Voice Check

Install in one command

mkdir -p ~/.claude/skills/voice-check && curl -fsSL https://eliteaiadvantage.com/skills/voice-check/SKILL.md -o ~/.claude/skills/voice-check/SKILL.md

Voice Check is the gate between 'I think this is done' and 'I'm about to publish.' Run any draft through it (blog post, white paper, LinkedIn longform, IG carousel, X thread, FB post, email, sales copy) and it returns a clean PASS or NEEDS REVISION verdict, names every voice violation by exact excerpt, and proposes the rewrite for each one.

It loads a documented brand voice from a Markdown file (one per brand) and checks the draft against the voice doc's hard rules (em-dashes, emojis, banned phrases, hedging, hype language, closing fluff) and the soft rules (Voice DNA mismatches, audience drift, anti-pattern matches, surface-specific length and structure). It does not rewrite for argument or developmental edits. One job: catch off-voice slips and propose the exact fix.

Built for solo operators and small teams running multiple brands. Default brand is Elite AI Advantage; switchable via flag to any other documented brand voice (Boxpress.io, FUEL UP Your Marketing, or your own).

• →Catches every em-dash, emoji, and banned phrase before they ship, even the ones that slip past your own re-read.
• →Names exact excerpts and proposes the exact rewrite for each violation, so revising is a 30-second copy-paste rather than a full re-edit.
• →Loads your documented brand voice from a Markdown file you control, so the rules are your rules and changes propagate everywhere.
• →Supports multiple brands via a single flag, so a solo operator running three brands gets one tool instead of three checklists.
• →Returns a clean binary verdict (PASS or NEEDS REVISION), no caveats, so you know whether to ship or revise.
• →Surface-aware checks (longform vs blog vs carousel vs X thread) catch length and structure mismatches that brand-only checks miss.

• •Right before publishing any blog post, white paper, LinkedIn longform, social post, or email.
• •When you've just finished a draft and want a second pass before it ships.
• •When you're generating content with another AI and want to filter the output through your brand voice before it goes live.
• •When a teammate or contractor sends you a draft that needs to land in your voice.
• •Anytime you've written something and aren't sure whether it sounds like you or like a generic consultant.

Without this skill, Claude tends to give general-purpose feedback on a draft (it's a bit long, the hook could be stronger, etc.) and may use em-dashes or polished consultant-speak in its own output. Voice Check reshapes the response into a deterministic verdict against a documented voice, with named violations and exact rewrites. The output is something you act on in 30 seconds, not a coaching session you have to translate into edits.