An Agent Skill (SKILL.md) for 40+ AI agents

Maya Ortiz
@mayaortiz
1You ask

2AI writes
mayaortiz3You post
A skill for the AI agent you already use. It asks about you, writes a carousel built to be swiped to the end, and hands you the finished slides in your name.
LightFit/Works with
It's an Agent Skill — the open SKILL.md format Anthropic published, and OpenAI, Google, GitHub, Cursor and Nous Research adopted. Pick yours to see where it goes.
$
Unzip into your skills folder. Then ask for a carousel, or call it by name.
+ every agent on the open standard's list








Writing works in all of them. Drawing the PNGs needs an agent that can run Python — most can. About Agent Skills
Download lightfit-carousel.zip, then put it where your agent reads skills. Ask for a carousel and it starts with a few questions about you.
| Agent | Where the skill goes |
|---|---|
| Claude Code | unzip lightfit-carousel.zip -d ~/.claude/skills/Unzip into your skills folder. |
| Claude.ai & desktop | Settings → Capabilities → Skills → Upload → lightfit-carousel.zipUpload the zip under Skills in Settings, with code execution on. |
| ChatGPT & Codex | unzip lightfit-carousel.zip -d ~/.agents/skills/Codex CLI, IDE and app read your agents folder; the ChatGPT desktop app lists it under Skills. |
| Gemini CLI | unzip lightfit-carousel.zip -d ~/.gemini/skills/Unzip into Gemini CLI's skills folder. |
| Cursor | unzip lightfit-carousel.zip -d ~/.cursor/skills/Unzip into Cursor's skills folder. |
| GitHub Copilot & VS Code | unzip lightfit-carousel.zip -d ~/.copilot/skills/Unzip into Copilot's skills folder, or .github/skills/ in a repo. |
| Hermes Agent | unzip lightfit-carousel.zip -d ~/.hermes/skills/Unzip into Hermes' skills folder; it starts in your next session. |
| OpenCode | unzip lightfit-carousel.zip -d ~/.config/opencode/skills/Unzip into OpenCode's skills folder. |
| Any other skills agent | Its skills folder; most also read ~/.agents/skills/. |
LightFit/Examples
Pick a use case. The message is exactly what was sent; the post is what the skill rendered from it. Tap any slide to see it big.

LightFit/How it works
You answer once: name, face, audience, offer and the ask. It builds the post from hook to action, never invents a number, and renders it in thread or quote-over-photo, light, paper, night or black, 4:5, 1:1 or 9:16.
1You answer
2It thinks it through
3In your look
LightFit/Pricing
LightFit Carousel Skill
$70one payment
A download you keep. No refunds once bought.
After paying, it downloads right there and lands in your inbox too.
It's a digital download, so it isn't refundable. Look through the examples above first. Refunds policy
LightFit/FAQ
Any agent that reads Agent Skills, the open SKILL.md standard: Claude Code, Claude.ai and the Claude desktop app, ChatGPT and Codex, Gemini CLI, Cursor, GitHub Copilot and VS Code, Hermes Agent, OpenCode, OpenClaw, Goose, Amp, Kiro and 40+ more. Install it once and it works the same in each.
A folder with a SKILL.md file that teaches an AI agent how to do one job well, plus any scripts it needs. Anthropic published the format as an open standard (agentskills.io), and OpenAI, Google, GitHub, Cursor, Nous Research and others adopted it, so one skill works across their agents. The LightFit Carousel Skill teaches yours to write and render carousels.
Unzip lightfit-carousel.zip into ~/.claude/skills/ (or .claude/skills/ inside a project), start a new session and ask "Write me a carousel about …". On Claude.ai, upload the zip under Settings → Capabilities → Skills with code execution on. Every other agent's folder is in the install guide on this page.
Yes, one that runs inside the AI you already use, with no separate app or account. It asks about you, your audience and your offer, writes the carousel from hook to call to action in your voice, and renders the slides as 1080px PNG images for Instagram, LinkedIn, X and Threads.
Yes. Put the folder in ~/.agents/skills/ and Codex (CLI, IDE extension and app) picks it up; the ChatGPT desktop app shows it under Skills. Type @lightfit-carousel, or just ask for a carousel.
Yes. Hermes reads skills from ~/.hermes/skills/ and follows the same open standard. Unzip it there and start a new session.
Both, wherever the agent can run Python with Pillow — Claude Code, Codex, Gemini CLI, Cursor, Hermes and Claude.ai with code execution on. Where it cannot, it still writes the whole carousel, slide by slide, ready to paste into a design tool.
No. Every number, result and quote has to come from you. Anything missing is left as a marked placeholder, and it won't render a slide with a placeholder on it unless you say so.
No. The skill is a file you download and keep, so it is not refundable once bought. Look through the example carousels and prompts on this page before you buy. If the file never arrives, will not open, or you were charged twice, write to email@lightfitstudios.com and we will put it right.
No. The skill renders its own images in your agent. If you want any Google font, gradients, presets and photo cropping on top, open the same slides in the LightFit Studios carousel studio.
Ask in plain words — "shorter", "make slide 3 the hook", "dark and square for LinkedIn" — and it rewrites and renders again.
Right where you paid. The moment the payment goes through you land on our page with a Download button, and lightfit-carousel.zip is emailed to you automatically too. Lost it? Open lightfitstudios.com/thanks and paste the payment ID from your receipt.
One file, lightfit-carousel.zip: the skill in the standard Agent Skills format (its instructions, the renderer and the Inter font, so slides look the same on every machine). Unzip it into your agent's skills folder, or upload it as it is on Claude.ai. Install once, use for every post.