What problem does it solve? Language tutors and AI practice sessions produce inconsistent feedback, which breaks learner mental models and makes session files impossible to parse for error mining and spaced repetition. This Skill enforces one canonical feedback template across every Fluent practice session. ## Core Features & Use Cases - Standardized Feedback Template: Renders every graded answer with celebration of correct parts, corrections with category and explanation, the full correct version, and a score out of 10. - Severity Classification: Tags every error as critical (🔴), moderate (🟡), or minor (🟢) to drive spaced-repetition priority. - Error Categorization: Labels mistakes with fixed categories (grammar, formal_informal, vocabulary, spelling, prepositions, articles, missing) that feed mistakes-db.json. - Use Case: After a learner submits a Dutch writing exercise in a fluent-writing session, grade the answer with the template, tag each error's severity and category, score it, and hand the score to the SM-2 calculator. ## Quick Start Grade my Dutch sentence using the standard feedback template with severity tags, error categories, and a score out of 10.
