agentleFS
murathany7murathany7/reel-studio/.agents/skills/new-brand/SKILL.md

new-brand

Skill5 starsChanged 8 days agoMITView on GitHub

Set up a new brand (product, app, side project) in brands/<slug>/ so videos can be made for it - facts and their sources, look (colours, fonts, logo, mascot), voice, compliance and default engines. Done as a short conversation with the person, reading their website and files first. Use before the first video for a brand, or when asked to add a brand.

The file

Files
SKILL.md

Adding a brand

The person is usually a marketer or founder. Treat this as getting to know their product, not filling in a form. Plain words, 2–3 questions per message, no file paths or jargon (see make-video, How to talk).

The conversation

  1. Ask for what you can read: "What's the product called, and is there a website, app store page or folder with your logo and past ads I can look at?" Then read all of it before asking anything else.
  2. Play back what you found, short: "So it's X for Y, the colours look like these, the font is Z. Right?" Show a quick brand still (logo, colours, a headline in each font) instead of listing hex codes.
  3. Ask only what you couldn't find, a few at a time:
    • Who is it for, in their customer's words? What does that customer struggle with?
    • Facts you're allowed to say (numbers, prices, ratings): "Only what's on your site, or are there others I can use? Where do they come from?"
    • Words or claims to never use (legal, sector rules, things the team hates).
    • Is there a mascot, character or a person who'd be happy to appear and talk? (Reels with a voice and a face reach further: worth asking now.)
    • A few accounts or ads they love, and ones they don't want to look like.
  4. Show the brand in 2–3 different looks (each look's starter with their colours, fonts and a real line) and ask which feel like them and which don't. Their answer fills the Look range.

Writing it down (quiet work)

  1. cp -R brands/_template brands/<slug> (lowercase-kebab).
  2. Facts from sources, never from memory. Record each source in Fact sources; anything the person told you is a source too ("told by <name>, <date>").
  3. README.md: who it is, audience, positioning, identity (colours, fonts, logo, mascot or presenter), voice, Look range (fits, ask first, never). Identity and art direction stay apart: the brand never fixes one layout for every piece. Logo, fonts, references → assets/ (scripts/svg-to-png.sh for SVG logos).
  4. tokens.css: real colours and fonts (plus optional look tokens, looks/README.md). Render the check still and look at it.
  5. compliance.md: keep Always, add the brand's own rules.
  6. Style frames from step 4 → assets/reference/looks/.
  7. STATUS.md (Brands) with open questions; the brand table in README.md.

A brand is ready when no ❓ or placeholder is left and the check still looks right. Then: "Ready. Want to start on the first video? Tell me what it's for."

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.