generate-shadowing-drill

name	generate-shadowing-drill
description	Create a shadowing drill from listening or speaking content, with sentence-by-sentence guidance, rhythm cues, and repetition instructions.

generate-shadowing-drill

When to Use

Invoke this skill when the learner runs /jp-shadow or when generate-daily-tasks assigns a shadowing task during grammar+shadowing weeks (weeks 9–16). Shadowing bridges listening comprehension to speaking production.

Inputs

Parameter	Type	Required	Description
level	string	yes	Learner's JLPT level: "N5", "N4", "N3"
topic	string	no	Topic preference (e.g., "travel", "greetings")
source_item_id	string	no	Specific listening or speaking item ID to derive from. If omitted, auto-selected.

Outputs

A shadowing drill object:

Field	Type	Description
source_item_id	string	The listening or speaking item used as source
source_type	string	"listening" or "speaking"
title	string	Descriptive title
sentences	object[]	Sentence-by-sentence drill entries (see Sentence Object)
overall_rhythm_note	string	General pacing guidance for this passage
repetition_instructions	string	How many times to repeat each sentence and in what mode

Sentence Object

Field	Type	Description
text_jp	string	The sentence in Japanese
text_zh	string	Chinese translation for self-check
rhythm_note	string	Specific pacing note (e.g., "pause after よろしく", "rising intonation on ですか")
focus_pattern	string	Grammar or phrase the learner should notice in this sentence

Logic

Source Selection

1. If source_item_id provided, load from data/listening.json or data/speaking.json.
2. Otherwise, select from data/listening.json by level and topic (prefer items completed in recent sessions for reinforcement).
3. For speaking source items, use sample_answer_jp as the drill text.

Sentence Extraction

• From listening items: split transcript_jp into individual sentences.
• From speaking items: split sample_answer_jp into individual sentences.
• Limit to 5–8 sentences per drill to avoid fatigue.

Rhythm Notes

Generate rhythm guidance based on:

• Sentence-final particles: か (rising intonation), ね/よ (softer falling tone)
• Long noun phrases: mark natural pause points
• Verb conjugations: note consonant clusters that need clear articulation
• Polite forms: steady, even pacing on 〜ます, 〜ます

Repetition Protocol

Standard protocol:

1. Listen mode: Read the sentence silently, note rhythm note.
2. Echo mode: Say the sentence immediately after reading (no pause).
3. Shadow mode: Say the sentence while reading it simultaneously.
4. Recall mode: Cover the text; say the sentence from memory.

Total: 4 repetitions per sentence. Reduce to 2 for learners with limited time.

Example

Output (abbreviated):

{
  "source_item_id": "listen_001",
  "source_type": "listening",
  "title": "自己紹介シャドーイング — Self-Intro Shadowing",
  "overall_rhythm_note": "Keep a steady, polite pace. The speaker pauses briefly after each sentence.",
  "repetition_instructions": "For each sentence: 1) Read silently, 2) Echo immediately, 3) Shadow while reading, 4) Say from memory.",
  "sentences": [
    {
      "text_jp": "はじめまして。",
      "text_zh": "初次見面。",
      "rhythm_note": "Short and crisp. Falling tone on て.",
      "focus_pattern": "Greeting formula — memorize as a complete chunk."
    },
    {
      "text_jp": "わたしはリンと申します。",
      "text_zh": "我叫林。",
      "rhythm_note": "Pause briefly after わたしは. Gentle drop on 申します.",
      "focus_pattern": "〜と申します (humble form of 〜と言います)"
    }
  ]
}