Nextcloud Forms JSON Creator

This skill enables creation of Nextcloud Forms JSON files from natural language descriptions, markdown files, or structured text.

JSON Structure Overview

json

{
    "title": "Form Title",
    "description": "Optional form description",
    "share_with_group": "admin",
    "questions": [
        {
            "text": "Question text",
            "type": "question_type",
            "isRequired": true/false,
            "options": ["Option1", "Option2"]  // only for choice types
        }
    ]
}

Question Types Reference

Type	Description	Has Options	Use Case
`short`	Single-line text input	No	Names, short answers
`long`	Multi-line text area	No	Essays, detailed responses
`multiple`	Multiple checkboxes	Yes	Select multiple items
`multiple_unique`	Radio buttons	Yes	Select exactly one item
`dropdown`	Dropdown menu	Yes	Select one from many options
`date`	Date picker	No	Birthdays, deadlines
`time`	Time picker	No	Appointments, schedules
`file`	File upload	No	Document submissions
`linearscale`	Rating scale	No	Likert scales, ratings
`color`	Color picker	No	Color preferences

Text to JSON Mapping Rules

Identifying Question Types

Apply these rules in order to determine the correct type:

•
Keywords → Type mapping:
- •"Name", "Email", "ID", "kurz", "short" → short
- •"Beschreib", "Erkläre", "Essay", "ausführlich", "lang" → long
- •"Mehrfach", "alle zutreffenden", "multiple choice" → multiple
- •"Wähle eine", "eine Option", "entweder oder" → multiple_unique
- •"Auswahl aus Liste", "Dropdown" → dropdown
- •"Datum", "Geburtstag", "date" → date
- •"Uhrzeit", "Zeit", "time" → time
- •"Datei", "Upload", "Hochladen" → file
- •"Bewerte", "Skala", "1-5", "Rating" → linearscale
- •"Farbe", "color" → color
•
Context clues:
- •Multiple possible answers listed → multiple or multiple_unique
- •Explicit "nur eine Antwort" → multiple_unique
- •No explicit restriction → multiple
- •Long list (>5 items) → prefer dropdown
•
Default fallback:
- •If unclear and expects short text → short
- •If unclear and expects longer text → long

Required vs Optional

•Default: Questions are isRequired: false unless explicitly stated
•
Make required when:
- •Text contains "Pflichtfeld", "erforderlich", "required", "must"
- •Critical information (name, contact in contact forms)
- •Context implies necessity (e.g., "Your name" in signup form)

Option Extraction

When extracting options from text:

•

List formats:

code

- Option A     →  ["Option A", "Option B", "Option C"]
- Option B
- Option C

•

Inline formats:

code

(Ja/Nein/Vielleicht)  →  ["Ja", "Nein", "Vielleicht"]

•
Common patterns:
- •Yes/No questions → ["Ja", "Nein"] or locale-appropriate
- •Agreement scales → ["Stimme voll zu", "Stimme zu", "Neutral", "Stimme nicht zu", "Stimme gar nicht zu"]
- •Never/Sometimes/Always → ["Nie", "Selten", "Manchmal", "Oft", "Immer"]

Validation Rules

Before outputting JSON, verify:

•
Structure:
- •Top-level has title, questions array
- •Optional: description, share_with_group
•
Each question has:
- •text (string, not empty)
- •type (one of the 10 valid types)
- •isRequired (boolean)
- •options (array) if type is multiple, multiple_unique, or dropdown
•
Options array:
- •Only present for choice types
- •Contains at least 2 items
- •All items are non-empty strings
•
Common mistakes to avoid:
- •Don't add options to non-choice types
- •Don't forget isRequired field (always include it)
- •Don't use invalid type names

Example Conversions

Example 1: Simple Contact Form

Input:

code

Contact Form

Name (required)
Email address (required)
Your message

Output:

json

{
    "title": "Contact Form",
    "description": "",
    "share_with_group": "admin",
    "questions": [
        {
            "text": "Name",
            "type": "short",
            "isRequired": true
        },
        {
            "text": "Email address",
            "type": "short",
            "isRequired": true
        },
        {
            "text": "Your message",
            "type": "long",
            "isRequired": false
        }
    ]
}

Example 2: Survey with Choices

Input:

code

Customer Satisfaction Survey

How satisfied are you? (Rate 1-5)
What did you like? (Select all that apply)
- Product quality
- Customer service
- Delivery speed
- Price

Would you recommend us? (Yes/No/Maybe)

Output:

json

{
    "title": "Customer Satisfaction Survey",
    "description": "",
    "share_with_group": "admin",
    "questions": [
        {
            "text": "How satisfied are you? (Rate 1-5)",
            "type": "linearscale",
            "isRequired": false
        },
        {
            "text": "What did you like? (Select all that apply)",
            "type": "multiple",
            "isRequired": false,
            "options": [
                "Product quality",
                "Customer service",
                "Delivery speed",
                "Price"
            ]
        },
        {
            "text": "Would you recommend us?",
            "type": "multiple_unique",
            "isRequired": false,
            "options": ["Yes", "No", "Maybe"]
        }
    ]
}

Example 3: Event Registration

Input:

markdown

# Workshop Registration

## Personal Information
- Full Name (Pflichtfeld)
- Date of Birth
- Email (required)

## Workshop Selection
Choose one workshop:
1. Python for Beginners
2. Advanced JavaScript
3. Data Science Basics

## Dietary Requirements
Select all that apply:
- Vegetarian
- Vegan
- Gluten-free
- No restrictions

## Additional Notes
Any other information we should know?

Output:

json

{
    "title": "Workshop Registration",
    "description": "",
    "share_with_group": "admin",
    "questions": [
        {
            "text": "Full Name",
            "type": "short",
            "isRequired": true
        },
        {
            "text": "Date of Birth",
            "type": "date",
            "isRequired": false
        },
        {
            "text": "Email",
            "type": "short",
            "isRequired": true
        },
        {
            "text": "Choose one workshop",
            "type": "multiple_unique",
            "isRequired": false,
            "options": [
                "Python for Beginners",
                "Advanced JavaScript",
                "Data Science Basics"
            ]
        },
        {
            "text": "Dietary Requirements (Select all that apply)",
            "type": "multiple",
            "isRequired": false,
            "options": [
                "Vegetarian",
                "Vegan",
                "Gluten-free",
                "No restrictions"
            ]
        },
        {
            "text": "Any other information we should know?",
            "type": "long",
            "isRequired": false
        }
    ]
}

Workflow

•Parse input: Identify form title, questions, and structure
•Map questions: Apply type identification rules for each question
•Extract options: For choice types, carefully extract all options
•Set requirements: Determine isRequired based on keywords and context
•Validate: Check structure against validation rules
•Output: Create clean, properly formatted JSON
•Save: Write to /mnt/user-data/outputs/ with descriptive filename (e.g., customer_survey.json)

Quality Standards

•Preserve intent: Don't over-interpret vague descriptions; ask for clarification if needed
•Clean text: Remove markdown formatting from question text (no #, *, etc.)
•Consistent formatting: Use proper indentation (4 spaces)
•German/English: Handle both languages naturally, preserve the language of questions
•Comments: Don't add _comment fields unless user explicitly requests them
•ASCII-only quotes: CRITICAL - Use only standard ASCII quotes (" for JSON, ' for nested quotes). Never use typographic quotes („", "", «») as they break JSON parsing

Common Pitfalls

❌ Wrong: Adding options to text fields

json

{
    "text": "Your name",
    "type": "short",
    "options": []  // ← Remove this!
}

✅ Correct:

json

{
    "text": "Your name",
    "type": "short",
    "isRequired": false
}

❌ Wrong: Using invalid type names

json

{"type": "text"}  // ← Should be "short" or "long"
{"type": "checkbox"}  // ← Should be "multiple" or "multiple_unique"
{"type": "select"}  // ← Should be "dropdown"

✅ Correct:

json

{"type": "short"}
{"type": "multiple"}
{"type": "dropdown"}

❌ CRITICAL: Using typographic quotation marks in JSON strings

json

{
    "description": "Die Skala reicht von „stimme zu" bis „stimme nicht zu""
    // ← Wrong! Uses „ and " instead of regular quotes
}

✅ Correct:

json

{
    "description": "Die Skala reicht von 'stimme zu' bis 'stimme nicht zu'"
    // ← Use single quotes (') for quotes within strings, or escape double quotes
}

IMPORTANT: Always use standard ASCII quotation marks in JSON:

•JSON string delimiters: " (U+0022)
•For quotes within strings: use ' (single quote) or escape with \"
•NEVER use typographic quotes: „" (German), "" (curly), «» (guillemets)
•These will cause JSON parsing errors!