| name | email-personalizer |
| description | Writes one personalized cold email from a completed account research brief. References a specific trigger from the last 90 days in the first two sentences. Under 120 words. Ends with a low-friction question CTA. Returns the draft for human approval — never sends. Flags banned words before output. |
| allowed-tools | Read, Write |
| effort | medium |
Email Personalizer
When to activate
After completing account research and confirming GO or CAUTION ICP score. Use this to generate the first-touch email in a sequence. Also use to rewrite a failed email after a human requests changes.
When NOT to use
Do not use before running account research — there must be a research brief with a specific trigger to reference. Do not use for follow-up emails in an existing sequence (use follow-up-scheduler instead). Do not use if the prospect is on an opt-out list.
Instructions
- Read the account research brief for this prospect. Identify the strongest trigger from the last 90 days (funding, exec hire, product launch, press mention, job posting signal).
- Identify the prospect's likely pain point based on their role, company stage, and tech stack.
- Write the email following these rules:
- Subject line: Specific to the trigger or their current context. Not "Quick question" or "Re: [Company]". Max 7 words.
- First sentence: States a business observation tied to the trigger. Never "My name is" or "I hope you're well."
- Trigger reference: Within first two sentences.
- Value prop: One sentence. Result-focused, not feature-focused. With proof if available.
- CTA: A specific low-friction question. "Worth a 20-minute call Thursday?" not "Let me know if you'd like to connect."
- Total length: Under 120 words.
- Tone: Direct, human, not salesy.
- Check output against banned words list in CLAUDE.md. Flag any violations before returning.
- Return draft with metadata header (see Output Format). Mark status as PENDING APPROVAL.
Output Format
EMAIL DRAFT — PENDING APPROVAL
Prospect: [Name, Title, Company]
ICP Score: [X]/100
Trigger: [specific trigger referenced]
Word Count: [X]/120
---
Subject: [subject line]
[Email body]
[Sender name]
[Title]
[Company]
---
Banned word check: PASS / FAIL ([flagged words if any])
Next: Human review required before send.
Example
Input: Alex Kim, CTO, Stackline (Series A, 85 employees, launched new API product 6 weeks ago, uses AWS + GitHub + Datadog + PagerDuty)
EMAIL DRAFT — PENDING APPROVAL
Prospect: Alex Kim, CTO, Stackline
ICP Score: 85/100 — GO
Trigger: New API product launch (6 weeks ago)
Word Count: 97/120
---
Subject: Stackline's API launch + developer onboarding
Six weeks after launching your new API, the #1 support volume spike is usually authentication edge cases and rate limit confusion.
We help API-first teams cut developer onboarding time by 40% with interactive docs and sandbox environments — Stripe and Twilio both used this pattern post-launch.
Worth a 20-minute call this week to see if it fits where you are?
[Sender name]
[Title]
[Company]
---
Banned word check: PASS
Next: Human review required before send.
