ontledingstrainer

Inspecteer de Ontleedlab-repo op labelset, annotatieformaat, chunk-conventies en feedbackstructuur vóórdat nieuwe oefeningen of generatorlogica worden toegevoegd. Gebruik bij elke taak die nieuwe zinnen, validatie of feedbackregels wil maken of aanpassen.

Maak didactisch bruikbare, eenduidig ontleedbare Nederlandse oefenzinnen voor Ontleedlab met een constraint-based aanpak (blueprints, validatie, uitleg en moeilijkheidsopbouw). Gebruik na repo-inspectie wanneer nieuwe oefenitems, sentenceBlueprints, difficultyModel of validatieregels nodig zijn.

Use when working on Dutch zinsontleding pedagogy: feedback quality, scaffolding, sentence difficulty calibration, misconception handling (PV/WG, OW/LV, LV/MV), and educational content tone for students age 12-15.

Use for auditing UI accessibility (WCAG-oriented), keyboard navigation, semantics, ARIA usage, contrast, focus management, and inclusive interaction design.

Use when creating or reviewing Dutch sentence datasets, role labels, difficulty progression, and metadata quality in `data/sentences-level-*.json`.

Use for React+TypeScript frontend implementation, component architecture, performance tuning, and maintainable UI code in this Vite/Tailwind project.

Use when analyzing learner interaction logs, usage trends, mistake patterns, and session outcomes to improve pedagogy and product decisions.

Use for writing and improving technical documentation, README sections, guides, migration notes, and developer-facing explanations with high clarity and structure.

Use when adding or updating tests (Vitest), improving test coverage, validating regression risk, and designing focused test cases for TypeScript logic and React UI flows.

Use when adding brand personality, delightful microcopy, playful interactions, or low-risk gamification while preserving usability, accessibility, and performance.