What problem does it solve? In modular monoliths and bounded-context architectures, vocabulary from one module silently leaks into another through string literals, event names, and API calls, eroding linguistic boundaries. This Skill performs a strict read-only audit of those boundaries so violations are found with concrete evidence before they harden into coupling. ## Core Features & Use Cases - Cross-module boundary audit: Parses language.md files to build vocabulary inventories and relationship graphs, then detects string leaks, wrongly-directed API calls, foreign event consumption, and unpublished internal vocabulary usage. - Single-module PR check: Diffs a pull request with the --pr flag and flags new concepts that break the module's linguistic space or leak downstream language upstream. - Evidence-based reporting: Produces a linguistic-boundary-report.md with ASCII architecture diagrams, violation tables citing exact code locations, and before/after fix proposals confirmed with the user. - Use Case: Before merging a PR that adds a new event handler in the Orders module referencing Payments terminology, run the verifier in PR mode to confirm the new concept fits the Orders linguistic space. ## Quick Start Ask the assistant to run the linguistic boundary verifier on all modules, or on a single module with the --pr flag to check new concepts in a pull request.
