What problem does it solve? Newcomers and AI assistants waste weeks reverse-engineering a codebase's core nouns — what a Thread, Session, or Rollout actually means — and silently misread systems where one concept has two names or one name has two meanings. This scanner writes that domain language down from primary sources so the vocabulary is documented, owned, and checked for drift. ## Core Features & Use Cases - Code-anchored glossary: Defines the 8–15 concepts that carry the system from their defining types, doc comments, identity, lifecycle, and owning component — not from README aspirations. - Bounded context and capability maps: Shows which components own which vocabulary, where concepts are shared or translated, and which nouns each product capability family rides on. - Language drift detection: Finds synonym pairs, homonyms, misleading names, undefined jargon, and renames caught mid-flight, each with the visible canonical direction. - Use Case: After running a Sokrates analysis on a repository, ask for the domain model; the scanner reads the protocol/API types and produces a validated findings file plus an interactive HTML explorer a new hire can read in one sitting. ## Quick Start Ask the AI to run the domain language scan on this repository's _sokrates analysis and produce a glossary of its core concepts with any naming drift.
