Cold Email Checker

Before: Draft opening: "Subject: Re: quick question about your growth ... I saw you just launched Hemingway Editor Plus last week and it looks amazing. I've helped 40 software companies double their signups ..."

After (skill output, from a real test run): Test run on a draft we wrote with common flaws, against text from a real public software homepage. Verdict, unedited: "Do not send. The email claims results you can't prove, says the prospect launched something last week when their page doesn't say so, and makes four asks." It marked the launch claim Unsupported, the results claim Needs proof, flagged "Re:" on a first message, listed the missing postal address, opt-out and ad disclosure, and returned a tightened version with [CONFIRM] markers.

Give it a cold email draft and the text of the prospect's page. It tests every claim about the prospect against that text, flags claims about you that need proof, checks the subject line, the single ask and the length, and lists the commercial email basics from the US FTC's CAN-SPAM guide that the draft is missing. You get a verdict, a claims table, quoted issues and a tightened version with unsupported claims removed.

1 run on a draft we wrote with common flaws and text from a real public homepage (Claude Sonnet).

1 run on a flawed draft we wrote and text from a real public homepage (Claude Sonnet). The draft is ours, the page text is real. A second, harder run (2026-10-04, Claude Sonnet, an email with an invented award and a site line suggesting the reader is not a dentist): it flagged every unsupported claim and the missing commercial-email basics. No defects found.

Tested on a real example before release. The example is shown in the Before and after section above.

How we test, what we do not do, and which assistants we ran it on.

• It checks claims against the text you give it. Without the page text it marks them "not checked".
• The legal list comes from one US guide and is not legal advice. It does not say whether a law applies to you. Rules in other countries differ.
• It does not find addresses, check deliverability or send anything.
• It cannot know whether your proof is true. You confirm it.

Works with any assistant that reads SKILL.md skills (Claude, Codex, Gemini CLI, Cursor and others), plus a paste-in prompt version for ChatGPT or any chatbot. Gets a draft and page text. Checking against a page needs a web tool or pasted text.

1. Run npx skills add proskillpacks/skills in your project and pick this skill. Or copy the cold-email-checker folder from GitHub (https://github.com/proskillpacks/skills/tree/main/skills/sales/cold-email-checker).
2. Put the cold-email-checker folder where your assistant looks for skills. The generic path is .agents/skills/ (or npx skills add from a folder). Tool-specific paths:
   • Claude Code: ~/.claude/skills/ (or .claude/skills/ inside one project)
   • OpenAI Codex, Gemini CLI, Cursor and others: .agents/skills/ or the skills folder their docs name
   • claude.ai: zip the skill folder yourself, then Customize, Skills, the plus button, Create skill, Upload a skill (code execution must be on in Settings, Capabilities)
   • ChatGPT or any chatbot: open the skill's prompt.md, paste it into the chat, then paste your own text or link
   • Steps for each tool, with links to its official docs: how to install Agent Skills
3. Ask for the job in plain words, for example: "Check my cold email draft before I send it. Here is the draft and the prospect's page text.". The skill loads by itself.