Cold Email Personalizer

Most "personalised" cold emails swap in a name and a company and hope for the best. Readers can tell, and the fake touches ("as we discussed", a "Re:" subject on a first email, a made-up result) do more harm than good.

This skill gives your AI a method for honest one-to-one outreach. It researches each prospect's recent, professional public information (from your notes or with its web tools), records every fact with a source link and date, and leaves out anything private. It then writes a first line about the prospect, a problem hypothesis phrased as a guess, a one-sentence offer with only the proof you supplied, a soft call to action, and two follow-ups that each add something useful. Every email stays under 120 words and has an opt-out line. When there is nothing worth mentioning, it writes an honest role-based opener and labels it.

Inside: SKILL.md, a personalization research checklist, writing rules with before/after examples, a spam-word and claim reference, compliance basics for the US, UK, EU, Canada and Australia, an output template, and a Node checker that flags long bodies, missing opt-outs, placeholders, fake replies, spam phrases, unproven claims, pushy asks and guilt-trip lines. The example covers three fictional prospects.

For founders, consultants and small sales teams. It drafts but never sends. The compliance notes are general information, not legal advice.