AI-powered marketing at scale — vibe marketing workflows for rapid content generation, creative iteration, campaign launches, and growth experiments using ChatGPT, Claude, Midjourney, Runway, HeyGen, ElevenLabs, and AI video tools. Based on the "vibe coding" philosophy applied to marketing. Use when scaling content production with AI, launching campaigns fast, generating creative variants, or building a marketing engine that runs on prompts. Triggers on: "vibe marketing", "AI marketing", "AI content", "scale content with AI", "AI campaign generation".
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
vibe-marketing
description
AI-powered marketing at scale — vibe marketing workflows for rapid content generation, creative iteration, campaign launches, and growth experiments using ChatGPT, Claude, Midjourney, Runway, HeyGen, ElevenLabs, and AI video tools. Based on the "vibe coding" philosophy applied to marketing. Use when scaling content production with AI, launching campaigns fast, generating creative variants, or building a marketing engine that runs on prompts. Triggers on: "vibe marketing", "AI marketing", "AI content", "scale content with AI", "AI campaign generation".
license
MIT
compatibility
Claude Code, Jesse, Codex, Hermes, Windsurf, OpenCode, Gemini CLI, Copilot, Zed, VS Code, Goose
metadata
{"version":"1.0.0","author":"LeadMagic","category":"creative","tags":["vibe-marketing","ai-marketing","ai-content","chatgpt","claude","midjourney","runway","heygen"],"related_skills":["vibe-coding","ai-content-creation","ai-video-creation","content-marketing","social-media-strategy","copywriting","ad-creative-strategy","growth-hacking-tactics"],"frameworks":["Andrej Karpathy — 'Vibe coding' applied to marketing: describe, generate, iterate","Gary Vaynerchuk — Content at scale: 64 pieces of content per day per platform","Dave Gerhardt (Exit Five) — B2B marketing that doesn't feel like marketing","Amanda Natividad (SparkToro) — Zero-click content, audience-first AI","Harry Dry (Marketing Examples) — Show, don't tell (visual proof)","Justin Welsh — Content operating system, repurposing lattice"]}
Vibe Marketing
Overview
"Vibe marketing" is vibe coding applied to marketing: describe what you want
to an AI, iterate conversationally, and ship campaigns in hours instead of
weeks. The old model: copywriter drafts → designer creates → manager reviews →
revisions → launch (2-4 weeks). The vibe marketing model: prompt → AI
generates 5 variants → pick 2 → polish → launch (2-4 hours). This skill covers
the complete AI marketing stack: content generation, creative production,
campaign ideation, and the "prompt-to-publish" workflow that lets solo
marketers operate at team scale.
Authoritative Foundations
Andrej Karpathy — 'Vibe coding' applied to marketing: describe, generate, iterate — 'Vibe coding' applied to marketing: describe, generate, iterate
Gary Vaynerchuk — Content at scale: 64 pieces of content per day per platform — Content at scale: 64 pieces of content per day per platform
Dave Gerhardt (Exit Five) — B2B marketing that doesn't feel like marketing — B2B marketing that doesn't feel like marketing
Amanda Natividad (SparkToro) — Zero-click content, audience-first AI — Zero-click content, audience-first AI
Phase 1: Research & Strategy (30 min — AI-assisted)
AI prompt chain for campaign strategy:
Prompt 1 (Research):
"You are a B2B SaaS marketing strategist. Research the top 5 competitors
for [my company] in the [space] market. For each: positioning, content
strategy, social presence, and what's working for them."
Prompt 2 (Strategy):
"Based on the competitive landscape, propose a 30-day content campaign
for [my company] targeting [ICP]. Include: content pillars, channel mix,
posting cadence, and key metrics."
Prompt 3 (Refinement):
"Narrow to the top 3 highest-ROI tactics. For each: specific execution
steps, AI tools to use, and expected outcomes."
Phase 2: Content Generation (1-2 hours — AI bulk generation)
The "prompt once, generate many" pattern:
Prompt:
"You are a B2B SaaS copywriter specializing in [industry]. Using the
following brand voice guide, generate a 30-day content calendar for
LinkedIn with:
- 5 long-form posts (educational frameworks)
- 5 short-form posts (contrarian takes)
- 5 engagement posts (questions, polls)
- 5 case-study style posts (customer wins)
- 5 carousel outlines (step-by-step guides)
- 5 video scripts (60-second tutorials)
Brand voice: [paste voice guide]
ICP: [describe target]
Key topics: [list 3-5 topics]"
This generates 30 pieces of raw content in one prompt. Of those 30:
10 will be great with minor edits
10 will be usable with significant edits
10 will be garbage
Your job: identify the 10 great ones. Polish them in 15 minutes each.
You've just created 2 weeks of content in 3 hours.
Phase 3: Visual Generation (1-2 hours — AI creative)
Social graphics pipeline:
1. Canva AI: "Generate a LinkedIn carousel about [topic]. 6 slides.
Style: [brand colors], [typography], clean B2B aesthetic."
2. Midjourney: "Professional B2B illustration for [topic]. Clean style,
[color palette], no text in image. --ar 16:9"
3. Remove.bg + Canva: Combine AI image with copy overlay
Ad creative pipeline (30 min per variant):
1. Claude: "Write 5 LinkedIn ad copy variants for [offer]. Each:
headline (under 150 chars), body (under 300 chars), CTA."
2. Midjourney: Generate 3 image variants per copy variant
3. Combine in Canva: image + copy overlay = ad creative
4. Launch all 5 variants. Kill underperformers after 48 hours.
Phase 4: Video Generation (2-4 hours — AI video)
Short-form video workflow:
1. Claude: "Write a 60-second script about [topic]. Hook in first 3 seconds.
One key point. CTA at end."
2. ElevenLabs: Generate voiceover from script (pick natural voice)
3. Runway/Pika: Generate B-roll footage for key moments
4. Descript/CapCut: Edit together — voiceover + B-roll + captions + branding
5. Opus Clip: Auto-generate 5 short clips from a longer video
AI spokesperson video workflow (HeyGen/Synthesia):
1. Claude: "Write a personalized demo video script for [customer name],
[title] at [company]. Address their specific [pain point]. 90 seconds."
2. HeyGen: Create AI avatar video. Upload script. Select avatar.
Personalize with customer name on-screen.
3. Send as personalized outreach — feels custom, costs $0.10 to produce.
Phase 5: Distribution & Testing (ongoing)
AI-powered distribution:
Prompt:
"You are a social media distribution strategist. Here is our content
calendar for the week. Generate:
1. 3 X/Twitter thread versions of each long-form post
2. 2 LinkedIn post variants (different hooks)
3. 5 comment templates for engaging with other accounts
4. 1 newsletter edition summarizing the week's insights"
AI-powered iteration:
After 1 week of posting:
Prompt:
"Here are the analytics from last week's content: [paste data].
Which posts performed best? Why? Generate next week's calendar
doubling down on what worked and cutting what didn't."
The "Prompt-to-Publish" Pipeline (End-to-End)
A complete vibe marketing campaign in 4 hours:
HOUR 1 — Strategy + Copy:
- Claude: Generate campaign strategy, ICP targeting, channel plan
- Claude: Write all copy — landing page, emails, social posts, ad copy
- Claude: Write nurture sequence for leads
HOUR 2 — Visuals:
- Midjourney/Canva AI: Generate hero images, ad creatives, social graphics
- Gamma: Generate pitch deck / one-pager for sales enablement
- v0/Lovable: Build landing page (or update existing)
HOUR 3 — Video (if needed):
- Claude: Write video scripts
- ElevenLabs: Generate voiceover
- HeyGen/Synthesia: Generate AI spokesperson video (demo, explainer)
- Runway: Generate B-roll
HOUR 4 — Launch & Automate:
- Schedule all content (Buffer, Hootsuite, or native schedulers)
- Set up email sequence (ConvertKit, HubSpot, Customer.io)
- Launch ads (LinkedIn, Meta, Google)
- Set up analytics and tracking
This is 2-4 weeks of traditional marketing work, compressed into 4 hours.
A team of 1 with AI operates like a team of 5 without.
CORE ASSET: 2,000-word blog post / framework / report
ATOMIZATION:
1. LinkedIn post (long-form) — extract the framework
2. LinkedIn post (short) — extract the contrarian take
3. LinkedIn carousel (6 slides) — step-by-step visual
4. X/Twitter thread (10 tweets) — break down the argument
5. Newsletter edition — deeper dive with personal story
6. YouTube Short / TikTok (60 sec) — say the one key insight
7. LinkedIn video (3 min) — walk through the framework
8. Podcast episode (15 min) — discuss with examples
9. Instagram/TikTok carousel — visual version
10. Reddit post — ask a question based on the insight
11. Quora answer — answer a question using the framework
12. Email to list — "I just published this"
13. Ad creative — turn the framework into a lead magnet ad
14. Sales one-pager — framework as leave-behind
15. Cold email hook — use the insight as an opener
16. Webinar topic — expand into a 30-min presentation
17. Guest post pitch — offer the framework to a publication
18. SlideShare / SpeakerDeck — upload the carousel as a deck
19. Infographic — visual summary (Canva AI)
20. Community post — share in Slack/Discord groups
AI can generate every one of these from the core asset in under 3 hours.
Expert Vibe Marketers to Follow
Expert
Platform
Vibe Marketing Specialty
Dave Gerhardt
@davegerhardt (X)
B2B marketing that feels human, not corporate
Amanda Natividad
@amandanat (X)
Zero-click content, audience-first strategy
Justin Welsh
@thejustinwelsh (X)
Content operating system, repurposing at scale
Harry Dry
@harrydry (X)
Visual proof marketing, Marketing Examples
Ross Simmonds
@TheCoolestCool (X)
Content distribution, repurposing, AI content
Blake Emal
@heyblake (X)
Copywriting frameworks, AI-assisted copy
Alex Garcia
@alexgarcia_atx (X)
Marketing experiments, AI tools
Aadit Sheth
@aaditsh (X)
AI content workflows, prompt engineering
Kieran Flanagan
@searchbrat (X)
AI + marketing (Zapier CMO)
Output Format
VIBE MARKETING SPRINT — [Week/Campaign]
GOAL: [specific outcome: signups, demos, revenue]
ICP: [target audience]
TIMEFRAME: [1-week sprint]
DAY 1 — Strategy & Copy:
- AI tools used: [Claude / ChatGPT]
- Assets created: [landing page copy, 10 social posts, 3 email drafts]
DAY 2 — Visuals:
- AI tools: [Midjourney / Canva AI / Gamma]
- Assets: [hero image, 10 social graphics, pitch deck]
DAY 3 — Video:
- AI tools: [HeyGen / Runway / ElevenLabs / Descript]
- Assets: [1 explainer video, 3 short clips, voiceover]
DAY 4 — Launch:
- Schedule: [all content scheduled, emails queued, ads live]
- Analytics: [tracking URLs, UTM parameters, dashboard]
DAY 5 — Review & Iterate:
- Top performers: [which content won]
- Next sprint: [what to double down on]
Implementation Checklist
AI-generated copy reviewed by human (AI writes, human edits)
AI-generated images checked for brand consistency
All claims fact-checked (AI hallucinates statistics)
Video content reviewed for accuracy before publishing
No AI "slop" — every piece earns its place or is cut
UTM parameters on all links for attribution
A/B test variants generated (at least 2 per ad/post)
Analytics dashboard live before launch
Quality Check
Before delivering, verify:
Output matches the user's stated request
Named frameworks or sources are reflected in the recommendation
The deliverable is specific enough for an agent to execute
Any assumptions, risks, or dependencies are explicit
No unsupported claims, invented facts, or private/internal references are included
Common Pitfalls
AI slop at scale. "30 AI blog posts in 1 hour!" = 30 pieces of
unreadable garbage. Fix: AI writes first drafts. Humans edit for voice,
accuracy, and insight. 10 great posts > 100 AI-slop posts.
No fact-checking. AI confidently hallucinates statistics, customer
quotes, and case studies. Fix: Every stat needs a source. Every quote
needs attribution. Every claim needs verification.
Brand voice dilution. "Write a LinkedIn post" without brand voice
context = generic AI voice. Fix: Paste your brand voice guide into every
content prompt. Better yet: save it as a custom GPT.
One-and-done prompting. First AI output is never the best. Fix:
"That's good. Now write 4 more variants with different hooks." Pick
the best. Iterate. AI gets better with direction.
AI-only content with no human insight. AI can summarize what's
already been said. Only humans can say something genuinely new. Fix:
Your unique data, experiences, and opinions are the moat. AI is the
amplifier, not the source.
Ignoring distribution. AI makes creation 10x faster. Distribution
is still the bottleneck. Fix: Spend 20% of time creating, 80%
distributing. AI can help with distribution copy too.
Execution Artifacts
references/framework-notes.md — Named frameworks and reference tables
templates/output-template.md — Deliverable shell for agent output
