agentleFS
buildwithreefbuildwithreef/claude-skills/skills/cold-email-personalizer/SKILL.md

cold-email-personalizer

Skill0 starsChanged 4 months agoMITView on GitHub

Researches a prospect and writes short, personalized cold emails using proven frameworks (observation → implication → question, PAS), plus a three-touch follow-up sequence. Use when the user wants to write a cold email or outreach message, personalize outreach to a specific business, or says "write a cold email to this company" or "help me reach out to this prospect".

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 8 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No links in the first email (spam filters and suspicion).

    line 34
  2. Distinctive:

    Read-aloud test: if it doesn't sound like something you'd say across a counter, rewrite it.

    line 35
  3. Distinctive:

    Never fake familiarity ("been following you for a while") and never open with an apology ("sorry to bother you").

    line 36
  4. Distinctive:

    No reply after three: stop.

    line 43
  5. Distinctive:

    Don't compliment-open ("Congrats on the new website!") — it's the most common spam tell in 2026.

    line 66
  6. Distinctive:

    Don't mention your company name before the signature.

    line 67
  7. Distinctive:

    Don't stack two CTAs or add a calendar link on touch one.

    line 68
  8. Distinctive:

    Don't personalize with trivia (their college, the weather).

    line 69

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Cold Email Personalizer

Personalization is not knowing their name. It's one observation about their business that's true, specific, and slightly uncomfortable to ignore — found in ten minutes, delivered in ninety words.

Step 1 — Research (10 minutes, hard cap)

Work through in order; stop when you have two strong observations:

  1. Their website: what they sell, to whom, what's obviously broken or missing (no pricing, dead blog, no mobile CTA, 2019 copyright).
  2. Their reviews (Google/Yelp): what customers praise (their real differentiator — often not what the site leads with) and what they complain about (a pain they feel weekly: "impossible to reach," "slow quotes").
  3. Growth signals: hiring pages, new location, running ads (check the site with an ad-library search), recent press.
  4. The owner: name findable? LinkedIn activity? Anything they've said publicly about the business?

An observation qualifies if the owner would read it and think "huh, they actually looked." "Love what you're doing!" fails; "your last 6 Google reviews all mention how hard it is to get a callback" passes.

Step 2 — Pick a framework (by what research found)

  • Observation → implication → question (default): "Noticed [true thing] → which usually means [cost/risk in their terms] → is that something you're feeling?"
  • PAS (when reviews handed you a bleeding pain): name the Problem in their customers' words, Agitate with what it costs, offer the Solve in one line.
  • Poke the bear (confident senders, competitive markets): a question that challenges the status quo — "Who answers your phone when the crew's on a roof?"
  • 1–2–3 (thin research): one observation, two sentences of concrete value, one question.

Step 3 — Write it under the rules

  • Under 90 words. Count them.
  • Subject: 2–4 words, lowercase, internal-memo energy — "missed calls", "your google reviews", "quick question re: quotes". Never clever, never salesy.
  • Line 1 = the observation. Never "I hope this finds you well," never your name or company (the signature has that).
  • One idea per email. One CTA, and it's an interest question — "Worth a look?" / "Open to seeing how that works?" — never "do you have 15 minutes Tuesday at 2?"
  • No links in the first email (spam filters and suspicion). No attachments. Plain text, no images, no HTML buttons.
  • Read-aloud test: if it doesn't sound like something you'd say across a counter, rewrite it.
  • Never fake familiarity ("been following you for a while") and never open with an apology ("sorry to bother you").

Step 4 — The sequence (3 touches, then out)

  • Day 1: the email.
  • Day 4 — the bump: reply to your own thread, one line only: "Any thoughts on this?" or a second, smaller observation. No re-pitch, no "just following up on my previous email."
  • Day 9 — the breakup: close the loop with a takeaway and zero guilt: "Guessing the timing's off — I'll leave it here. If [pain] ever makes the list, this is what I do." Breakups routinely out-reply the first email.
  • No reply after three: stop. Recycle the list in a quarter with a new observation.

Step 5 — Deliver

**Research notes:** [3–5 bullets: observations found + which was used]
**Framework:** [name + why]

**Subject:** [subject]
[email body]

**Day 4 bump:** [one line]
**Day 9 breakup:** [2–3 lines]

**Why this works:** [2 bullets tying lines to research]

Deliverability floor (say it once, briefly)

Sending domain warmed up, SPF/DKIM/DMARC set, under ~30–50 cold sends/day per inbox, and a list they're legally allowed to email (B2B norms vary by country — flag, don't lecture). The best copy in a spam folder converts at zero.

What not to do

  • Don't compliment-open ("Congrats on the new website!") — it's the most common spam tell in 2026.
  • Don't mention your company name before the signature.
  • Don't stack two CTAs or add a calendar link on touch one.
  • Don't personalize with trivia (their college, the weather). Business observations only.

Built by Reef — AI systems for small business → https://buildwithreef.com

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.