Generate JLPT N5 text-level grammar (Mondai 3) questions in JSON format. Use this skill when the user asks to create, generate, or produce N5 text-level grammar questions, text integration exercises, or "metin tamamlama" questions. Output is a JSON object with a "textFlow" type, containing "title", "textSegments", and "blanks" array. Trigger on requests like "N5 Mondai 3 sorusu üret", "create N5 text integration questions", "generate text-level grammar", or "JLPT N5 metin bütünleme sorusu oluştur".
Generate N5-level text-level grammar (Mondai 3) questions as JSON.
Determine the theme for the short text (e.g., daily routine, email, announcement, short story) from user request or context.
Read references/json-schema.md for textFlow field specifications.
Read references/n5-grammar-points.md when selecting connectors (shikashi, soshite, etc.) or verb forms.
Generate the question:
textSegments array. The number of segments will be number of blanks + 1.
blankNumber starting from 1 for each text (1, 2, 3...).position (0, 1, 2...).id as n5_grammer_text_integration_XXX where XXX is a zero-padded sequential number. Continue from the last existing ID in the file.Append the generated question to the questions array in backend/grammar/data/n5_text_integration.json. Read the file first, then add the new question to the end.
Validate: After saving, run the jlpt-n5-grammer-text-integration-question-tester skill to validate the JSON file. If any questions fail, fix them and re-run the tester until all pass.
Upload prompt: After all questions pass validation, ask the user whether to upload to Firebase. If yes, run: python3 backend/grammar/scripts/upload_grammar_questions.py n5 --type text-integration
Output a summary of the text created and the grammar points tested.