What problem does it solve? Sokrates tracks only TODOs by default, so questions like "how much code handles secrets" or "where are the feature flags" go unanswered. This Skill defines the concern groups (features of interest) in _sokrates/config.json and verifies each regex against the real codebase so the concern views are meaningful. ## Core Features & Use Cases - Candidate measurement: The propose_concerns.py script scans the main files against a catalog of ~30 generic concerns (technical debt, security, robustness, data, integration, observability, configuration, lifecycle) plus the repository's most imported external libraries, reporting file counts, LOC percentages, matched tokens, and stratified samples per candidate. - Concern authoring guidance: Enforces Sokrates regex semantics (contentPattern must match an entire line, pathPattern the entire path, exception vetoes, metaConcerns with extract/replace name operations) and selection rules such as dropping candidates that touch more than 60% of files. - Use Case: A team migrating off a legacy framework defines legacy: javax.servlet and new: spring concerns, verifies them with the preview script, and tracks both curves across analyses to make the migration measurable over time. ## Quick Start Ask the AI to propose and configure features of interest for the repository by running the propose_concerns.py script against _sokrates/config.json and updating the concernGroups.
