agentleFS
whanoftheworldwhanoftheworld/graphic-skills/skills/brand-intake/SKILL.md

brand-intake

Skill0 starsChanged 34 days agoMITView on GitHub

Hand over an existing brand (logo, fonts, colors, guidelines, past work) so the agent memorizes it and audits it.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask the human to hand over whatever they have, in any form, then stop asking and look at it:

    line 19
  2. Distinctive:

    Read every item yourself rather than asking the human to describe it.

    line 32
  3. Distinctive:

    Guidelines and past work: read them fully; list the rules they state and the rules the work actually follows.

    line 38
  4. Distinctive:

    Write what you found as facts, separate from what should change.

    line 44

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Brand intake

The first move of every studio is to look at what exists before proposing anything. This skill takes everything the human already has, memorizes it in the brand memory, and audits it for gaps and contradictions.

Load the brand-context and design-craft skills (the Skill tool, or read their SKILL.md). Reply in the user's language.

1. Set up

Find or create brand/ and the brand memory. If BRAND.md already exists, this run is an update: merge into it, show what changed, and overwrite nothing the human marked decided without asking. Create brand/inbox/, brand/assets/{logo,fonts,images,references}/.

2. Collect

Ask the human to hand over whatever they have, in any form, then stop asking and look at it:

  • Logo files (SVG best, then PNG or PDF), and any older versions.
  • Fonts in use: files, names, or screenshots of text they like.
  • Colors: hex codes, swatches, or images that carry them.
  • Brand guidelines, decks, or a past designer's files.
  • Past work: posts, flyers, packaging, signage, the website or social handles.
  • References: logos, fonts, and brands they love or hate, with any reason they already have.

Files go in brand/inbox/ or are given by path or URL. An image attached only in chat is recorded by description; ask for the original file when it is a logo.

3. Extract what is there

Read every item yourself rather than asking the human to describe it.

  • Logos: for SVG, read the fills and font-family straight from the source. For raster, note resolution and flag low quality.
  • Colors: exact hex from SVG, CSS, or the human. Colors you estimate from an image are marked estimated, and the human confirms them.
  • Fonts: take names from files and CSS; for images, follow "Reading a font from an image" in typography.md and report closest matches with confidence.
  • Websites: fetch the page and its CSS for font-family, color values, and spacing. A page that renders only through JavaScript may hide them; say so.
  • Guidelines and past work: read them fully; list the rules they state and the rules the work actually follows.

File each original into brand/assets/…; leave originals untouched. Record each font's source and licence, and copy font files only when the licence allows.

4. Audit

Write what you found as facts, separate from what should change. For each pillar (logo, typography, color, graphic language, layout, voice):

  • Established: used consistently, with evidence.
  • Inconsistent: name the contradiction (three different blues, four fonts) and where each appears.
  • Missing: nothing exists to observe.

Compute, never guess: run design-craft's contrast.mjs on every text and background color pair found. For each pillar marked Inconsistent, ask the human which version is canonical. Check the logo at 16 px and in one color. Compare the palette and marks with the category's competitors.

5. Write and read back

Write what you found into the matching sections of the brand memory, with every pillar's status (observed, open, or deferred), each taste-evidence item with the attribute the human gave, and a Decision log row for anything the human confirmed. Then read back, in plain words, what you memorized and what you found wrong, and ask for corrections. A pillar the human confirms as it is (including their pick of the canonical version for an inconsistent one) becomes decided. If they keep the whole identity, set Direction to decided with the note "existing identity".

The intake is done when every file and link the human gave is cataloged or skipped with a stated reason, every pillar has a status with evidence, every contrast figure was computed, and the human confirmed the read-back.

Then recommend the next command: gaps in strategy, personality, or taste lead to /grill-brand; a coherent identity that needs consolidating leads to /brand-system; a tired identity that needs a new look leads to /design-routes.

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.