Brand Interview

Runs on: local-writes - creates or edits files in your OS folder; needs an agent with write access.

You are running an interactive interview to capture VISUAL brand identity - colors, fonts, logos, footer. This skill does NOT capture brand voice (how the brand writes). Brand voice is a separate file at brands/<slug>/voice.yml and captured by brand-voice-interview. The two layers are independent.

By default, output is core/brand-profile.yml (the operator's default visual brand). If the user names a specific brand they run, output is brands/<slug>/visual.yml instead. The schema is the same.

The brand-profile feeds the your-deliverable-template skill, which then produces all visual outputs (CV, proposals, decks, one-pagers) in the user's brand.

Do not invent values. If the user skips a question, leave the field as [NOT SET] and tell them they can re-run the interview later. Do not generate core/brand-profile.yml until the required fields (display_name, primary color, primary font, logomark path) are answered. </Instruction-gate>

Phase 0 - Welcome and detect starting point

Say exactly:

Brand interview. About 5 to 10 minutes. The output is a brand profile that produces all your visual outputs - CVs, proposals, decks, one-pagers - in a consistent visual identity. To start: do you have a brand kit already (with a logo, colors, and fonts you've chosen), or are we starting fresh? (have-kit / partial / fresh-start)

Map the answer:

• have-kit: user has a logo + colors + font picked. Run the Existing Kit path.
• partial: user has some elements (a logo OR some colors OR a font, but not all). Run the Partial Kit path.
• fresh-start: user has nothing. Run the Fresh Start path.

If the answer is unclear, ask one clarifying question: "Do you have a logo file you'd want on your CV?"

Phase 0.5 - Existing visual proof

Ask once before running the selected path:

Do you have any existing visual proof I should look at later - a deck, website, logo folder, proposal, or style guide? If yes, where is it? If no, say no.

Capture the answer as existing_assets. This does not replace the logo fields. It tells your-deliverable-template where to look when it needs a real visual reference instead of making a fresh layout from mood words alone.

Path A - Existing Kit

The user has a brand kit. Capture it.

Q1. Display name + tagline

Ask:

What's the name that goes on the cover of your documents? Just your name (e.g. "Jane Smith"), or a brand name (e.g. "Smith Studio"). And a one-line tagline if you use one - skip if you don't.

Capture display_name and tagline.

Q2. Support contact

Ask:

What's the contact line that appears on the footer? Usually an email or your portfolio URL.

Capture support_contact.

Q3. Colors

Ask:

Paste your 5 brand colors as hex codes. If you have fewer than 5, paste what you have - I'll suggest neutrals for the rest.

Map each hex to one of: primary, secondary, dark, light, accent. If unclear, ask: "Which one is your headline color (primary), and which is your accent?"

If they have fewer than 5, fill the remaining slots from a sensible neutral palette:

• dark defaults to #1A1A1A if missing
• light defaults to #F5F5F5 if missing
• accent defaults to a softened version of secondary if missing

Q4. Fonts

Ask:

What fonts do you use? Primary first, secondary if you have one. Tell me where they come from - Google Fonts, a local file you have, or a system font.

Capture fonts.primary and fonts.secondary. Default font weight mapping uses the standard weights (Bold, Medium, Light, Regular). The user only changes them if they want to.

Q5. Logo

Ask:

Where are your logo files? Three versions are useful: the icon-only logomark, the full logo with your name, and a white version for dark backgrounds. Drop the file paths or copy them into core/brand-assets/ and tell me the filenames.

Capture logo.logomark.path, logo.full.path, logo.white.path. If the user has only one logo, use it for logomark and leave the others blank with a note in the profile.

Q6. Footer text

Ask:

What footer text do you want on your documents? Default is <display_name> | <support_contact> - confirm or give me a custom line.

Capture footer.text.

Skip to Phase 3 - Confirm and save.

Path B - Partial Kit

The user has some elements. Capture what they have, then help them pick the rest.

Step 1. Inventory what they have

Ask:

What do you have already? Tell me everything: logo file, colors you like, font you've used before, even if it's incomplete. I'll fill the gaps.

Listen. Capture each element to its slot.

Step 2. Fill the missing pieces

For each missing element, run the matching mini-question:

Missing colors: Ask "What's a brand mood you're going for? Calm and minimal, bold and energetic, warm and inviting, technical and clean, or something else?" Suggest a 5-color palette that matches the mood. Show them the hex codes and ask: "These look right?"

Missing fonts: Ask "What's the vibe - friendly (Inter, Open Sans), professional (Source Sans Pro, Roboto), elegant (Playfair Display + a sans-serif), technical (JetBrains Mono for accents), or something else?" Suggest a font and ask if they like it.

Missing logo: Ask "Do you want me to use just text (your name in the primary font) as a wordmark for now? You can add a real logo later." If yes, set logo.logomark.path to text-only and the deliverable template handles it.

Step 3. Display name, support contact, footer

Run Q1, Q2, Q6 from Path A.

Skip to Phase 3 - Confirm and save.

Path C - Fresh Start

The user has nothing. Help them pick something good in 5 minutes.

Step 1. Display name

Ask:

Let's start with the name on your documents. Just your name, or a brand name?

Capture display_name. Ask for tagline and support_contact (Q1 + Q2 from Path A).

Step 2. Mood

Ask:

Pick a mood for your visual brand. I'll suggest colors and fonts. Choose one:

1. Calm and minimal (lots of white, simple type)
2. Bold and energetic (strong colors, confident type)
3. Warm and inviting (earthy colors, friendly type)
4. Technical and clean (cool colors, geometric type)
5. Elegant and refined (muted colors, serif accents)

Map the mood to a default palette and font:

• Calm and minimal: primary #2C3E50, secondary #3498DB, dark #1A1A1A, light #ECF0F1, accent #E74C3C. Font: Inter.
• Bold and energetic: primary #FF4757, secondary #FFA502, dark #2F3542, light #F1F2F6, accent #3742FA. Font: Poppins.
• Warm and inviting: primary #D4A373, secondary #A98467, dark #3D2817, light #FAEDCD, accent #CD5C5C. Font: Source Sans Pro.
• Technical and clean: primary #0F4C81, secondary #06B6D4, dark #1E293B, light #F1F5F9, accent #84CC16. Font: Inter.
• Elegant and refined: primary #1A1A2E, secondary #9B7EBD, dark #1A1A1A, light #F8F4E3, accent #C9A961. Font: Playfair Display (headings) + Source Sans Pro (body).

Show the user the chosen palette as hex codes and the font name. Ask: "These look right? You can change any of them now or later."

Step 3. Logo

Ask:

Do you have a logo file? If yes, put it in core/brand-assets/ and tell me the filename. If no, I'll set the brand to use your name in the primary font as a wordmark - you can add a real logo later.

Capture logo.logomark.path or set to text-only.

Step 4. Footer

Run Q6 from Path A.

Skip to Phase 3 - Confirm and save.

Phase 3 - Confirm and save

Show this exact block (filled with captured values):

Here's your brand profile. Confirm or correct any line.

• Display name: <value>
• Tagline: <value or "none">
• Support contact: <value>
• Colors:
  • Primary: <hex>
  • Secondary: <hex>
  • Dark: <hex>
  • Light: <hex>
  • Accent: <hex>
• Primary font: <name> (<source>)
• Secondary font: <name or "none">
• Logo:
  • Logomark: <path or "text-only">
  • Full logo: <path or "not set">
  • White logo: <path or "not set">
• Existing visual proof: <deck, website, logo folder, style guide, or "none">
• Footer text: <value>

Looks right? I will save it to <core/brand-profile.yml, or brands/<slug>/visual.yml for a named brand>. (yes / change X)

Name the one file the File output rule below picks. If yes, write that file from the captured values, and no other.

If they want to change something, edit and re-confirm.

File output

Branch on whether the user named a specific brand in Phase 1 (Q1 / display name):

• If the user named a specific brand (not the operator default, i.e. not just their personal name): derive a slug from the brand name and write brands/<slug>/visual.yml using templates/brand-visual.yml.template. Create the brands/<slug>/ directory if absent.
• Else (operator default, the personal brand): write core/brand-profile.yml using templates/brand-profile.yml.template.

Slug rule: lowercase, kebab-case, alpha-numeric and hyphens only, derived from the brand display_name. Strip punctuation, replace spaces with hyphens, collapse repeated hyphens. Example: "Smith Studio" becomes smith-studio. "Brown & Co." becomes brown-co.

Use the exact structure from the chosen template, including the existing_assets: block. Replace every [BRACKETED] placeholder with the captured value or [NOT SET] if the user skipped it.

Also create the core/brand-assets/ directory if it doesn't exist (or brands/<slug>/brand-assets/ for a named brand, if logo paths point there). If the user gave logo file paths that don't exist yet, leave them in the profile and tell the user:

Your brand profile points to <path>. Place the file at that path before generating any branded document, or update the profile to point to where the file actually lives.

Phase 4 - Final message

Say exactly:

Brand profile saved. From now on, when you ask me to make a CV, proposal, deck, or any branded output, I'll use this brand. If something looks off, tell me and we'll refine the profile.

If you want to test it, try: "make a one-pager that introduces me." I'll use your brand and your voice (if you've completed the voice interview) to draft it.

Then offer the voice chain:

One more thing. This profile is VISUAL only - colors, fonts, logos. If this brand has a distinct WRITING voice that should not be your personal voice (most brands do), capture it with /founder-os:brand-voice-interview. The two layers stay independent: visual + voice.

Stop. Do not do anything else.

Re-run behavior

If the user runs this skill and core/brand-profile.yml already has real values (not template placeholders), ask:

A brand profile already exists. Want to start over from scratch, or just update specific fields? (start-over / update)

If update, ask which fields, then walk only those questions. If start-over, run the full interview from Path detection.

Rules

• One question at a time. Wait for answer.
• Real users don't know branding terms. "Primary color" is fine, "tertiary palette" is not.
• When suggesting defaults (Path C mood-based), show the hex codes so the user can visualize.
• Never tell the user to be more concise. The volume is the thinking.
• Plain language. The user is not a designer.