Update to AI Roleplay Chat Simulator

{
  "description": "Explain the tradeoffs among the current Conversation app reply choices while leaving the decision to the user. Use when the user asks to compare options, understand how a reply may land, or draft an alternative before submitting a response. Reuse visible continuation tool inputs and the user's goal. Do not choose, submit, edit widget state, or advance the role-play merely to provide coaching; request the option text if it is unavailable.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 302
    }
  ],
  "name": "response-strategy-walkthrough",
  "skill_md_contents": "---\nname: response-strategy-walkthrough\ndescription: Explain the tradeoffs among the current Conversation app reply choices while leaving the decision to the user. Use when the user asks to compare options, understand how a reply may land, or draft an alternative before submitting a response. Reuse visible continuation tool inputs and the user's goal. Do not choose, submit, edit widget state, or advance the role-play merely to provide coaching; request the option text if it is unavailable.\n---\n\n# Response Strategy Walkthrough\n\n## Inputs\n\nUse the current other-person message, available reply options, and the user's\ngoal or boundary. Obtain options from visible `continue-conversation` input\n`possibleUserResponses`; they are not duplicated in structured tool output.\nUse the current message from visible output or model-visible state. If the\nuser refers to edited wording or unavailable options, ask them to provide the\ntext. Do not infer an unsubmitted widget draft.\n\n## Workflow\n\n1. Confirm which turn and options the user means. Prefer current visible\n   context and the user's supplied text over an older picker. If a response\n   has already been submitted, explain it retrospectively and do not imply\n   that the selection can be undone.\n2. Establish the user's intended outcome. Use an existing goal when clear;\n   ask one short question when different goals would change the comparison.\n3. Explain each option's main conversational move, what it could achieve,\n   and its main tradeoff. Consider clarity, acknowledgement, information\n   gathering, commitments, and consistency with the user's limits. Describe\n   possible reactions as uncertain rather than predicting the real person.\n4. Keep the comparison neutral and concrete. If the user requests a\n   recommendation, give one conditional on their goal with a short reason.\n   Otherwise help them judge the options without selecting on their behalf.\n5. If none fits, draft one concise alternative using only known facts and\n   intended commitments. Explain which tradeoff it addresses. Avoid an\n   unsolicited list of generic communication tips.\n6. Tell the user how to use their chosen wording when helpful: select an\n   existing reply, use its edit control, or choose “write your own” and submit.\n   Leave those widget actions to the user; do not claim to update the picker.\n7. Wait for the actual submitted reply or an explicit request to continue.\n   Only then generate the next in-character message and call\n   `continue-conversation` with `otherPersonMessage` and 2–4 distinct nonempty\n   `possibleUserResponses`. End the assistant turn after that call.\n8. If the user finishes instead, give the requested assessment in chat using\n   actual submitted responses. Include a score with a stated scale if the\n   user or End follow-up requests it; do not treat compared options as choices.\n\n## Deliverable\n\nProvide a compact comparison of the available options and, when useful, one\ncustom reply. State how the alternatives relate to the user's goal and leave\nthe final decision with them.\n\n## App boundaries\n\n- Give the walkthrough in chat. Do not call `start-conversation` or\n  `continue-conversation` solely to explain, replace, or re-render choices.\n- Use `continue-conversation` only for a requested next role-play turn; add\n  no analysis, recommendation, selection, or editing arguments.\n- Do not assume access to private metadata, widget-only drafts, or all prior\n  turns. Use visible `userSelectedResponse` only as evidence of a submission.\n- Do not invoke an unregistered tool, control Start or End, or submit replies.\n  Respect requests for plain text or free-form continuation without a widget.\n"
}