What problem does it solve? Scattered research notes and one-off chat summaries get lost between sessions. This Skill builds long-lived, structured knowledge bases on any topic—papers, products, people, codebases—stored as markdown on disk and indexed in a SQLite FTS5 database so any future session can recall them. ## Core Features & Use Cases - Wiki scaffolding: Initialize a wiki with a standard layout (raw sources, compiled pages, logs, sources table) in one of nine flavors such as research, paper, product, codebase, or incident. - FTS5 indexing: Every page written is upserted into a SQLite full-text index via the wiki-cli, keeping search in sync with the markdown files. - Citation discipline: Enforces a sources.md registry so every claim in compiled pages cites a tracked source, with lint prompts to catch orphans, broken links, and uncited claims. - Use Case: Ask to start a wiki on a research topic, ingest papers and links over weeks, then query the accumulated knowledge base from any later session without re-reading the sources. ## Quick Start Start a research wiki on transformer interpretability and add my first source page to it.
