20 commands · 3 skills · 2 agents · 16,714 tokens measured

Marketing that sounds like you, not like AI

Content that sounds like you wrote it: a voice engine built from your real posts, a humanizer that strips every AI tell, and 20 commands for the daily content grind.

See every command ↓

14-day refunds · works on 3 devices · plain Markdown you can read

claudekit · /mkt
20

slash commands

3

skills, auto-loaded

2

specialist agents

16,714

tokens, measured at pack time

Command explorer

Every one of the 20 commands

Nothing hidden behind a demo call. This is the complete /mkt command list, the same files that land in your .claude directory.

  • /mkt auditResearch

    Review your past posts and metrics: what worked, the patterns behind it, and what to double down on.

  • /mkt bioWrite

    Write your X or LinkedIn profile: header line, bio, and pinned post. The page people check before following.

  • /mkt calendarCampaigns

    Build a 30-day content calendar from your pillars, batched into 4 writing sessions instead of 30 daily scrambles.

  • /mkt competitorsResearch

    Tear down a competitor's content engine: their hooks, cadence, formats, and what's worth stealing.

  • /mkt ctaWrite

    Write CTA and offer variants that sell without sounding like a banner ad. Scored, with placement advice.

  • /mkt hooksWrite

    Get 10 first-line variants for your post, each scored 1–5 on specificity and curiosity with reasons.

  • /mkt humanizeVoice

    Strip the AI smell from any draft, then rewrite it in your voice. Shows the worst 5 fixes as before/after.

  • /mkt ideasWrite

    Generate N post-ready insight bombs, each with a named example and a number. No abstract advice.

  • /mkt landingWrite

    Write landing page copy that sounds human: hero, subhead, proof section, CTA, and FAQ.

  • /mkt launchCampaigns

    Write the full launch set: X post, LinkedIn post, Product Hunt tagline + description, and launch email.

  • /mkt leadmagnetCampaigns

    Design a lead magnet people actually want: the concept, full outline, and the landing blurb to promote it.

  • /mkt newsletterWrite

    Write a weekly letter (800–1,500 words) with the claim-concede-refine structure that keeps people reading.

  • /mkt postWrite

    Turn one idea into one post for X or LinkedIn — written in your voice, not AI voice.

  • /mkt repliesWrite

    Write engagement replies in your voice: add to the conversation, get noticed, never sound like a reply guy.

  • /mkt repurposeCampaigns

    Turn one piece into 5 formats: thread, LinkedIn post, carousel script, newsletter section, 60-sec video script.

  • /mkt rewriteVoice

    Rewrite any copy in a chosen mode: your voice, isenberg, koe, or a classic framework like AIDA or PAS.

  • /mkt swipeVoice

    Paste a viral post, learn exactly why it worked, and get a fill-in template you can reuse.

  • /mkt threadWrite

    Turn one idea into a numbered X thread where every item stands alone and the hook earns the click.

  • /mkt voiceVoice

    Build a voice file from 5–10 of your real posts so every command writes like you, not like AI.

  • /mkt welcomeWrite

    Write a 5-email welcome sequence that turns new subscribers into readers who actually open.

Flagship: /mkt humanize

Strip the AI smell, keep the argument

Strips all 14 AI tells (em dashes, 'It's not X, it's Y', banned words) from any draft, then rewrites it in your voice. Shows the worst 5 fixes as before/after.

1

Sweep all 14 tells

Em dashes as connective tissue, "It's not X, it's Y", rule-of-three lists, 34 banned words from delve to game-changer, stock openers, fake suspense. Every hit gets fixed, not flagged.

2

Litmus test per paragraph

Each paragraph needs at least one number, one name, or one opinion. A paragraph with none gets the most specific fact from elsewhere in the draft, or gets cut. Vague mush doesn't survive.

3

Re-voice with receipts

Rewrites against your VOICE.md: sentence stats, owned vocabulary, your openers and closers. Then prints the worst 5 fixes as before/after lines and a final tally. Done without counts is not evidence.

claudekit · /mkt humanize

Under the hood

3 skills, 2 agents

Skills are knowledge files Claude Code loads automatically when the topic comes up. Agents are read-only specialists a command can dispatch for research or review without polluting your main context.

Skills (3)

Auto-loaded context: the rules, benchmarks, and playbooks the commands cite. Plain Markdown you can open and read.

mkt-context

The kit's shared memory files (VOICE.md, AUDIENCE.md, PILLARS.md). Read when present, offer to create when missing, never block.

mkt-formats

Platform format specs for X, LinkedIn, carousels, newsletters, and Product Hunt. Read before formatting any output.

mkt-voice-engine

The anti-AI-tell rules and the two creator voice specs. Read before writing or reviewing any prose for humans.

Agents (2)

Dispatched by commands for focused work. Read-only by design: they research and report, they never edit your files.

content-researcher

Fetches competitor content and viral posts, returns structured patterns (hook shapes, cadence, formats). Read-only.

copy-critic

Reviews a draft against the AI-tell list and the user's voice spec. Returns the 5 worst violations with fixes. Never rewrites wholesale.

Install

One command, ready in your next session

Authenticate the CLI with your license key, then install. The files land in ./.claude (or ~/.claude with -g) and every /mkt command shows up immediately.

$ ck install marketingkit
  • ·Prefer plugins? The kit also loads via Claude Code’s /plugin install flow; both paths ship the same files.
  • ·One license covers 3 devices. Activate and deactivate from your dashboard whenever you switch machines.
  • ·Install prints a per-file token ledger, so you see exactly what 16,714 measured tokens means before your first session pays for it.

Use cases

What people actually use MarketingKit for

How do I make AI writing sound human?

/mkt humanize sweeps all 14 AI tells (em dashes as glue, negative parallelism, 34 banned words, stock openers) and fixes every hit, then rewrites the draft in your voice. It shows the worst 5 fixes as before/after lines and ends with a tally, so you can verify the sweep instead of trusting it.

How does Claude Code learn my writing voice?

/mkt voice builds a voice file from 5 to 10 of your real posts: sentence stats, owned vocabulary, how you open and close. Every other command reads that file, so /mkt post and /mkt thread write like you on day one instead of generic AI voice.

Can Claude Code plan my content calendar?

/mkt calendar builds a 30-day calendar from your content pillars, batched into 4 writing sessions instead of 30 daily scrambles. /mkt repurpose then turns each finished piece into 5 formats: thread, LinkedIn post, carousel script, newsletter section, and a 60-second video script.

How do I write a product launch with Claude Code?

/mkt launch writes the full set in one run: the X post, the LinkedIn post, the Product Hunt tagline and description, and the launch email. All of it goes through the voice engine, so the launch sounds like you wrote it on a good day.

Pricing

Two ways to get it

Both come with 14-day refunds and licenses that cover 3 devices.

MarketingKit

One-time, the kit as shipped

$99once

  • All 20 commands, 3 skills, 2 agents
  • Yours forever, exactly as shipped
  • Plain Markdown: read and edit everything
  • 14-day refunds, no forms to fight

All-Access

BEST VALUE

Every kit, every update, while active

$49.99/mo

  • All 5 kits: 101 commands, 19 skills, 13 agents
  • Every update and new release while subscribed
  • Swap focus between kits as your work changes
  • 14-day refunds, cancel anytime

FAQ

MarketingKit questions, answered

20 slash commands, 3 skills, and 2 agents as plain Markdown files, 16,714 tokens measured at pack time. Install with ck install marketingkit or as a Claude Code plugin. The voice engine, the 14-tell list, and the platform format specs are all readable text you can audit.