agentleFS
megandmartinmegandmartin/agent-skills-repo/skills/content-marketing/email-nurture-sequence/SKILL.md

email-nurture-sequence

Skill4 starsChanged 4 months agoMITView on GitHub

Drafts a 5-email nurture sequence where each email has exactly ONE goal, every subject line ships as an A/B pair, styling is plain-text-first, and unsubscribe intent is respected — drafts only, a human loads and sends. Use when the user asks for a "nurture sequence", "welcome series", "email drip", "onboarding emails", or "5-email campaign". Don't use for a single standalone newsletter issue — that's newsletter-assembler.

The file

Files
SKILL.md

Email Nurture Sequence

Produces a 5-email sequence that earns its place in an inbox: one job per email, subject lines tested not guessed, formatting that reads like a person wrote it, and a hard rule that leaving must always be easy. This skill outputs drafts and a load plan — it never connects to an ESP and never sends.

When to Use

  • The user needs a welcome, onboarding, lead-magnet follow-up, or pre-launch sequence.
  • An existing drip underperforms and needs a one-goal-per-email restructure.
  • The user asks for "subject line A/B tests" for a sequence.
  • Not for: a single recurring newsletter issue (newsletter-assembler), the page the sequence points to (landing-page-copywriter), or actually sending — no ESP access, drafts only, human loads and presses send.

Quick Reference

ActionRule
One goal per emailEach email drives exactly one click/reply/action; a second ask means a sixth email, not a longer fifth
Subject pairsA/B per email: A = clarity, B = curiosity; both ≤50 chars, both honest about contents
Plain-text-firstWrite as if no HTML exists: short lines, no image-dependent meaning, ≤2 links, one formatting device max
Sequence arc1 deliver+welcome → 2 quick win → 3 story/proof → 4 objection → 5 direct ask
Unsubscribe respectFooter slot in every draft; no guilt copy; opt-down offered before the ask email
SendingNEVER. Deliver drafts + load checklist; human owns the ESP

Procedure

  1. Gather — confirm: who enters the sequence and from where (lead magnet, signup, purchase), the sequence's end goal (the one conversion email 5 asks for), sending cadence (default: days 0, 2, 4, 7, 10), sender name/voice, and links available. End goal missing → stop and ask; a sequence without a destination is just noise.
  2. Map one goal per email — write a 5-row plan first: email #, single goal, the one link/action, arc role (per Quick Reference). If the user's wishlist has 7 goals, cut or defer 2 — do not double-load an email. Get a nod on the map when the stakes are high; otherwise proceed.
  3. Draft in arc order — each email: subject pair, preview text (≤80 chars, extends A not repeats it), body in plain-text style — greeting, 60–150 words that earn the one ask, single CTA as a text link, human sign-off. Emails 1–2 give before anything asks; email 4 names the biggest objection honestly; email 5 asks directly, once.
  4. Subject A/B discipline — for each email: A states the value plainly ("Your template is inside"), B opens a loop ("The mistake I made first"). Both must be true to the body — curiosity that the email doesn't pay off trains people to ignore you.
  5. Respect-the-exit pass — every draft ends with an [UNSUBSCRIBE] placeholder line; email 4 or 5 includes a one-line opt-down ("just want the monthly digest? click here" placeholder). Strip any guilt phrasing ("we'll be sad to see you go" is fine; "you'll miss out forever" is not).
  6. Deliver drafts + load plan — assemble the Output Template with the schedule table and an ESP load checklist. State explicitly: these are drafts; nothing has been sent.

Output Template

# Nurture Sequence: [name] — 5 emails, drafts only
**Entry:** [trigger] | **End goal:** [email 5's ask] | **Cadence:** D0/D2/D4/D7/D10

| # | Day | Goal (one) | Arc role |
|---|---|---|---|

## Email [n] — [goal]
Subject A: ... | Subject B: ...
Preview: ...
---
[body, plain-text style, one CTA link]
—[sender]
[UNSUBSCRIBE] · [opt-down, emails 4–5 only]

## Load checklist (human)
- [ ] Paste into ESP, map merge fields, verify unsubscribe link renders
- [ ] Set A/B split on subjects, schedule per cadence table
- [ ] Send test to self before activating

Pitfalls

  • Two asks in one email — a "quick win" email that also pitches the offer, so neither lands. Recovery: return to the step 2 map; move the second ask to its rightful email or cut it. One email, one goal, no exceptions.
  • Bait subject, mismatched body — subject B promises a story the email doesn't tell. Recovery: in step 4 read each body asking "does this pay off both subjects?"; rewrite whichever side lies.
  • Newsletter-shaped nurture — emails balloon to 400 words of sections and links. Recovery: cut to the 60–150 word band; if material genuinely deserves more, it's a link to a post, not inline copy.
  • Guilt-trip exit copy — retention dark patterns creep into the footer. Recovery: run the step 5 pass; unsubscribe is one neutral line, opt-down is offered once, nothing shames the reader.
  • The agent "helpfully" sends — drafts get loaded or dispatched because ESP access happened to exist. Recovery: this skill has no send step by design; if a send was made, report it immediately to the human with the audience affected — never quietly continue.

Verification

  • Exactly 5 emails; each maps to one goal on the step 2 table; no email carries two asks
  • Every email has subject A + B (≤50 chars each) and preview text; both subjects truthful to the body
  • Bodies 60–150 words, ≤2 links, single CTA, plain-text-first formatting
  • [UNSUBSCRIBE] in all 5; opt-down present in email 4 or 5; zero guilt phrasing
  • Deliverable states "drafts only — human loads and sends"; nothing was sent

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.