prompt-x03 — Structured Editor
Nine semantic fields guide every prompt
Role, Tone, Context, Task, Reasoning, Examples, Output Format, Constraints, and Tools. Each field answers a specific question about your prompt. Start with 3 fields or use all 9.
Why structure matters
Without structure
A wall of prose. The AI guesses your intent. Some instructions contradict each other. You get inconsistent results across platforms and wonder why Tuesday's prompt stopped working on Friday.
With the 9-field anatomy
Each field answers one question. Nothing is ambiguous. The compiler knows exactly how to format for each platform. You get the same quality output from Claude, GPT-4, and Gemini.
The 9 fields
Derived from Anthropic's prompting research and battle-tested across thousands of production prompts. The order follows a cognitive flow — each field builds on the previous one.
Role
Who the AI becomes. Not just a job title — a complete professional identity with experience level, domain expertise, and communication style.
Example
"Senior technical writer with 10 years of API documentation experience at developer-tools companies."
Tone
How the AI communicates. Calibrates formality, precision, and personality across the entire response.
Example
"Direct and technically precise. Avoid filler words. Use active voice."
Context
What the AI needs to know before starting. Background information, constraints, audience, and domain-specific knowledge.
Example
"The API serves B2B SaaS companies. Users are senior engineers. Documentation follows Stripe's style guide."
Task
What the AI actually does. The core instruction. Specific, measurable, and unambiguous.
Example
"Write a migration guide from v2 to v3 of the Authentication API, covering all breaking changes."
Reasoning
How the AI should think. Chain-of-thought instructions, decision frameworks, or step-by-step approaches.
Example
"For each breaking change: explain what changed, why, and provide before/after code samples."
Examples
What good output looks like. Few-shot examples that anchor quality and format expectations.
Example
"See the attached v1→v2 migration guide as a reference for structure and depth."
Output Format
How the response should be structured. Markdown, JSON, numbered steps, tables — explicit format rules.
Example
"Markdown with H2 per breaking change. Each section: Description, Before, After, Migration Steps."
Constraints
What the AI must avoid. Guardrails that prevent common failure modes and keep output focused.
Example
"Do not reference internal codenames. Do not suggest workarounds — only official migration paths."
Tools
What external capabilities the AI can use. Function calls, code execution, file access, web search.
Example
"Access to the v3 OpenAPI spec. Can execute Python for schema validation examples."
Cognitive flow
The field order isn't arbitrary. It mirrors how you'd brief a human expert — identity first, then context, then instructions.
Three complexity levels
Not every prompt needs 9 fields. Start simple and add structure when the task demands it.
Simple
3 fields
- Role
- Task
- Output Format
- Tone
- Context
- Reasoning
- Examples
- Constraints
- Tools
Quick tasks, one-shot queries, chat-style interactions
Standard
5 fields
- Role
- Tone
- Context
- Task
- Output Format
- Reasoning
- Examples
- Constraints
- Tools
Professional workflows, repeatable processes, team prompts
Advanced
9 fields
- All 9 fields
Production systems, API-served prompts, high-stakes outputs
Works with
Structure your first prompt
Start with 3 fields. Add more when the task demands it. The editor guides you either way.
Start free.