design-brand
Skill4 starsChanged 12 days agoMITView on GitHub
Use this skill when the user says things like "create a brand", "build our brand", "brand book", "import my brand guidelines", "bring our brand book in", "logo", "design a logo", "brand colours", "brand voice", "write our positioning", "refresh our brand", or when scaffold or design-ui hands off because a project needs a brand made or imported before its UI is designed. Not for designing screens, mockups, or DESIGN.md — use design-ui. Not for writing a project's agent docs — use scaffold. Imports a finished brand book and logo pack, auditing contrast, font licences, claims, and logo files — or builds one from the product story, one choice at a time. Writes docs/brand/, a builder-facing README with positioning, promise, voice, distinctive assets, a palette with roles and measured contrast, licensed type, logo rules, icons, imagery, and motion. What an agent can't make — a professional logo pack, a photo shoot, a trademark check — becomes a brief and a pending decision, never a fake asset.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- asv-digital/56-skills-claude-codetheir own version, first seen Apr 7, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- asv-digital/56-skills-claude-code★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- design-uiSkill
- execute-planSkill
- plan-moduleSkill
- scaffoldSkill
- test-and-verifySkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 0 are distinctive, and 7 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
7
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
After scaffold, before design-ui, on a project with a public face — a product people choose, a landing page, a store listing — built or …
- Very specific: line 31
Read the root context file, docs/product.md, docs/decisions.md, docs/conventions.md, and docs/development-roadmap.md.
- Very specific: line 32
Look for docs/brand/, brand files anywhere else in the repo or handed over in the conversation, and DESIGN.md.
- Very specific: line 42
Read references/brand-system.md.
- Very specific: line 51
Read references/brand-strategy.md.
- Very specific: line 59
Read references/brand-system.md.
- Very specific: line 68
Write docs/brand/README.md in the format of references/brand-files.md: only what was confirmed, every override named, and every brief and …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
design-brand
Settles who the product is — its story, its voice, and the marks and colours that carry them — in files every later agent reads, before design-ui maps it onto screens.
When this runs
- After
scaffold, beforedesign-ui, on a project with a public face — a product people choose, a landing page, a store listing — built or accepted indocs/decisions.md - A brand book, logo pack, or guidelines arrive from founders or an agency, and need importing as the source
- A refresh or a rebrand of a project that already has
docs/brand/— see Step 0 - Not when
docs/brand/exists and the user confirms it is current (go todesign-ui), for an internal tool, for a project with no UI and no public surface, or with nodocs/development-roadmap.md(go toscaffold). Parked, it leaves a dated pending decision anddesign-uiworks from its own visual thesis
Core philosophy
- Who comes before how it looks. A palette with no story behind it is decoration, and nobody can map decoration onto a UI with reasons. Strategy, voice, and distinctive assets are settled before a single colour.
- The brand outlives the UI. It covers the site, the store listing, email, decks, and the booth.
docs/brand/is its source of truth;DESIGN.mdis one mapping of it, and belongs todesign-ui. - Nothing is settled until the user says so. One decision at a time, each with three exits: accept, override, or defer. An imported book is confirmed once, explicitly, as the source; then every audit finding is confirmed on its own.
- Every claim must be true today. A promise, proof line, certification, number, or customer name is checked against
docs/product.md,docs/decisions.md, and the code. A false one becomes a never-say rule — a brand book is not evidence. - Assets are files, never retyped. A logo is used from its file, never set in a font or redrawn. A wrong file in a pack is corrected and recorded, or listed as don't-use — never silently fixed.
- Make only what you can honestly make. Strategy, voice, palette, type, and simple wordmark explorations, yes. A professional logo pack, a photo shoot, a trademark search, legal wording, and market tests become briefs and pending decisions — never stand-ins presented as finished.
- Precedence is explicit. A dated
docs/decisions.mdentry beats the source book, and the book beats the README's summary. The README names every override and the decision behind it.
Workflow
Step 0 — Read the project, and check what exists
- Read the root context file,
docs/product.md,docs/decisions.md,docs/conventions.md, anddocs/development-roadmap.md. Who the product serves, and what it truly does today, come from there. - Look for
docs/brand/, brand files anywhere else in the repo or handed over in the conversation, andDESIGN.md. Say whether this is a new brand, an import, or a re-run. - Surface any
Status: Pendingbrand decision first — decide it now, or keep it parked. - On a re-run, classify the change with
references/brand-strategy.md: a refresh (colour, type, a cleaner mark, same story) or a rebrand (a new name, promise, or audience). Default to refresh. A rebrand is confirmed explicitly, after listing every file it reaches downstream.
Step 1 — Find the source
- A brand exists outside the repo — a book, guidelines, a logo pack, a design file through a connector the environment offers. Save every file unchanged under
docs/brand/(layout inreferences/brand-files.md) — a correction confirmed later is made in place, with the old value recorded — and extract from the files, never from memory. Ask once which file and version is the source, even when there is only one. Then Step 2. - Part of one exists — a logo with no story, a palette with no voice. Import what exists through Step 2, then build the rest through Steps 3–4. A complete-looking book usually still lacks some of Step 3 (voice dials, a trust-moment rule, competitors' claims): after the audit, name the gaps in one question — build them now, or leave them out — then confirm each one built on its own.
- Nothing — Steps 3 and 4.
Step 2 — Import: audit, then confirm each finding
Read references/brand-system.md. Extract the book into the README's sections, then audit it. Present each finding on its own:
- Contrast — measure every text pairing the book allows. Each failure becomes a rule ("never X as text on Y").
- Font licences — every face, for embedding on web and in apps. An unlicensed face gets a licensed substitute recommended.
- Claims — every promise, proof line, and number against what is true. False ones become never-say rules.
- Logo files — every file in the pack against the book's own rules. Colour corrections go in
CORRECTIONS.md; anything not corrected is listed as don't-use. Letterforms are never redrawn. - Imagery — a licence recorded per file, and third-party marks found.
- The book's open questions — ask whether each is still open; each that is becomes a pending decision.
Step 3 — Build: strategy, voice, distinctive assets
Read references/brand-strategy.md. Pre-fill each answer from the plan and docs/product.md, say where it came from, and confirm one at a time:
- Always: positioning ("the only ___ that ___"), the value proposition in one breath, a promise the product keeps today, and what competitors already say.
- Only when it applies: brand architecture (two or more products or audiences), a new category (only if the user claims one, and parked with its test until validated), and an archetype as a shorthand for the voice.
- Voice: three dials, three "we are / we are not" lines, and a say/don't-say table in the product's own moments. Trust moments — security, pricing, consent, errors — are always plain.
- Distinctive assets: name the two the brand will own for years. The palette and the mark are built from them.
- Surfaces: list where the brand has to work. No surface, no rule for it.
Step 4 — Build: the identity system
Read references/brand-system.md. For each part, propose two or three genuinely distinct options drawn from Step 3, recommend one, and confirm it before the next:
- Palette — every colour with a role and a reason traced to the story, measured contrast, failures as rules.
- Type — licence verified before a face is offered, with its source recorded.
- Logo — two or three SVG wordmark explorations under
docs/brand/logo/explorations/, each labelled an exploration. A chosen one becomes a working mark, with a pending decision for the professional pack. - Icons, imagery, motion, and sound — direction and rules, never generated photographs presented as brand photography.
Preview however the environment allows, and always give the file path — the file is written either way.
Step 5 — Write and hand off
- Write
docs/brand/README.mdin the format ofreferences/brand-files.md: only what was confirmed, every override named, and every brief and open item listed. - Append dated entries to
docs/decisions.mdin the shapescaffoldwrites — accepted or pending, never editing a past one. A deferred choice also gets an item in the context file's pending-decisions section, with the heading and placementscaffold'sreferences/templates.mdspecifies, created if absent. Add one## Related docsline pointing atdocs/brand/README.md. - Those are the only writes outside
docs/brand/.DESIGN.md,docs/design/, the roadmap, plan files, and app code are never touched — bundling fonts or logos into the app is a build task. A finding that should become a root rule, like a claim never to make, is reported forscaffold, not written. - Report what was written, what is deferred, each brief, and every conflict with the project's own docs — a binding palette in
docs/conventions.mdthe brand replaces, a claim a root rule should carry — naming which skill owns the fix. Then hand off todesign-uito map the brand onto the UI — or, after a refresh, to re-map it.
Reference files
references/brand-strategy.md— positioning, value proposition, promise, differentiation, architecture, category, archetype, voice, distinctive assets, refresh versus rebrand, and what only the market can measure. Read this in Steps 0 and 3.references/brand-system.md— palette and contrast, type and licences, the logo and its pack, iconography, imagery, motion and sound, and what an agent can't make. Read this in Steps 2 and 4.references/brand-files.md— thedocs/brand/layout, the README,CORRECTIONS.md, and brief formats, and whatdesign-uireads from them. Read this in Steps 1 and 5.
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
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.

