End-to-end marketing campaign planning and execution — audience research, positioning, campaign angle, landing page copy, email sequences, social posts, ad copy, video scripts, and content calendars. Use when orchestrating a multi-channel product launch from one brief. Do NOT use for a single asset (content-engine for a post, article-writing for a piece, crosspost for distribution only) or for the underlying market study (market-research).
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Um comando direto ignora o prompt de revisão. Verifique a origem antes de executá-lo.
Instruções da origem · Visualização somente leitura
name
marketing-campaign
description
End-to-end marketing campaign planning and execution — audience research, positioning, campaign angle, landing page copy, email sequences, social posts, ad copy, video scripts, and content calendars. Use when orchestrating a multi-channel product launch from one brief. Do NOT use for a single asset (content-engine for a post, article-writing for a piece, crosspost for distribution only) or for the underlying market study (market-research).
origin
ECC
Marketing Campaign
Plan and execute launch campaigns that convert — not just campaigns that ship.
When to use
planning a product, collection, or feature launch
building a full content suite from a single product brief
defining positioning and campaign angle before any copy is written
orchestrating multiple content types across channels
reviewing an existing copy set for conversion quality and brand consistency
When NOT to use:
the ask is one post, one thread, or one article — use content-engine / article-writing
the ask is publishing already-written copy — use crosspost
positioning is not settled and the user will not settle it. Copy written before the angle is
approved gets thrown away; that is the most expensive failure in this skill
Inputs
Required before Phase 3 — absent input = stop, never assume:
Product truth. For SkyyRose: SKU, name, price, and availability come from
wordpress-theme/skyyrose-flagship/data/skyyrose-catalog.csv; imagery resolves only via
/ (filenames are not identity); collection
language only from . Never a product fact from memory.
data/sot-images.json
skyyrose.core.sot_images
docs/brand/collection-stories.md
Audience research — from market-research: jobs-to-be-done, fears, language, alternatives.
Approved positioning and angle (Phase 2 output). No approval → no copy.
Voice profile — from brand-voice, so every channel sounds like one author.
A real deadline if urgency will be used. Fake urgency is banned; without a real date the
urgency beat is cut, not invented.
Procedure
Phase 1 — Research
Use market-research to profile the audience (jobs-to-be-done, fears, language, alternatives),
map 3+ direct or adjacent competitors (positioning, gaps, messaging weaknesses), and identify
1–3 insights the angle will exploit.
Deliverable: a research brief — audience profile + competitive summary + key insights.
Phase 2 — Positioning
Produce: core benefit statement (one sentence, no feature list, no jargon) · positioning formula
[Product] helps [audience] [achieve outcome] by [mechanism] · campaign angle (the specific
tension, insight, or moment the campaign lives in) · tone profile (delegate to brand-voice).
Do not write any copy until positioning and angle are approved.
Phase 3 — Content production
In this order — each layer informs the next:
Landing page copy — hero, problem, solution, features, how it works, proof, CTA
Email sequence — one purpose per email; arc: problem → education → agitation → solution →
proof → urgency → final CTA
Social posts — platform-native via content-engine (LinkedIn and X are different formats,
not the same copy resized)
Short-form video scripts — timestamp-blocked, written for screen and ear
Ad copy variants — 3–4 testing different angles or audience segments
Write each deliverable to its own file so the Verification checks have real targets.
Phase 4 — Review, then hand off
Gate every deliverable: 5-second test on all hero / above-fold copy · CTA audit (one per piece,
specific, earned) · tone consistency across channels · claim audit (every claim specific and
supportable) · cross-channel consistency (ad claims match landing page; email body matches
subject). Anything paid, published, or written to production goes through the owning skill's
STOP-AND-SHOW gate — this skill produces copy, it does not spend money or publish.
Output contract
Positioning brief — angle, core benefit statement, tone profile
Landing page copy — all sections
Email sequence — subject + preview + body + CTA per email, labelled by day and purpose
PASS: the row exists and its name/price match every mention in the campaign copy. No row
→ the product claim is deleted, not approximated. [repo]
Prove check 1 can fail before trusting it (rule 3): copy one hero line verbatim into an email,
confirm uniq -d returns it, restore. Observed in the worked example.
A SKIP is not a PASS. The 5-second test and tone-consistency read are human judgments — no
command grades them. State explicitly that they are open and name the reviewer who closes them,
rather than letting the mechanical greens imply the whole gate passed.
Worked example
Real run, 2026-07-28, a Black Rose one-run drop campaign. Two deliverables written to the session
scratchpad: landing-hero.md ("Armor for the people the concrete raised. / Black Rose. One run.
No restock.") and email-01.md ("You stood up first. The coat is the receipt. / Black Rose goes
live tonight at 7pm PT."). The hero line is derived from Black Rose canon in
docs/brand/collection-stories.md[repo].
Zero verbatim overlap [repro] — same angle, different sentences, which is the contract.
The number-sourcing gate was proven red first in the same run, on a sibling research file:
$ grep -nE '[0-9]+(\.[0-9]+)?%|\$[0-9]' research-brief.md | grep -v 'Source:'
3:- Average luxury-streetwear AOV is $240
After labelling that line (Source: operator-supplied; unverified), the same command exits 1
[repro]. Open items from this run: the 5-second test and tone read on the hero are unclosed and
owned by the requester — mechanical greens do not close them.
