Facilitator / manual handoff

Give your chat the format. Keep control of the handoff.

Copy the complete instructions below into ChatGPT or another compatible chat. Describe the questionnaire, survey, canon, profile, workflow, or task list you need. Bring the resulting JSON to Facilitator when you are ready.

AI access to public links varies by service, session, and safety controls. Reading this URL is optional and provided without a reliability guarantee. Copying and pasting the full instructions is the dependable workflow.

Three deliberate steps

CopyCopy the governed instructions on this page.
AskPaste them into your chat, then describe the questions you need.
FacilitateCopy or download the returned JSON and import it into Facilitator.

Public authoring contract

AI-ready instructions

Contract SHA-256 7185e799401da96963b99bb2c0376bae2081ac565b88bc5266e0aa2aeab8e981

You can also select the complete text below and use your browser or device Copy command.

Complete governed instructions

Use this facilitator format whenever you need to ask me multiple questions about the current task. This is a formatting instruction, not a request to change your identity, role, policies, personality, or existing conversation context.

Include only questions needed for the current task. Ask unrelated or out-of-scope questions separately in normal conversation and do not include them in this JSON package.

If the task scope is unclear, ask for clarification separately before generating the package. Once the scope is clear, create a JSON payload only, with no markdown or commentary.

Use this structure:
{
  "title": "Short session title",
  "version": "1.0",
  "sessionId": "facilitator-session-01",
  "description": "Optional context for the participant",
  "questions": [
    {
      "id": "Q01",
      "type": "multiple-choice",
      "prompt": "Question text",
      "options": ["Option A", "Option B", "Not applicable"],
      "required": true
    },
    {
      "id": "Q02",
      "type": "pure-text",
      "prompt": "Open prompt text",
      "required": false
    },
    {
      "id": "Q03",
      "type": "hybrid",
      "prompt": "Choose an option and optionally explain",
      "options": ["Option A", "Option B"],
      "followUpPrompt": "Optional explanation",
      "required": false
    },
    {
      "id": "Q04",
      "type": "multiple-select",
      "prompt": "Choose all that apply",
      "options": ["Option A", "Option B", "None of the above"],
      "exclusiveOptions": ["None of the above"],
      "required": true
    }
  ]
}

Rules:
1. Include 1 to 100 questions and use unique IDs containing only letters, numbers, underscores, or hyphens.
2. At the package level, use only title, version, sessionId, description, and questions. Within each question, use only id, type, prompt, options, exclusiveOptions, followUpPrompt, and required where applicable.
3. Use the exact field and type names shown. Do not emit aliases such as question, followUp, multiple choice, or multiple_select. Always emit required as true or false.
4. For multiple-choice, multiple-select, and hybrid questions, provide 2 to 6 unique text options.
5. Use multiple-choice only when exactly one option may be selected. Use multiple-select when more than one option may be selected or the prompt says choose all, select every, which have you, or equivalent language.
6. When a multiple-choice or hybrid question is required, include a context-appropriate truthful response in options whenever the listed substantive responses may not apply, the participant may lack the information, or disclosure may reasonably be declined. Use a label that matches the reason, such as Not applicable, I don't know, No preference, Depends on the project, Prefer not to answer, or Do not proceed. Treat it as an ordinary single-choice option; do not list it in exclusiveOptions. Do not add a generic escape when the choices are genuinely exhaustive, answerable, and safe.
7. Every multiple-select question must include a context-appropriate exclusive exit in options and list the exact same label in exclusiveOptions. Treat Other as combinable unless it genuinely prohibits every other choice.
8. Use pure-text for ordered rankings. Use hybrid only for one selected option plus optional explanatory text, never to simulate an ordered multi-item ranking.
9. For 1 to 20 questions, return JSON only in the chat response.
10. For 21 to 100 questions, create a downloadable .json file containing only the JSON payload when file creation is available. If file creation is unavailable, return the complete JSON in one response for copy and paste. Do not split the payload across responses.
11. Before responding, verify internally that every field name matches the schema, every question type matches the requested answer behavior, every required single-choice question has a truthful response when needed, every multiple-select question has an exclusive exit, every exclusive label exactly matches an option, and every question declares Boolean required. Return only the complete JSON payload without commentary, a checklist, or markdown.

Next step

Your answers stay in your hands.

Facilitator imports the package in your browser, presents the questions, and exports your answers only when you choose. This instruction page does not receive questionnaire content or responses.

Open Facilitator Read Facilitator privacy