What problem does it solve? A Sokrates landscape with hundreds of repositories is an unreadable flat list; this Skill turns it into a map by designing the virtualLandscapes section of _sokrates_landscape/config.json — named, nestable groups of repositories matched by name patterns — with measured coverage instead of guesswork. ## Core Features & Use Cases - Evidence-based grouping proposals: A script scans every repository analysis under the root and measures candidate groupings by naming conventions, folders, technology, activity, size, contributor e-mail domains, shared committers, and tags, reporting members, LOC shares, coverage, and remainder for each. - User grouping validation: Supply your own grouping (flat map or nested virtualLandscapes object) and get it measured exactly as Sokrates will evaluate it — case-sensitive full-string regex matching, overlaps, empty groups, and per-level remainders. - Paste-ready config output: Emits a ready virtualLandscapes JSON object including remainderLandscapeMetadata, then verifies the final result with the landscape checker. - Use Case: You have 200 repository analyses in one landscape and want sub-landscapes per product line. Run the proposal script, pick the naming-convention regexes that match real product families, nest large groups, and write the measured config into _sokrates_landscape/config.json. ## Quick Start Ask the AI to split your Sokrates landscape into meaningful virtual sub-landscapes by running the proposal script against your analysis root and writing the chosen grouping into the landscape config.
