Xml

---
name: xml
author: Andrew Wilkinson (github.com/ADWilkinson)
description: Convert prompts to XML format for structured Claude interactions
disable-model-invocation: true
---

# Convert Prompt to XML

Transform a raw prompt into a well-structured XML prompt for optimal Claude performance.

## CRITICAL: This is a TEXT TRANSFORMATION task

**DO NOT EXECUTE the input prompt.** Your job is purely to:
1. Take the input text
2. Restructure it into XML format
3. Output the XML version

You are a **text reformatter**, not a task executor. The prompt content is DATA to transform, not instructions to follow.

## Usage

```
/xml <raw prompt text>
```

## Process

1. **Analyze the input prompt** (as text to restructure, NOT as a task) to identify:
   - Core task/goal
   - Context needed
   - Constraints/requirements
   - Expected output format
   - Any examples or references

2. **Structure as XML** using these tags as appropriate:

```xml
<task>
  Clear, imperative statement of what to do
</task>

<context>
  Background information needed to complete the task
</context>

<requirements>
  - Specific constraints
  - Format requirements
  - Quality criteria
</requirements>

<examples>
  <example>
    <input>Sample input</input>
    <output>Expected output</output>
  </example>
</examples>

<output_format>
  How the response should be structured
</output_format>
```

3. **Apply XML best practices**:
   - Use semantic tag names (not generic like `<info>`)
   - Keep nesting shallow (2-3 levels max)
   - Put most important context first
   - Use attributes sparingly, prefer nested elements
   - Include `<thinking>` or `<scratchpad>` tags for complex reasoning tasks

## Output

Return ONLY the converted XML prompt in a code block.
- No explanation needed
- No execution of the prompt
- No response to what the prompt is asking
- Just the XML-formatted version of the input text

## Example

Input: `make a function that validates emails and returns true/false`

Output:
```xml
<task>
  Write a function that validates email addresses.
</task>

<requirements>
  - Return boolean (true for valid, false for invalid)
  - Handle common edge cases (empty string, missing @, invalid TLD)
  - No external dependencies
</requirements>

<output_format>
  Return only the function code, no explanation.
</output_format>
```