wiki

To maintain persistent project documentation across multiple AI sessions, use a markdown knowledge base that stores pages in a local directory with an auto-maintained index. This ensures your session learning and architecture notes compound over time without being lost.

You can find specific architecture notes or debugging records by querying a markdown knowledge base using keyword and tag matching. This mechanism retrieves titled pages with snippets for citation-based synthesis, operating without vector embeddings.

The best way to detect broken cross-references and stale content in a markdown wiki is to run a linting health check. This process identifies orphan pages, oversized pages, structural contradictions, and stale content to ensure ongoing refactoring quality.

Yes, you can create cross-references between decision logs in a local knowledge base using the [[page-name]] link syntax. This connects your categorized markdown pages, allowing you to navigate related architecture notes and session learning records.

To quickly add a single page to a project knowledge base for a debugging record, use the fast page creation feature. This allows you to instantly capture a decision, pattern, or debugging note as a titled and tagged markdown page.