motionvector.dev

music_explain

SAFETY: changes the UI immediately — writes your text into the visible, attributed agent-advice panel, no human confirmation step (the text itself has no other effect: it does not alter the lesson, exercise, or player state). Call this to surface advice you have already formed from music_get_state and, if relevant, lastCompletedTake; this tool does not call a model or generate the advice itself, it only displays text you supply. Ground the explanation in evidence: cite a concrete field from music_get_state's live block (e.g. "you held up 3 fingers") or from lastCompletedTake (e.g. "3 of 4 notes were in time"), not a generic tip. Rendered as plain text, never HTML. Returns {displayed: true}.

Page: /air-instruments/

Input schema

{
  "type": "object",
  "additionalProperties": false,
  "required": [
    "message"
  ],
  "properties": {
    "message": {
      "type": "string",
      "minLength": 1,
      "maxLength": 500,
      "description": "Advice text to display verbatim, as plain text (never HTML), in the agent-advice panel. Cite a specific live or lastCompletedTake fact rather than generic encouragement. 1-500 characters. Required."
    }
  }
}