Update to AI Roleplay Chat Simulator

{
  "description": "Practise a second language through a graded text conversation in the Conversation app. Use when the user wants a language role-play, everyday conversation rehearsal, or progressively less help with replies at a chosen proficiency level. Establish a target language, situation, and feedback preference, then review submitted wording and provide reusable phrases. Do not claim audio, pronunciation assessment, formal proficiency certification, or access to unseen drafts.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 326
    }
  ],
  "name": "second-language-conversation-practice",
  "skill_md_contents": "---\nname: second-language-conversation-practice\ndescription: Practise a second language through a graded text conversation in the Conversation app. Use when the user wants a language role-play, everyday conversation rehearsal, or progressively less help with replies at a chosen proficiency level. Establish a target language, situation, and feedback preference, then review submitted wording and provide reusable phrases. Do not claim audio, pronunciation assessment, formal proficiency certification, or access to unseen drafts.\n---\n\n# Second Language Conversation Practice\n\n## Inputs\n\nEstablish the target language, approximate proficiency, a concrete situation,\nand the language the user prefers for explanations. Use plain descriptions\nsuch as beginner if no level is known. Ask about missing language or situation;\ndefault to a short exchange with feedback after finishing unless the user\nrequests another approach. Use a dialect or register only when known or needed.\n\n## Workflow\n\n1. Choose one practical communication goal, such as ordering and clarifying\n   a meal, arranging a meeting, or asking for directions. Match vocabulary,\n   sentence length, and register to the user's stated level.\n2. Agree on support using available preferences: simple reply choices first,\n   brief explanations on request, and an invitation to edit or write a reply\n   later. Do not require a formal level assessment before starting.\n3. Call `start-conversation` with known participant details and the other\n   person's role and personality. Describe the situation and target language\n   in `conversation.topic`. Use `userInitialMessage` only for a supplied or\n   agreed opening. Keep proficiency and feedback preferences in chat context.\n4. Wait for the user's Start follow-up and use current model-visible setup\n   edits. Generate a natural in-character message in the target language and\n   call `continue-conversation` with `otherPersonMessage` and 2–4 distinct\n   nonempty `possibleUserResponses`. End the turn after the tool call.\n5. Keep choices at the learner's level and make them meaningful conversational\n   replies rather than disguised grammar quizzes. Include a way to request\n   clarification when useful. Let the learner edit a suggestion or submit\n   their own text; never claim the app can hide or remove all required choices.\n6. Continue from actual submitted replies. Gradually reduce linguistic help\n   and encourage custom wording when the learner appears ready or asks for a\n   challenge. Retain 2–4 choices on every continuation call. Ease the language\n   when the learner requests help; do not infer ability from unchosen options.\n7. Address a comprehension-blocking issue with a brief natural clarification\n   in character. For an explicit coaching pause, explain the issue in the\n   preferred explanation language in chat. Resume only after a submitted reply\n   or continuation request; do not append teaching text after a tool call.\n8. When the user finishes, review only submitted wording. Separate errors\n   from optional style improvements and distinguish copied suggestions from\n   independent wording when that distinction is visible. Show the original\n   phrase, a natural revision, and a short reason. Include an overall score\n   with a stated scale when requested by the user or End follow-up, describing\n   it as feedback on this exchange rather than certified language proficiency.\n\n## Deliverable\n\nProvide up to three focused corrections, three reusable phrases from the\nexchange, and one practical next exercise. Prefer useful progress feedback\nover correcting every minor issue, and preserve the learner's intended meaning.\n\n## App boundaries\n\n- Use only `start-conversation` and `continue-conversation` for practice.\n  Add no language, proficiency, translation, audio, or score parameters.\n- Generate target-language text through the host model. Do not promise that\n  the skill changes widget interface labels or controls the app's locale.\n- Use visible chat, tool context, and model-visible scenario or submitted reply\n  state. Do not access private metadata, drafts, stored transcripts, or scores.\n- Leave Start, editing, selection, and End to the learner. Stop when requested\n  and honour plain text or free-form practice without widgets.\n- Limit assessment to text; do not claim pronunciation, listening, or speech\n  capabilities that the app does not expose.\n"
}