Loading...
The prompt structure Moonrank's LinkedIn Agent uses to brief image models: one visual metaphor, exact text, and a house style that makes every post read as part of the same series.
Download the prompt structure, see how the agent picks the metaphor, or run it inside the LinkedIn Agent.
The LinkedIn Post Visual skill turns a finished post into a prompt an image model can follow: one visual metaphor, the exact words to render, and a consistent house style. It also tells you when a real screenshot beats a generated image.
It's written as a SKILL.md, a plain-language instruction format that people and AI agents can both follow. Use it with ChatGPT or Gemini, hand it to a designer, or let the LinkedIn Agent write the prompt with each draft.
Downloading the playbook is free with a Moonrank account. Pairing every draft with a visual prompt is what the LinkedIn Agent does for you.
“My AI-generated images are cluttered and the text is always misspelled. How do I brief the model better?”
Your prompts describe a mood, not a layout. The rewrite gives the model a fixed structure (headline, one hero metaphor, one row of steps, footer), lists every word exactly, locks your background and accent color, and ends with an instruction to spell the text exactly as written.
The final text, and your brand colors.
The payoff, the proof, or the offer. Never a copy of the hook.
A single visual idea that makes the point at a glance.
Format, style, exact text, and a 16:9 variant for X.
Use a real screenshot when you have one. Generate only when you don't.
The image says what the reader gets, not what the hook said.
A single concrete idea instead of a busy collage.
Same background, font, and accent every week.
Every word listed so the model spells it right.
Portrait for LinkedIn, landscape for X.