Skip to content
Back to skills

Thumbgate Brand Voice

ASecurity

Make any ThumbGate-facing copy — landing/pricing/compare/guide pages, README, postinstall banners, CLI receipts, Reddit/LinkedIn/Bluesky/Threads posts, docs, outreach drafts, blog/launch content — sound like ThumbGate: direct, technical, honest, anti-hype. Use BEFORE writing or editing customer-facing or marketing text, and when the user says "write a landing page", "draft a Reddit/LinkedIn post", "write the README", "make this on-brand", "punch up this copy", or "review this copy for voice"....

  • 27 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 3, 2026
ai-agentsrustgo

Works with

  • cli

Security analysis

A100/100

Pro scans all 7 files and shows the line behind each finding

Scanned September 3, 2026

npx -y skills add IgorGanapolsky/ThumbGate --skill thumbgate-brand-voice --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Thumbgate Brand Voice?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Thumbgate Brand Voice
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/igorganapolsky-thumbgate-brand-voice/badge)](https://www.skillsdirectory.com/skills/igorganapolsky-thumbgate-brand-voice)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: ThumbGate Brand Voice
description: Make any ThumbGate-facing copy — landing/pricing/compare/guide pages, README, postinstall banners, CLI receipts, Reddit/LinkedIn/Bluesky/Threads posts, docs, outreach drafts, blog/launch content — sound like ThumbGate: direct, technical, honest, anti-hype. Use BEFORE writing or editing customer-facing or marketing text, and when the user says "write a landing page", "draft a Reddit/LinkedIn post", "write the README", "make this on-brand", "punch up this copy", or "review this copy for voice". Do NOT use for code comments, commit messages, internal notes, or non-ThumbGate copy — and never to add fabricated traction or unsupported claims (honesty is the brand).
---

# ThumbGate Brand Voice

Orchestrator skill for producing on-brand ThumbGate copy. Method adapted from
the "Claude brand skill" approach (contrast-based learning: show the model a
good line, a line that goes too far, and a line that dies of beige).

## When to use

Use this whenever you are about to write or edit copy a human will read as
ThumbGate: landing/pricing/compare/guide pages, README, postinstall banners,
CLI receipts, social posts (Reddit, LinkedIn, Bluesky, Threads), outreach
drafts, or blog/launch content. Do **not** use it for code comments, commit
messages, or internal notes.

## How to use (in order)

1. Read [`brand-foundation.md`](./brand-foundation.md) — who ThumbGate is, the
   5 personality traits, and the explicit "what we aren't" list.
2. Read [`voice-and-tone.md`](./voice-and-tone.md) — per-trait good / too-far /
   too-flat rewrites. Match the "good" column; never the other two.
3. Read [`content-formats.md`](./content-formats.md) — which traits amplify or
   recede per channel, with on/off-brand examples per channel.
4. Read [`voice_profile.md`](./voice_profile.md) — for specific pacing, tone, and vocabulary guidelines.
5. Read [`body_of_work.md`](./body_of_work.md) — to anchor the core technical thesis and foundational opinions of ThumbGate.
6. Consult [`tokens.json`](./tokens.json) — for color palettes, typography, and styling constraints when generating assets.
7. Draft. Then **stress-test**: reread your draft against the "what we aren't"
   list and the channel rules. If a sentence could appear verbatim on a generic
   "AI governance platform" site, rewrite it until it could only be ThumbGate.

## The one rule that overrides everything

**Never overclaim.** ThumbGate is pre-revenue and honesty is the brand. No
fabricated traction ("trusted by N teams"), no benchmark claims without a cited
source, no "live" integration we don't ship. Honest-and-specific beats
impressive-and-vague every time.

Files in this skill

  • SKILL.md2.6 KB
  • body_of_work.md2 KB
  • brand-foundation.md2.9 KB
  • content-formats.md2.6 KB
  • tokens.json1.1 KB
  • voice-and-tone.md2.6 KB
  • voice_profile.md1.8 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.