What does this skill do, and when should you use it?
Email Creative Builder produces a complete single-email creative: subject-line variants, preheader, body copy, one primary CTA, and a plain-text alternate. It supports B2C promotional or lifecycle email, B2B cold outbound, and newsletter modes. The skill anchors the message to the destination page and traces claim-bearing copy to approved narrative and claims records. It does not pre-score subject lines, calculate EQS, design sequences, or create A/B test plans.
Reads the offer or topic, destination URL, email mode, audience or lifecycle stage, existing copy, narrative and claims projections, and relevant consent or suppression state; reads the destination page's headline, value proposition, offer or claim, and CTA; drafts 3–5 labeled subject lines, a matched preheader, structured body copy, one primary CTA, and a plain-text alternate; annotates claim-bearing lines with message-match notes and preserves [needs source] flags for unresolved claims; with permission, writes a WARM artifact or saves the creative to the documented memory path.
- A marketer needs a B2C promotional or lifecycle email tied to an existing landing page.
- A B2B sales team needs a low-pressure cold-outbound email personalized from a named signal.
- A newsletter editor needs value-led editorial copy with a distinct sponsor or monetization slot.
- An email team wants to replace losing subject lines while preserving message-match to the destination page.
How do you install this skill?
- Before use, verify access to the destination URL, claims projection, consent/suppression state, and any required connector; do not mark the output ready to send otherwise.
- The cold-outbound CAN-SPAM, opt-out, and consent guidance does not replace legal review for the target jurisdictions.
- Resolve the governance mismatch between skill version 18.0.0 and the security policy's supported 17.x line, and add key-path tests for this skill.
- Shell / CLI
- Network access
- Local filesystem
The skill is located at email/engage/email-creative-builder/SKILL.md in the repository. To install the full collection in Claude Code, run: /plugin marketplace add aaron-he-zhu/aaron-marketing-skills, then /plugin install aaron-marketing@aaron. For a generic Agent Skills-compatible host, run: npx skills add aaron-he-zhu/aaron-marketing-skills. The supplied material does not document a separate single-skill install command for this specific skill.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/aaron-he-zhu/aaron-marketing-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/email/engage/email-creative-builder" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Supply the offer or topic, destination landing-page URL, mode, and audience details. Example: Write 5 subject lines + preheader + body + CTA for [offer], destination [URL], promo mode. Another supported prompt is: Draft a cold-outbound email to [persona] for [offer]; personalize on [signal]; destination [URL]. If the URL is missing, the skill should return NEEDS_INPUT because message-match cannot be enforced.
What are this skill's strengths and limitations?
- Produces a complete, clearly structured creative unit with variants, preheader, body, one CTA, and plain text.
- Supports promo/lifecycle, cold-outbound, and newsletter patterns.
- Traces claims to approved records and the destination page, while prohibiting invented statistics, prices, guarantees, discounts, or testimonials.
- Works with Claude Code and compatible Agent Skills hosts; the repository is Apache-2.0 licensed.
- Does not pre-score subject lines for spam triggers, truncation, length, emoji count, or inbox-preview rendering.
- Does not score SEND dimensions, calculate EQS, or run the email auditor's vetoes.
- Missing approved claims can leave the draft exploratory rather than ready to send; material unresolved claims block ready-to-send status.
- The supplied material does not document a dedicated test suite or validation on specific ESP platforms for this skill.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Email Creative Builder this page | 58 · Recommended | ★ 3k | 12d ago | Apache-2.0 |
| Subject Line Lab | 56 · Use with care | ★ 3k | 12d ago | Apache-2.0 |
| Ad Creative Builder | 60 · Recommended | ★ 3k | 12d ago | Apache-2.0 |
| Email Marketing Bible | 60 · Recommended | ★ 342 | 14d ago | MIT |
| Proof Point Packager | 60 · Recommended | ★ 3k | 12d ago | Apache-2.0 |
Unlike subject-line-lab, this skill drafts subject variants but does not pre-score or rank them. Unlike email-quality-auditor, it does not score the email, calculate EQS, or run vetoes. Unlike email-sequence-designer, it builds one email rather than a multi-step flow. Unlike send-experiment-designer, it does not design A/B or send-time tests.
How did FollowSkills review this skill?
The skill treats exports, web content, and personalization signals as untrusted, forbids credential storage and silent ledger edits, and requires confirmation before saving; however, it still reads destination URLs, CRM/consent state, and optional connectors, while complete data-flow disclosure, rollback, and runtime isolation are not established within this skill, so 7 points are deducted.
Input confirmation, NEEDS_INPUT handling for missing URLs, mode routing, blocking conditions, and quality gates are internally coherent on the happy path; however, there are no skill-specific key-path tests, complete diagnostics for unavailable web/registry dependencies, or executable reproduction evidence, so the static score remains below the cap.
The audience, single-email use case, three operating modes, and non-fit boundaries for scoring, sequencing, and experimentation are clearly stated; however, semantic triggering still depends heavily on natural-language context, Chinese-language output capability is not explicit, and the legal guidance is jurisdiction-limited, so 5 points are deducted.
Front matter, version, Apache-2.0 license, quick starts, input/output contract, handoff fields, limitations, and next-skill routing are well documented; however, maintenance ownership, changelog, and update path are unclear, and SECURITY.md supports 17.x while the skill declares 18.0.0, indicating governance inconsistency, so 5 points are deducted.
The expected deliverable is concrete and includes subject variants, preheader, body, one primary CTA, plain-text alternative, claim-level message matching, and source flags, making the core task plausibly directly usable; however, static files do not verify output quality or marginal benefit over manual drafting, and readiness depends on live destination, approved claims, and runtime state, so the static maximum of 7 is used.
The skill, its two supplied references, and repository-level CI/tests provide auditable specification evidence and distinguish approved, fallback, and blocked states; however, there are no dedicated tests for this skill, third-party execution results, or independent corroboration, so the static maximum of 5 is used.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →