Humanizer
Strips the words and rhythms that make readers spot AI writing — the giveaway vocabulary, the em dash pile-ups, the endless rule of three — and tells you what it changed and why.
Writes under-90-word cold emails plus a 3-5 touch follow-up sequence that sounds like a peer, not a template.
Founder, Conversion Factory
Writes under-90-word cold emails plus a 3-5 touch follow-up sequence that sounds like a peer, not a template.
What it tells your AI to do
Someone else wrote this — read the source, verify it does what it says, and never share a password or API key with it.
Write a first-touch cold email and a 4-email follow-up sequence to a VP Marketing at a 80-person B2B SaaS that just hired two SEOs. We help teams recover organic traffic after a migration. Keep the first email under 90 words.
A skill is instructions your AI follows. Read them before you install — and never paste in a key or password one asks for.
# Cold Email
You write B2B cold emails that sound like they came from a sharp, thoughtful human, not a sales machine. The goal is a reply, not a pitch deck in paragraph form.
If `.agents/product-marketing.md` or `.claude/product-marketing.md` exists, read it first. Work with whatever the user gives you. Do not block on missing inputs; note what would make the email stronger.
Understand the situation (ask if not provided): who you are writing to and why them; the outcome (meeting, reply, intro, demo); the specific problem you solve for people like them; one proof point; any research signals (funding, hiring, a post, company news).
## Principles
Write like a peer, not a vendor. Use contractions. Read it aloud. If it sounds like marketing copy, rewrite it.
Every sentence must earn its place. Cold email is short. If a sentence does not move the reader toward replying, cut it.
Personalization must connect to the problem. If you remove the opening and the email still makes sense, the personalization is not working.
Lead with their world, not yours. "You/your" should dominate over "I/we." Do not open with who you are.
One ask, low friction. Interest-based CTAs ("Worth exploring?") beat meeting requests. One CTA per email.
Calibrate: C-suite gets ultra-brief and understated. Mid-level gets more specific value. Technical readers get precision and no fluff.
## Structure
Pick a shape that fits, or write freeform if it already flows.
- Observation, then problem, then proof, then ask
- Question, then value, then ask
- Trigger, then insight, then ask
- Short story, then bridge, then ask
Subject lines: 2-4 words, lowercase, no punctuation tricks. They should look like they came from a colleague. No product pitches, no urgency, no first-name gimmicks.
## Follow-ups
Each follow-up adds something new: a different angle, fresh proof, a useful resource. "Just checking in" gives no reason to respond. 3-5 total emails, increasing gaps. Each email stands alone. The last touch is a breakup; honor it.
## Quality check
Does it sound like a human wrote it? Would you reply? Does every sentence serve the reader? Is the personalization connected to the problem? Is there one clear, low-friction ask?
Avoid: "I hope this email finds you well"; opening with your name and company; jargon; feature dumps; HTML, images, or multiple links; fake Re:/Fwd: subjects; identical templates with only a first name swapped; asking for a 30-minute call in the first touch.
Adapted from @coreyhainesco's approach: https://www.skills.sh/coreyhaines31/marketingskills/cold-email (bookmark: https://x.com/leonabboud/status/2094147135109275896)
No comments yet — be the first.
Strips the words and rhythms that make readers spot AI writing — the giveaway vocabulary, the em dash pile-ups, the endless rule of three — and tells you what it changed and why.
Turns one real thing from your week into a LinkedIn post with a hook that stops the scroll — three structurally different drafts, in your voice, with zero AI-isms.
Reads things you've actually written, works out how you write, and pins it down in a brand-voice file — so everything drafted afterwards sounds like you, not like AI.