown wiki for contradictions. The agent-driven curation process lints the knowledge base, highlights contradictions across pages, and answers questions using cross-referenced insights.

Yes, you need a specific directory structure to maintain Markdown wiki provenance. The required layout features a wiki directory containing schema, raw sources, and cross-linked pages to ensure robust provenance and maintainability.

Research wikis lose information when updated from multiple sources due to messy cross-references and lack of structured curation. Applying an agent-driven ingestion process with a three-layer Markdown structure prevents this loss and keeps the wiki current.