You saw the LinkedIn post. Here is the kit. This is the exact folder my Image Agent uses to generate one infographic every day. Open it in Claude Code. It reads 764 lines of creative system before touching a single pixel. Visual metaphors, density rules, 12 design moves, anti-template checks, quality gate. The same brain my agent uses. How it works: You send your post caption. That is it. The agent handles everything. Picks the visual metaphor. Writes the Gemini prompt. Generates the image. Runs quality checks. Delivers the PNG. You do not pick layouts. You do not write prompts. You do not give design instructions. The agent figures it out. What is in the kit: - CLAUDE.md -- the agent brain. Full execution checklist it follows for every image. - scroll-stopping-visuals.skill -- 764 lines. Visual metaphors, 12 design moves, density standards, quality gate. The agent reads this before every image. No exceptions. - 5 example briefs -- one per visual metaphor (hourglass, cheat sheet, comparison, pyramid, metro map). So the agent knows how to think about different topics. - feedback-log.md -- this is how the agent learns your taste. Say "headline too small" and it adjusts next time. - README.md -- setup in 3 steps. Setup: 1. Get a Gemini API key (free at aistudio.google.com/apikey) 2. Paste it in the .env file 3. Open the folder in Claude Code First message to send: Set up my brand The agent asks 5 questions. Your name, handle, brand colors, vibe. Takes 2 minutes. Saves your brand config. Updates the skill file to match YOUR style. After that, just paste your caption: "Here is my LinkedIn post: [paste]. Make me an infographic for this." Or keep it simple: "make me an infographic about 5 ways to save time with AI" Cost is about $0.04 per image. A dollar a month for daily posting. The file is below. Download it. Ship tonight. Drop a comment when you generate your first one. I want to see what you make.