skillZs
★ LIVE SKILL TAGS ★
>>> LIVE SKILLS INDEX <<<
* OPEN SOURCE *
NO LOGIN, NO TRACKING
※ REAL INSTALL DATA ※
← back to all skills
cofoundy/brand-skills108 installs

brand-mod

Put a brand inside Claude Code — a branded spinner for the terminal and the desktop app (brand colors, the brand's own words, its logo), previewed in HTML for approval before it is generated as an installable mod. Use when the user says "brand our Claude Code", "branded spinner", "Claude Code mod with our brand", "custom spinner", "make the terminal on-brand", "brand the agent", "spinner with our logo", "show our brand while Claude works", or wants their team's or a client's Claude Code to look like their brand (for screen recordings, demos, or daily use). Claude Code only — mods don't exist in other agents.

How do I install this agent skill?

npx skills add https://github.com/cofoundy/brand-skills --skill brand-mod
view source ↗

Is this agent skill safe to install?

  • Gen Agent Trust Hubpass

    The skill is a developer utility for creating branded Claude Code mods. It uses a Python script to generate assets and plugin files using standard CLI tools. All behaviors are consistent with its documented purpose as an automated branding tool.

  • Socketpass

    No alerts

  • Snykwarn

    Risk: MEDIUM · 1 issue

What does this agent skill do?

Brand Mod

<p align="center"><img src="examples/cofoundy/mod/cofoundy-spinner/media/terminal.gif" alt="Branded spinner in the terminal" width="640"></p>

You turn a finished brand into a Claude Code mod: while Claude works, the terminal and the desktop app show the brand's colors, words and logo. You never ship it blind. You show it in HTML, get approval, then generate.

Before You Start

Load the brand package first. Look for brand.yaml (in ./, ./brand/, or brands/<slug>/); read it, identity.md (palette) and voice.md (vocabulary) from the same folder. No package yet? Run brand-init, then brand-identity and brand-voice. You need at least a primary color and an SVG logo.

Confirm the team uses Claude Code (the terminal, or the Code tab of the desktop app). Mods are Claude Code plugins; they don't run in Cursor, claude.ai chat, or other agents. Ask; don't assume.

Steps

  1. Write mod.json in the brand package (schema in the docstring of scripts/brand-mod.py):
    • primary from identity.md (the script derives an 8-tone scale for the glint), or stops if the brand already has its scale.
    • logo_svg (color logo: desktop app, light theme) and logo_light_svg (light logo: terminal and desktop dark theme) from assets/.
    • words per turn phase and done for the end of turn — see Words below.
    • name, slug and the domain come from brand.yaml; don't repeat them.
  2. Generate: python3 scripts/brand-mod.py <package>/mod.json --media --check Writes <package>/mod/<slug>-spinner/: the installable plugin (also its own marketplace), preview.html, a README.md with the GIFs first, and media/. --media needs agent-browser and ffmpeg; --check runs claude plugin validate and claude plugin test. Both are optional; without rsvg-convert the terminal gets no logo.
  3. Show it: open or share preview.html (two windows: Ghostty and the desktop app, light/dark toggle, "inside herdr" toggle). If your host can publish HTML pages (e.g. Claude Artifacts), publish it and hand over the link.
  4. Approval gate: the owner approves the preview or asks for changes. Changes go into mod.json and you regenerate. Never edit hooks/brand.ts by hand, and never install or ship before approval.
  5. Ship: the generated folder goes into a repo. Its README carries the install steps and the /<slug>-spinner on|off command. Set artifacts.mod: true in brand.yaml.

Words

  • Gerunds the team actually says, in their language: 3–6 per phase (thinking, tool-use, tool-input, responding, requesting). Pull them from voice.md (vocabulary, tone), not from puns on the brand name.
  • Cut anything nobody says. In testing, invented phrases ("herding subagents", "big-braining") and brand puns ("Cofounded") read as forced the moment people saw them.
  • The end of turn reads «<word> for 12s» (the engine adds "for"): past participles work ("Shipped", "Cooked").
  • Have someone on the team read the list before shipping.

What the mod does

  • Terminal: a 2×2 square of three cubes with a turning gap, then the phase word; one glint in the brand scale crosses the square and the word. The logo sits at the right edge in Ghostty or kitty run directly (kitty graphics); in herdr, tmux, zellij or screen there is no logo and no stand-in.
  • Desktop app (Code tab): the logo with a soft SMIL sweep, color logo on light themes and the light one when the OS is dark, and the word in one readable color.
  • End of turn: the brand's done words instead of "Baked".
  • /<slug>-spinner off brings back Claude Code's own spinner and is remembered.

Design decisions and the failures behind them: method.md. Worked example: examples/cofoundy/ (package, mod.json, generated mod with its README).

Related skills

  • brand-identity — the palette and logo direction this skill reads.
  • brand-voice — the vocabulary the spinner words come from.
  • brand-guidelines — add the mod's preview to the brand book's applications section.

Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.

<a href="https://skillzs.dev/skills/cofoundy/brand-skills/brand-mod">View brand-mod on skillZs</a>