jlpt-n5-grammer-sentence-building-question-creator

name: jlpt-n5-grammer-sentence-building-question-creator description: Generate JLPT N5 sentence building (starQuestion) questions in JSON format. Use this skill when the user asks to create, generate, or produce N5 sentence building questions, word-order exercises, or scrambled sentence questions. Output is a JSON object containing a "questions" array with starQuestion type items, each including sentencePrefix, sentenceSuffix, 4 scrambledWords, starPosition, and correctOrder. Trigger on requests like "N5 cümle birleştirme sorusu üret", "create N5 sentence building questions", "generate word order questions", "JLPT N5 mondai2 soru oluştur", or "scrambled sentence questions".

JLPT N5 Sentence Building Question Creator

Generate N5-level sentence building (starQuestion) questions as JSON.

References

• Output format & field rules: See references/json-schema.md
• N5 grammar points and word-order rules: See references/n5-grammar-points.md

Workflow

1. Determine the grammar target(s) — from user request or source documents in context.
2. Read references/json-schema.md for field specifications.
3. Read references/n5-grammar-points.md for word-order rules and scramble patterns.
4. Generate each question:
   • Write a natural N5-level sentence split into: sentencePrefix + 4 words + sentenceSuffix.
   • The 4 words in scrambledWords must be in a shuffled order (not the correct order).
   • Set correctOrder as the array of indices mapping scrambled positions to the correct sequence.
   • Set starPosition (0–3) indicating which slot in the correct order is the marked answer.
   • Set id as n5_grammer_sentence_building_XXX where XXX is a zero-padded sequential number. Continue from the last existing ID in the file.
5. Append generated questions to the existing questions array in backend/grammar/data/n5_sentence_building.json. Read the file first, then add new questions to the end of the questions array.
6. Validate: After saving, run the jlpt-grammer-sentence-building-question-tester skill to validate the JSON file. If any questions fail, fix them and re-run the tester until all pass.
7. 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 sentence-building
8. Output a summary of what was generated, validated, and (optionally) uploaded.

Quality rules

• All vocabulary must be within N5 scope.
• Sentences should reflect real daily-life contexts.
• scrambledWords must NOT be in the correct order — always shuffle.
• correctOrder must be a valid permutation of [0, 1, 2, 3].
• The reconstructed sentence (prefix + ordered words + suffix) must be grammatically correct and natural.
• Ensure only one valid word order exists for each question.
• Default quantity: 5 questions unless user specifies otherwise.