What problem does it solve? Combining many Sokrates repository analyses into one landscape report requires configuration Sokrates cannot infer: how repositories group into sub-landscapes, which tags matter, who the people behind email addresses are, and which accounts are bots. Misconfigured regexes silently match nothing, duplicate repository names collide, and contributors land in an "Undefined Team" — all discovered only after a slow report run. ## Core Features & Use Cases - Landscape bootstrap and tuning: Creates and edits _sokrates_landscape/config.json plus config-tags.json, config-teams.json, and config-people.json, covering thresholds, virtual sub-landscapes by repository-name patterns, tags, teams, bots, and contributor identity merging. - Pre-run checker: The check_landscape.py script re-implements Sokrates' discovery and matching rules to preview which repositories are found or excluded, which tags and virtual landscapes each gets, and how the contributor pipeline (ignore → transform → people → threshold → bots → teams) resolves — before running updateLandscape. - Use Case: An engineering org with 80 analyzed repositories wants them grouped by team with merged contributor identities. The skill sets up virtual landscapes and people configs, then the checker flags two regexes that do not compile and five active contributors matching no team. ## Quick Start Ask the AI to set up a Sokrates landscape over your analysis root and run the checker to preview repository discovery, tags, and team assignments before generating the report.
