Compare commits

...
5 Commits
Author SHA1 Message Date
Corey Haines ba3c85c92a Merge pull request #426 from coreyhaines31/development
release: v2.8.5 — motion video ads, brand-flexible styles, carousel frameworks, iOS reveal surfaces
2026-07-09 17:08:57 -07:00
Corey HainesandClaude Fable 5 2b45f16275 feat(ad-creative): ChatGPT + Apple Notes reveal surfaces — iOS-native reveal family (2.5.0) (#425)
* feat(ad-creative): extend iMessage reveal reference to ChatGPT + Apple Notes surfaces (2.5.0)

Adds an Other iOS-Native Reveal Surfaces section: surface selection by
persuasion mechanic, ChatGPT reveal craft + strict compliance rules for
fabricated AI answers, and Apple Notes confession-format guidance.
Repo release 2.8.5. Closes #424.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs(ad-creative): repeat trade-dress note in Apple Notes subsection (codex suggestion)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-07-09 17:07:31 -07:00
Corey HainesandClaude Fable 5 3826a38f88 feat(social): carousel framework library — 5 slide-by-slide narrative architectures (2.2.0) (#423)
* feat(social): carousel framework library — 5 slide-by-slide narrative architectures (2.2.0)

Re-expresses and extends the five-framework carousel breakdown
(Value-Stack, Problem-Proof, Hack List, Rant Callout, Demo Walkthrough)
as references/carousel-frameworks.md with selection guidance, SaaS
examples, failure modes, platform notes, and measurement guidance.
Repo release 2.8.4. Closes #421.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(social): address codex review — description length, slide-range math, metrics qualifier

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(social): generalize Rant Callout table to cover 4-8 range; sync changelog trigger list

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-07-09 17:06:22 -07:00
Corey Haines 44eb148f67 feat(ad-creative): brand-flexible motion video style tier (2.4.0) (#422)
Four token-driven styles (monoline editorial, Swiss typographic, wireglow,
duotone screenprint) driven by a brand-slots contract (FIELD/INK/ACCENT/
TYPE FEEL + per-brand constraints) so any company's brand system drops in.
Per-style motion notes. All four verified with a real brand's tokens.
Repo release 2.8.3.
2026-07-09 14:53:42 -07:00
Corey Haines d95efdf556 feat(ad-creative): faceless motion-style video ad format (2.3.0) (#420)
* feat(ad-creative): faceless motion-style video ad format (2.3.0)

New references/motion-video-ads.md: fully generated 15-45s concept videos
(styled poster stills -> image-to-video living motion -> TTS -> word-timed
captions, ~$3-6 per video). Provider-agnostic pipeline with one-key Gemini
path, five-style visual library with fill-in prompt formulas, motion prompt
formula, and production QC gotchas earned shipping the format end-to-end.
Format credit: @borjafat / Bomx super-video-maker (re-expressed, extended).

Repo release 2.8.2.

* fix: preserve json formatting in version bump (no unicode escapes, keep newline)
2026-07-09 13:02:45 -07:00
11 changed files with 370 additions and 12 deletions
+1 -1
View File
@@ -6,7 +6,7 @@
},
"metadata": {
"description": "Marketing skills for AI agents — conversion optimization, copywriting, SEO, paid ads, and growth",
"version": "2.8.1",
"version": "2.8.5",
"repository": "https://github.com/coreyhaines31/marketingskills"
},
"plugins": [
+1 -1
View File
@@ -1,7 +1,7 @@
{
"name": "marketing-skills",
"description": "Marketing skills for AI agents — conversion optimization, copywriting, SEO, paid ads, ad creative, and growth",
"version": "2.8.1",
"version": "2.8.5",
"author": {
"name": "Corey Haines"
},
+18 -2
View File
@@ -5,7 +5,7 @@ Current versions of all skills. Agents can compare against local versions to che
| Skill | Version | Last Updated |
|-------|---------|--------------|
| ab-testing | 2.0.0 | 2026-05-05 |
| ad-creative | 2.2.0 | 2026-07-08 |
| ad-creative | 2.5.0 | 2026-07-09 |
| ai-seo | 2.1.0 | 2026-06-15 |
| analytics | 2.0.0 | 2026-05-05 |
| aso | 2.0.0 | 2026-05-05 |
@@ -49,11 +49,27 @@ Current versions of all skills. Agents can compare against local versions to che
| signup | 2.0.0 | 2026-05-05 |
| site-architecture | 2.0.0 | 2026-05-05 |
| sms | 1.0.0 | 2026-05-21 |
| social | 2.1.0 | 2026-06-10 |
| social | 2.2.0 | 2026-07-09 |
| video | 2.0.1 | 2026-05-18 |
## Recent Changes
### 2.8.5 (2026-07-09)
- **ad-creative** (2.4.0 → 2.5.0): extended the iMessage reveal reference into **iOS-native reveal video ads** — `references/imessage-video-ads.md` (filename unchanged; retitled) now covers three surfaces sharing one production architecture. New **Other iOS-Native Reveal Surfaces** section: a surface-selection table keyed to persuasion mechanic (iMessage = a friend's recommendation; ChatGPT = an authoritative answer to the viewer's own question; Apple Notes = a private confession made public), **ChatGPT reveals** (the typed question is the customer's verbatim question; stream the answer in word chunks — type like thumbs, stream like a model; ChatGPT-real formatting; OpenAI trade-dress legal note; and the strictest compliance rule in the family — the answer is ad copy wearing an authority costume, so health/medical/financial advice in a fabricated AI answer requires legal review and the exchange is never presented as a real unprompted ChatGPT endorsement), and **Apple Notes reveals** (confession-not-conversation: note title as the hook, first-person list typed live with keyboard-taps-only audio, one typo-and-correction, the product as the least enthusiastic line, realizations grounded in real reviews per Grounded Inputs). SKILL.md pointer updated to the three-surface framing and adds 'ChatGPT ad' and 'Apple Notes ad' triggers. New eval (id 8) covers the health-claims compliance flag, surface selection, and Notes craft rules. Format source credit unchanged (Gooseworks iOS Ads pack). Closes #424.
### 2.8.4 (2026-07-09)
- **social** (2.1.0 → 2.2.0): added a **carousel framework library**. New `references/carousel-frameworks.md` — five slide-by-slide narrative architectures for Instagram carousels and LinkedIn document posts, re-expressed and extended from a pattern circulating on marketing X (Vibe Marketers HQ's five-framework breakdown): **Value-Stack** (exact-count cover → one item per slide → action close; the count is a checkable promise), **Problem-Proof** (result-as-fact hook → reframed problem → named mechanism → literal detail → screenshot receipt that closes slide 1's loop), **Hack List** (contrarian stat hook → why the common approach fails → one *named* technique per slide → thesis + CTA), **Rant Callout** (provocative claim → specific escalation → the fairness pivot that keeps it conviction-not-bitterness → signed close), and **Demo Walkthrough** (outcome first → pain → numbered process overview *before* the detail to cut mid-swipe drop-off → one real screenshot per step → result + positioning). Includes a framework-selection table (match structure to content type), two cross-framework rules (slide 1 is the thumbnail; one visual template per carousel), per-framework SaaS examples and failure modes, platform notes (IG 4:5 specs, LinkedIn PDF document posts with the post text as a second hook, TikTok photo-mode pacing), a production checklist (real claims only, exact counts delivered, one CTA), and measurement guidance (judge on saves + completion, not likes). SKILL.md wires the reference next to post templates and adds 'carousel,' 'slide-by-slide,' and 'document post' triggers; post-templates.md's generic Carousel Hook now points to the full library. New eval (id 7) covers framework selection, named-technique slides, LinkedIn document-post guidance, and no-invented-stats. Closes #421.
### 2.8.3 (2026-07-09)
- **ad-creative** (2.3.0 → 2.4.0): added a **brand-flexible style tier** to `references/motion-video-ads.md` — four token-driven styles that absorb any company's brand system instead of imposing their own palette: **monoline editorial** (thin single-weight ink line-art on the brand neutral, one accent sweep — the most universally brandable), **Swiss typographic** (the label set enormous in the brand's type on a grid; works for any brand with a font and a color), **wireglow** (wireframe subject on near-black with one gradient accent beam — dev-tool/dark-mode brands), and **duotone screenprint** (photography reduced to ink + accent). All driven by a **brand slots contract** — FIELD (neutral ground), INK (drawing/type color), ACCENT (one element per frame, scarcity is the design), TYPE FEEL, plus per-brand constraints (e.g. "gradients only on edges") — resolvable from brand guidelines or `.agents/product-marketing.md`. Includes per-style motion notes (lines draw themselves, beams pulse, type settles) under the same composition-never-changes rules. All four verified to render with clean typography using a real brand's tokens.
### 2.8.2 (2026-07-09)
- **ad-creative** (2.2.0 → 2.3.0): added the **faceless motion-style video ad format**. New `references/motion-video-ads.md` (format popularized by @borjafat; original re-expression of the open Bomx `super-video-maker` motion-collage recipe, extended with production lessons from shipping it end-to-end): a fully generated 15–45s concept/explainer video — styled poster stills (one per beat, series-locked via reference images) → image-to-video "living" motion → TTS narration → whisper-word-timed captions — at roughly $3–6 and ~15 minutes per finished video. Covers the provider-agnostic pipeline with a one-key Gemini path (Nano Banana Pro + Veo 3.1 + Gemini TTS) and alternatives (GPT-Image/Flux, Seedance/Kling/Runway, ElevenLabs), a **five-style visual library** with fill-in prompt formulas (screen-print collage, flat vector explainer, papercraft diorama, pop-art comic, claymation), a motion prompt formula, hard-earned QC gotchas (image-to-video models inject photoreal "maker hands" — and negative prompts make it worse; always QC each clip's final 2 seconds; caption/label collision; TTS/whisper sound-alike words in CTAs), ad-specific rules (brand in the poster for sound-off autoplay, beat 1 is the 3-second hook, regenerate stills per aspect rather than cropping), and AI-content disclosure notes. SKILL.md wires the reference into Generating Ad Visuals and adds 'motion video ad,' 'faceless video ad,' 'animated explainer ad,' and 'motion collage ad' description triggers.
### 2.8.1 (2026-07-08)
- **ad-creative** (2.1.0 → 2.2.0): added the **iMessage chat-reveal video ad format**. New `references/imessage-video-ads.md` — an original distillation of the format popularized by Shiv Sakhuja / Gooseworks (their public `create-imessage-video-ad` skill; the repo carries no open-source license, so content was re-expressed with credit rather than copied): a 9:16 video where a scripted iMessage thread unfolds bubble-by-bubble — screenshot hook → friend reacts → "what app is that?" → brand + promo-code reveal → static end card. Covers when to use it and when not (considered B2B, nothing screenshot-able, regulated-industry compliance), a compliance/grounding section tying dramatized conversations to the skill's Grounded Inputs rules (facts in the thread must be real; never frame the thread as a real testimonial), the six concept angles (result-as-screenshot, setup flex, cancellation moment, feature-as-punchline, friend-asks-friend inverse, receipt-as-hook), ad anatomy with script + pacing rules (8–14 bubbles, texting voice, brand appears once and late, rhythm not metronome, link-underlined promo code), three production routes (off-the-shelf Gooseworks skill; code-based HTML mockup + one continuous Playwright recording + ffmpeg stitch; Remotion for templated scale), the craft rules that sell the illusion (the recognizable send/receive SFX — CC0 recordings on BigSoundBank, with a trade-dress legal-review note; silent typing indicators; composer-typed text must equal sent text; static end card with the real logo SVG; native 1080×1920 recording; HTML-mimic real app UIs for hook screenshots instead of AI-generating them), a ship checklist, and an iteration order (hook attachment first). SKILL.md wires the reference into Generating Ad Visuals and adds 'iMessage ad,' 'chat reveal ad,' 'text message ad,' and 'fake DM ad' description triggers. Closes #411.
+5 -3
View File
@@ -1,8 +1,8 @@
---
name: ad-creative
description: "When the user wants to generate, iterate, or scale ad creative — headlines, descriptions, primary text, or full ad variations — for any paid advertising platform. Also use when the user mentions 'ad copy variations,' 'ad creative,' 'generate headlines,' 'RSA headlines,' 'bulk ad copy,' 'ad iterations,' 'creative testing,' 'ad performance optimization,' 'write me some ads,' 'Facebook ad copy,' 'Google ad headlines,' 'LinkedIn ad text,' 'static ads,' 'static ad concepts,' 'ad templates,' 'iMessage ad,' 'chat reveal ad,' 'text message ad,' 'fake DM ad,' or 'I need more ad variations.' Use this whenever someone needs to produce ad copy at scale or iterate on existing ads. For campaign strategy and targeting, see ads. For landing page copy, see copywriting."
description: "When the user wants to generate, iterate, or scale ad creative — headlines, descriptions, primary text, or full ad variations — for any paid advertising platform. Also use when the user mentions 'ad copy variations,' 'ad creative,' 'generate headlines,' 'RSA headlines,' 'bulk ad copy,' 'ad iterations,' 'creative testing,' 'ad performance optimization,' 'write me some ads,' 'Facebook ad copy,' 'Google ad headlines,' 'LinkedIn ad text,' 'static ads,' 'static ad concepts,' 'ad templates,' 'iMessage ad,' 'chat reveal ad,' 'text message ad,' 'fake DM ad,' 'ChatGPT ad,' 'Apple Notes ad,' 'motion video ad,' 'faceless video ad,' 'animated explainer ad,' 'motion collage ad,' or 'I need more ad variations.' Use this whenever someone needs to produce ad copy at scale or iterate on existing ads. For campaign strategy and targeting, see ads. For landing page copy, see copywriting."
metadata:
version: 2.2.0
version: 2.5.0
---
# Ad Creative
@@ -150,7 +150,9 @@ For detailed specs and format variations, see [references/platform-specs.md](ref
**For static ad structure**, use the 15-template library in [references/static-ad-templates.md](references/static-ad-templates.md) — layout frameworks (Us vs. Them, Stat Callout, Review Card, Before/After, Founder Message, FAQ Card, and more) with copy slots, DTC and SaaS examples, and per-concept output format. Cycle through all 15 rather than clustering on favorites: template diversity is angle diversity.
**For iMessage chat-reveal video ads** — the 9:16 format where a scripted iMessage thread unfolds bubble-by-bubble (screenshot hook → friend asks "what app is that?" → brand + promo code reveal → end card) — see [references/imessage-video-ads.md](references/imessage-video-ads.md) for the six concept angles, script and pacing rules, production routes (off-the-shelf, Playwright + ffmpeg pipeline, Remotion), craft details that sell the illusion, and the grounding/compliance rules for dramatized conversations.
**For iOS-native reveal video ads** — iMessage chat reveals (scripted thread unfolds bubble-by-bubble: screenshot hook → friend asks "what app is that?" → brand + promo code reveal → end card), ChatGPT reveals (typed question → streaming answer), and Apple Notes reveals (a confessional note typed live) — see [references/imessage-video-ads.md](references/imessage-video-ads.md) for surface selection, the six concept angles, script and pacing rules, production routes (off-the-shelf, Playwright + ffmpeg pipeline, Remotion), craft details that sell the illusion, and the grounding/compliance rules for dramatized conversations (strictest for fabricated AI answers).
**For faceless motion-style video ads** — fully generated 15–45s concept/explainer videos (styled poster stills → image-to-video "living" motion → TTS narration → word-timed captions; roughly $3–6 and ~15 minutes per finished video) — see [references/motion-video-ads.md](references/motion-video-ads.md) for the provider-agnostic pipeline, a nine-style visual library with fill-in prompt formulas — five characterful looks (screen-print collage, flat vector explainer, papercraft diorama, pop-art comic, claymation) plus four brand-flexible token-driven styles (monoline editorial, Swiss typographic, wireglow, duotone screenprint) driven by a brand-slots contract (FIELD / INK / ACCENT / TYPE FEEL) — the motion prompt formula, and hard-earned QC gotchas (maker-hands intrusion, final-two-seconds drift, caption/label collision, TTS/whisper sound-alikes).
For image and video generation tools, see [references/generative-tools.md](references/generative-tools.md) for the complete guide covering:
+15
View File
@@ -100,6 +100,21 @@
"Mentions at least one production route and key craft rules (SFX, silent typing indicator, 9:16)"
],
"files": []
},
{
"id": 8,
"prompt": "We sell a menopause supplement. I saw those ads where someone asks ChatGPT a health question and the answer recommends the product — make one of those for us. Also curious about the Apple Notes version.",
"expected_output": "Should load references/imessage-video-ads.md and apply the Other iOS-Native Reveal Surfaces section. Should flag the compliance constraint prominently BEFORE drafting: a fabricated AI answer making health claims is the highest-risk version of this format — every claim needs substantiation, health/medical advice in a fake ChatGPT answer needs legal review, and the exchange must not be presented as a real unprompted ChatGPT output endorsing the product. May propose a compliant angle (mechanism education grounded in documented facts) or steer to the Apple Notes confession format as the lower-risk fit for a transformation story. For the Notes version: title-as-hook, first-person list with the product as the least enthusiastic line, keyboard-taps-only audio, grounding realizations in real reviews. Should apply surface-selection guidance rather than treating the three formats as interchangeable.",
"assertions": [
"Applies the iOS-native reveal surfaces section of the imessage-video-ads reference",
"Flags health-claim/substantiation risk for the fabricated ChatGPT answer before or while drafting",
"Does not present the ChatGPT exchange as a real unprompted output endorsing the product",
"Recommends legal review or a compliant reframe for health advice in the AI answer",
"Apple Notes guidance: title-as-hook, first-person confession, product as an understated list item, keyboard-taps-only audio",
"Grounds claims and realizations in documented facts/reviews (Grounded Inputs)",
"Gives surface-selection reasoning (ChatGPT vs Notes) instead of treating formats as interchangeable"
],
"files": []
}
]
}
@@ -1,6 +1,6 @@
# iMessage Chat-Reveal Video Ads
# iOS-Native Reveal Video Ads (iMessage, ChatGPT, Apple Notes)
A 9:16 social-native video format that recreates an iMessage conversation unfolding in real time: someone sends a screenshot of a result or product, a friend reacts and asks what it is, and the conversation reveals the brand — usually with a promo code. Message bubbles pop in over ~15–22 seconds with authentic send/receive sounds, then a static brand end card lands the CTA.
A family of 9:16 social-native video formats that recreate a familiar iOS surface in real time and let the brand emerge inside it. The flagship is the **iMessage chat reveal** — someone sends a screenshot of a result or product, a friend reacts and asks what it is, and the conversation reveals the brand, usually with a promo code. Message bubbles pop in over ~15–22 seconds with authentic send/receive sounds, then a static brand end card lands the CTA. The same architecture powers **ChatGPT reveals** and **Apple Notes reveals** — covered in [Other iOS-Native Reveal Surfaces](#other-ios-native-reveal-surfaces) below.
The format works because it borrows the most-read UI on earth. A chat thread is a familiar, high-attention dramatization — it mirrors how real recommendations happen, so the viewer leans in instead of scrolling past. The CTA arrives conversationally ("use code FREEPACK") instead of as a hard sell, which keeps the ad-skip reflex from firing until the pitch has already landed. Run it only as a clearly labeled paid placement (Meta's "Sponsored" tag does the disclosure work); never seed it organically as if it were a real leaked conversation.
@@ -144,4 +144,43 @@ Treat the thread as the variable and the pipeline as fixed. Test in this order
4. **Peer persona** — name, avatar, and texting style shift the perceived audience
5. **Length** — try a 12-bubble and an 8-bubble cut of the same script
The same architecture extends to other chat surfaces — a ChatGPT conversation reveal, WhatsApp, Slack — same timeline-driven recording, different UI shell. Start with iMessage; it has the highest recognition and the strongest audio identity.
The same architecture extends to further surfaces too — WhatsApp, Slack, a search box — same timeline-driven recording, different UI shell.
---
## Other iOS-Native Reveal Surfaces
Everything above about production (UI mockup → timeline-driven continuous recording → deterministic SFX cues → static end card), grounding, and disclosure carries over unchanged. What changes per surface is the *persuasion mechanic* and a handful of craft details.
| Surface | Persuasion mechanic | Reach for it when |
|---|---|---|
| **iMessage** | A friend's recommendation — social proof through dialogue | The product is discovered through results people share ("what app is that?") |
| **ChatGPT** | An authoritative answer to the viewer's own question | The problem is question-shaped — something people would literally type into ChatGPT |
| **Apple Notes** | A private confession made public — first-person, no dialogue | The angle is transformation or realization ("things nobody told me about 45") |
The strongest signal for choosing: which of these three surfaces already fills your audience's day. Recommendation products want iMessage; advice-seeking problems want ChatGPT; identity/transformation stories want Notes.
### ChatGPT Reveal
The viewer identifies with the *asker*. The typed question is the hook and must be the target customer's verbatim question — awkward phrasing and all ("why is my stomach so bloated all of a sudden at 47?"). The streaming answer names the problem's real mechanism, then the solution category; the brand lands in the answer's recommendation or in a typed follow-up ("what's the best one?").
**Craft details:**
- **Stream the answer in word chunks**, not character-by-character (that's typing, not generation) and not whole paragraphs at once. A subtle tick underneath the stream and a clean stop when the response completes; no iMessage tritones anywhere.
- **Type the question like thumbs, stream the answer like a model.** Two distinct rhythms — the contrast is what reads as "real ChatGPT."
- **Keep the answer scannable:** short paragraphs, a bolded phrase or a short list, exactly the way ChatGPT actually formats. A wall of text breaks the illusion and loses the viewer.
- OpenAI's interface is their trade dress — same legal-review posture as the Apple UI mimicry note above, with a generic "AI assistant" skin as the fallback.
**Compliance — stricter here than anywhere else in this family.** The "answer" is your ad copy wearing a lab coat: an authority costume. Every claim in it needs the same substantiation as a claim in your own voice, and the format's borrowed authority raises the bar, not lowers it. Do not put health, medical, or financial advice in a fabricated AI answer without legal review — that's the highest-risk version of this format. And never present the exchange as a real, unprompted ChatGPT output endorsing your product; it's a dramatization, same as the iMessage thread.
### Apple Notes Reveal
A different genre from the chat formats: **confession, not conversation.** The viewer watches someone type a private note — a list of realizations, a "things I wish I knew" entry — with the keyboard visible. The note's title is the hook and does the job slide 1 does in a carousel ("Things nobody told me about 45."). The product appears as one item in the list, named the way a person would actually write it to themselves — not the way a brand would.
**Craft details:**
- **Audio is keyboard taps only.** No chat SFX, no receive tones — a note has no other party. A quiet music bed still works underneath.
- **Type at real thumb pace with jitter**, same as the iMessage composer rule. One typo-and-correction reads as human; several read as staged.
- **Get the Notes chrome right:** title styled larger than body, the formatting bar above the keyboard, iOS-yellow accents. Same HTML-mimicry approach — and the same Apple trade-dress review note and generic-notes-app fallback — as everything else here.
- **Fit the note to the frame.** Write short enough that the whole note fits without scrolling, or scroll once, deliberately, late.
- **First person or it doesn't work.** The moment the note reads like ad copy ("[Brand] changed everything!"), the intimacy that makes the format convert is gone. The product mention should be the *least* enthusiastic line in the note.
The grounding rule hits differently here: the confession is a dramatization of a *composite, true* customer story — pull the realizations from real reviews and interviews (the Grounded Inputs corpus), and keep any numbers or outcomes to documented ones.
@@ -0,0 +1,126 @@
# Motion-Style Video Ads (Faceless, Fully Generated)
> Format popularized by Borja ([@borjafat](https://x.com/borjafat)) and the open `super-video-maker` motion-collage recipe by [Bomx](https://github.com/Bomx/super-video-maker-skill); this guide is an original re-expression of the method, extended with a multi-style library and production lessons from building and shipping it end-to-end.
Produce a 15–45s faceless video ad or explainer from nothing but a concept: a styled
poster still (image model) → brought to life with subtle motion (image-to-video model)
→ narrated (TTS) → word-timed captions. No footage, no presenter, no editor. Cost per
finished video is roughly $3–6 in API calls; wall-clock ~15 minutes.
The format works because the *still* carries the idea (one literal, slightly surreal
visual per beat) and the *motion* only makes it breathe. Resist the urge to make the
video do the storytelling — this is animated poster design, not filmmaking.
## When to use
- Concept/explainer ads: one idea made concrete ("your CRM is a junk drawer")
- Top-of-funnel social video (9:16 Reels/Shorts/TikTok, 4:5 and 1:1 feed)
- Brand-response hybrids where a distinctive owned style beats stock UGC
- NOT for: demo/proof ads (screen recordings win), testimonial/UGC formats,
anything requiring a real product shot as evidence
## Pipeline (provider-agnostic)
1. **Script** 3–6 beats, 20–45s of VO. One idea per beat. Calm and specific beats
hype. End on a single CTA line.
2. **Poster stills** — one per beat, using a *style formula* (below). Generate beat 1,
approve it, then pass it as a reference image for every later beat so the set reads
as one series. Fix garbled label text by regenerating with a shorter phrase.
3. **Animate** each approved still with an image-to-video model (5–8s per beat).
Motion belongs to the objects in the frame; the composition must not change.
4. **VO + captions**: one continuous TTS take, transcribe with word timestamps
(whisper), cut beats at sentence boundaries, burn 2–3-word caption groups.
5. **Assemble**: concat beats trimmed to their VO spans (hold the last frame to pad),
loudness-normalize to `I=-16:TP=-1.5:LRA=11`, export per-placement aspect.
**Provider options** (any combination works; the recipe is model-agnostic):
| Stage | One-key Gemini path | Alternatives |
|---|---|---|
| Stills | Nano Banana Pro (`gemini-3-pro-image-preview`) — excellent label typography | GPT-Image, Flux, Ideogram |
| Motion | Veo 3.1 fast image-to-video (note: 1080p requires 8s clips) | Seedance 2.0 via fal.ai, Kling, Runway |
| VO | Gemini TTS (calm voices: Charon/Kore) | ElevenLabs, OpenAI TTS |
| Captions | whisper word timings + PIL/ASS burn-in | CapCut, platform auto-captions |
## The style library
Five proven looks. Each is a fill-in-the-slots prompt formula; keep ONE style per
campaign so the account builds a recognizable visual identity. All five animate well.
### A. Screen-print collage (editorial, "In a Nutshell" docu energy)
> Flat screen-print collage poster, single saturated `<COLOR>` background, subtle newsprint grain. Centerpiece: a black-and-white halftone cutout of `<SUBJECT DOING THE LITERAL CONCEPT>`, treated as a paper sticker with a thin white die-cut outline, slightly torn edges, and a soft drop shadow. Visible halftone dot texture, vintage editorial photo feel, grayscale subject. Accent cutouts: 2–4 flat shapes (cream circle sun, black zigzag, scattered dots). A torn-paper label near the bottom with the words "`<LABEL>`" in bold condensed uppercase newspaper type. Matte printed risograph aesthetic, limited palette. No gradients, no glow, no 3D, no photorealism, no extra text.
### B. Flat vector explainer (clean, techy, infinitely brandable)
> Flat vector explainer illustration in the style of a premium animated science channel: a friendly simplified `<SUBJECT>`, bold flat shapes with clean rounded edges, solid `<BRAND COLOR>` background, limited palette of `<2-3 ACCENTS>`, flat geometric accents, soft long shadows, completely flat 2D design. A clean rectangular banner near the bottom reads "`<LABEL>`" in bold geometric sans-serif uppercase. No outlines, no 3D, no photorealism, no texture, no extra text.
### C. Papercraft diorama (warm, tactile, premium-crafty)
> Layered papercraft diorama: `<SUBJECT>`, every element hand-cut from colored construction paper with visible paper thickness and real drop shadows between layers, `<COLOR>` paper background with cut-paper accents, tactile handmade craft feel with slightly imperfect scissor cuts. A cut-paper banner near the bottom reads "`<LABEL>`" in chunky cut-out paper letters. Soft studio lighting on the paper layers. No digital gradients, no photorealistic humans, no extra text.
### D. Pop-art comic (loud, scroll-stopping, promo-friendly)
> Vintage pop-art comic panel: `<SUBJECT>`, bold black ink outlines, Ben-Day halftone dots shading, flat process colors (`<PALETTE>`), comic starburst accents, thick panel border, aged newsprint paper texture. A comic caption box near the bottom reads "`<LABEL>`" in bold comic lettering. 1960s printed comic aesthetic, slight ink misregistration. No 3D, no photorealism, no gradients, no extra text.
### E. Claymation (charming, high pattern-interrupt)
> Stop-motion claymation scene: a charming handmade plasticine `<SUBJECT>`, visible fingerprints and clay texture, `<COLOR>` clay backdrop and floor, chunky clay props, warm soft studio lighting like a stop-motion film set, shallow depth of field. A small clay sign near the bottom reads "`<LABEL>`" in hand-molded clay letters. Handcrafted miniature feel. No 2D illustration, no photorealistic humans, no extra text.
## Brand-flexible styles (token-driven)
The five looks above are *characterful* — they impose their own palette. This second
tier is *brand-first*: each style is defined by *slots*, so any company's tokens drop
in and the output reads as that brand's own design system.
**The brand slots contract.** Before generating, resolve these from the brand's
guidelines (or `.agents/product-marketing.md`):
- `FIELD` — the neutral ground (brand white/off-white, or brand dark)
- `INK` — the drawing/type color (brand gray/charcoal, near-black)
- `ACCENT` — ONE brand color or gradient, used sparingly (a rule, a beam, a square)
- `TYPE FEEL` — the brand's typographic voice ("clean modern grotesque sans", "geometric sans", "mono captions")
- Any per-brand constraints (e.g. "gradients only on borders/edges, never fills")
Keep the accent genuinely scarce — one element per frame. Scarcity is what makes
these read as designed rather than generated.
### F. Monoline editorial (the most universally brandable)
> Minimal editorial monoline illustration poster: `<SUBJECT>`, drawn entirely in elegant thin single-weight `<INK>` lines on a clean `<FIELD>` background, the style of a premium tech company blog illustration. Sparse composition with generous whitespace, a few small monoline accent details, and ONE restrained `<ACCENT>` element: `<a thin accent underline sweep / a small accent arc>`. A small caption near the bottom reads "`<LABEL>`" in `<TYPE FEEL>`, `<INK>`, letterspaced uppercase, with a thin `<ACCENT>` underline. Precise, technical, refined. No fills except the single accent, no gradients, no 3D, no photorealism, no texture, no extra text.
### G. Swiss typographic (type IS the visual — any brand with a font and a color)
> Swiss International Typographic Style poster: the words "`<LABEL>`" set enormous in a bold `<TYPE FEEL>`, `<INK>` on a `<FIELD>` background, filling the upper two thirds with tight leading and cropped edges. A small black-and-white photographic cutout of `<SUBJECT>` sits on a thin baseline grid in the lower third, aligned to an asymmetric grid with one thin `<ACCENT>` rule line and a small `<ACCENT>` square as the only color. Visible faint grid lines, precise margins, mathematical composition. Flat, printed, matte. No gradients, no 3D, no decoration, no extra text beyond the label and one small letterspaced caption line.
### H. Wireglow (dark keynote — dev-tool / dark-mode brands)
> Dark minimal tech-keynote poster: `<SUBJECT>` rendered as an elegant thin light-gray wireframe line drawing on a near-black `<FIELD>` background with subtle film grain. From `<the focal object>` emanates a soft narrow beam of glowing `<ACCENT>` gradient light, the only color, feathered and atmospheric. Faint thin concentric geometric guide circles. A caption near the bottom reads "`<LABEL>`" in `<TYPE FEEL>`, light gray, letterspaced uppercase, with a hairline gradient rule beneath it. Restrained, premium, technical. No photorealism, no 3D render look, no busy elements, no extra text.
### I. Duotone screenprint (photo brands — editorial punch from two tokens)
> Bold duotone screenprint photo poster: a dramatic photograph of `<SUBJECT>`, reproduced as a two-color screenprint — `<INK>` for the shadows and `<ACCENT>` for the highlights — on an off-white `<FIELD>` paper background with visible coarse halftone grain and slight ink misregistration. Strong diagonal composition, the figure large and cropped. A wide solid `<INK>` bar near the bottom carries the words "`<LABEL>`" reversed out in bold condensed `<TYPE FEEL>` uppercase, with a small `<ACCENT>` square bullet. Editorial poster energy, matte printed feel. No gradients beyond the duotone, no 3D, no extra text.
**Motion notes for this tier**: F/G animate as drawing motions (lines extend, the accent
sweep draws itself, type settles by a few pixels); H animates as beam pulse + slow
wireframe rotation feel; I as grain shimmer + slow push. Same hard rules apply — motion
belongs to existing elements, composition never changes.
## Motion prompt formula
> Subtle living-`<style>` motion of the existing elements only. `<ONE literal motion tied to the concept: the pile inflates / the arrow creeps higher / the megaphone trembles with each shout>`. `<Secondary ambient motion: accents drift, gentle push-in>`. Every element that is visible now is the only thing that ever appears; the composition stays exactly as it is. Everything stays `<style descriptor: a flat printed collage / flat 2D vector / cut paper / printed comic / handmade clay>`. No camera whip, no scene change, no morphing, no added text.
## Hard-earned gotchas
- **Video models love adding photoreal "maker hands"** reaching into frame, especially
on pressing/handling motions — and *negative prompts make it worse* ("no hands" is an
attention trap). Never mention hands; describe motion as belonging to the objects,
and include "the composition stays exactly as it is."
- **Always QC each clip's final 2 seconds** — that's where intruding objects and style
drift appear. Trim before them or regenerate; never ship a "realified" frame.
- **One dominant motion per beat.** Two motions read as chaos at feed speed.
- **TTS + whisper disagree on sound-alikes** ("laws" → "loss"). Read the transcript
against the script before burning captions; prefer phoneme-unambiguous CTA wording.
- **Keep captions clear of the label band** (captions ~60% height, label ~80%).
Clamp caption groups so two never overlap; shrink-to-fit long groups.
- **Ad-specific**: put the brand/label in the poster itself (it survives sound-off
autoplay), front-load the concept in beat 1 (the 3-second hook is the poster), and
export 9:16 + 4:5 + 1:1 from the same beats by regenerating stills per aspect
rather than cropping.
## Compliance
Fully synthetic characters — no likeness/UGC disclosure issues, but check platform
AI-content disclosure requirements (Meta and TikTok label AI-generated media).
Don't fabricate statistics or testimonials in the VO; ground every claim.
+4 -2
View File
@@ -1,8 +1,8 @@
---
name: social
description: "When the user wants help creating, scheduling, or optimizing social media content for LinkedIn, Twitter/X, Instagram, TikTok, Facebook, or other platforms, or wants to do social listening and engagement triage. Also use when the user mentions 'LinkedIn post,' 'Twitter thread,' 'social media,' 'content calendar,' 'social scheduling,' 'engagement,' 'viral content,' 'what should I post,' 'repurpose this content,' 'tweet ideas,' 'LinkedIn carousel,' 'social media strategy,' 'grow my following,' 'TikTok video,' 'Reels,' 'Shorts,' 'video script,' 'video hook,' 'short-form video,' 'create a reel,' 'social listening,' 'brand mentions,' 'competitor monitoring,' 'top posts to comment on,' or 'find people asking for.' Use this for social media content creation, repurposing, scheduling, short-form video scripting, and social listening. For broader content strategy, see content-strategy. For paid ads, see ad-creative. For earned media, see public-relations."
description: "When the user wants help creating, scheduling, or optimizing social media content for LinkedIn, Twitter/X, Instagram, TikTok, Facebook, or other platforms, or wants to do social listening and engagement triage. Also use when the user mentions 'LinkedIn post,' 'Twitter thread,' 'social media,' 'content calendar,' 'social scheduling,' 'engagement,' 'viral content,' 'what should I post,' 'repurpose this content,' 'tweet ideas,' 'LinkedIn carousel,' 'social media strategy,' 'grow my following,' 'TikTok video,' 'Reels,' 'Shorts,' 'video script,' 'video hook,' 'short-form video,' 'create a reel,' 'social listening,' 'brand mentions,' 'competitor monitoring,' 'top posts to comment on,' 'find people asking for,' 'carousel,' 'slide-by-slide,' or 'document post.' Use this for social media content creation, repurposing, scheduling, short-form video scripting, and social listening. For broader content strategy, see content-strategy. For paid ads, see ad-creative. For earned media, see public-relations."
metadata:
version: 2.1.0
version: 2.2.0
---
# Social Content
@@ -105,6 +105,8 @@ The first line determines whether anyone reads the rest.
**For post templates and more hooks**: See [references/post-templates.md](references/post-templates.md)
**For carousels** (Instagram carousels, LinkedIn document posts): See [references/carousel-frameworks.md](references/carousel-frameworks.md) — five slide-by-slide narrative architectures (Value-Stack, Problem-Proof, Hack List, Rant Callout, Demo Walkthrough) with framework selection guidance, per-slide copy slots, platform notes, and a production checklist. Pick the framework before writing slides.
---
## Content Repurposing System
+15
View File
@@ -87,6 +87,21 @@
"May note social-to-email bridge context"
],
"files": []
},
{
"id": 7,
"prompt": "I want to turn our new pricing guide into a LinkedIn carousel. It has 9 negotiation tactics for SaaS buyers. Help me structure the slides.",
"expected_output": "Should load references/carousel-frameworks.md and pick a framework BEFORE writing slides — a list of 9 named tactics maps to Hack List (or Value-Stack; should justify the choice), not Rant Callout or Demo Walkthrough. Should produce a slide-by-slide structure: contrarian or exact-count cover (slide 1 written as a standalone scroll-stopper), one named tactic per slide, and a synthesis + single-CTA close. Should apply LinkedIn document-post notes (PDF upload, denser text tolerated, post text written as its own hook rather than 'see attached'). Should keep any count promised on the cover exactly delivered and flag that stats/claims must be real. Should mention judging performance on saves/completion rather than likes.",
"assertions": [
"Loads or applies the carousel-frameworks reference",
"Selects an appropriate framework (Hack List or Value-Stack) with a stated reason before drafting slides",
"Slide 1 is written as a standalone scroll-stopper with the count or contrarian claim",
"One tactic/idea per slide, each named",
"Final slide has a synthesis and exactly one CTA",
"Applies LinkedIn document-post guidance (PDF, post text as its own hook)",
"Does not invent stats or pad the tactic count"
],
"files": []
}
]
}
@@ -0,0 +1,141 @@
# Carousel Frameworks
Five slide-by-slide narrative architectures for carousels — Instagram carousels and LinkedIn document posts. Each framework is a different *reason to keep swiping*; picking the right one for the content matters more than polishing individual slides.
A carousel is not a blog post chopped into squares. It's a swipe-through with two jobs per slide: deliver one idea, and make the next swipe irresistible. These frameworks encode structures that show up repeatedly in high-performing carousels; treat them as starting structures to adapt, not rigid formulas — and expect to validate against your own analytics.
## Picking a Framework
| Your content is... | Use | Why |
|---|---|---|
| A list of resources, tools, or tips | **A: Value-Stack** | Completeness is the promise; the count is the hook |
| A personal result with a system behind it | **B: Problem-Proof** | Proof opens and closes the loop; the system is the meat |
| Several named techniques on one theme | **C: Hack List** | Each hack re-earns the swipe independently |
| A strong opinion about a common practice | **D: Rant Callout** | Conviction is the content; structure keeps it fair |
| A product or workflow you can show | **E: Demo Walkthrough** | Seeing the steps is more persuasive than describing them |
Two cross-framework rules before the specifics:
- **Slide 1 is the thumbnail.** It competes in the feed alone, before anyone knows a carousel follows. Design and write it as a standalone scroll-stopper.
- **One visual template per carousel.** Same layout, type scale, and palette on every interior slide. Variety between slides reads as clutter; consistency lets the *content* change while the frame stays still.
---
## Framework A: Value-Stack (4–14 slides)
The "everything you need" carousel. Works because the cover makes a completeness claim — an exact count and an exact deliverable — and every swipe pays it down.
| Slide | Job | Pattern |
|---|---|---|
| 1 — Cover | State the exact count + exact deliverable. Specificity proves scale upfront. | "[N] [resource type] for [role/outcome]" |
| 2 to N−1 — Value delivery | One item or category per slide, 3–6 concrete sub-items each. Same visual template, zero filler slides. | "[Category/item] — [sub-items]" |
| N — Close | Convert the swipe-through into an action. | "[Action] for [the payoff]" — comment a keyword, follow, link in bio |
**Why the exact count matters:** "27 free tools" outperforms "the best free tools" because a number is a checkable promise — the reader can verify you delivered. Rounding up with padding breaks the trust the count created.
*SaaS example:* "12 ChatGPT prompts for SaaS onboarding emails" → one prompt per slide with the use case → "Comment PROMPTS and I'll send the full doc."
**Failure mode:** filler slides to hit a bigger number. Cut to the real count; a tight 8 beats a padded 14.
---
## Framework B: Problem-Proof (6–10 slides)
The "I did X, here's the system" carousel. The hook is a *result stated as fact*, not advice — and the final slide shows the receipt, closing the loop slide 1 opened.
| Slide | Job | Pattern |
|---|---|---|
| 1 — Hook | A specific personal claim with a number. A result, not a tip. | "[I/we did X]. [Specific result, with number]." |
| 2 — Reframe the problem | Name what's *actually* going wrong so the reader recognizes themselves. | "The real problem — [named issue]" |
| 3–4 — The mechanism | Show the real system: named tools, named steps. Concrete beats vague. | "The system — [named tools/steps]" |
| 5 to N−1 — The detail | The literal prompt text, template, or step-by-step. This is the save-worthy part. | "[The literal prompt or process detail]" |
| N — Proof | The actual output artifact. A screenshot, not a description. | "[Screenshot of output] — the receipt" |
**The open loop is the engine:** slide 1 makes a claim, the last slide proves it, and everything between explains how. Readers swipe to see whether the receipt is real.
*SaaS example:* "We cut churn 22% with one onboarding email." → the real problem (activation, not price) → the sequence structure → the literal email copy → the retention chart.
**Failure mode:** vague mechanism slides ("optimize your workflow") — if the middle slides don't name tools and steps, the proof slide reads as luck, not system.
---
## Framework C: Hack List (6–10 slides)
The "numbered techniques" carousel. A contrarian cover creates status anxiety — most people are doing this wrong — and each numbered hack independently re-earns the swipe.
| Slide | Job | Pattern |
|---|---|---|
| 1 — Contrarian hook | A stat or claim implying most people fail at this. | "[Stat implying most people fail at X]" |
| 2 — The problem | Why the common approach fails, ideally with an analogy. | "The problem — [why it fails]" |
| 3 to N−1 — Numbered hacks | One named technique per slide. Bad-vs-good contrast makes each hack land instantly. | "Hack #[n] — [named technique]" |
| N — Synthesis + close | A one-line thesis tying the hacks together, then a save/share/follow CTA. | "[Thesis tying hacks together]" + [CTA] |
**Name every technique.** "The 3-second rule" is shareable and memorable; "keep it short" is neither. Named techniques travel — people repeat them and credit the source.
*SaaS example:* "90% of trial emails never get opened." → why (they read like receipts, not messages) → Hack #1: The founder-from line, Hack #2: The one-question subject, ... → "Trial emails are conversations, not confirmations. Save this for your next sequence."
**Failure mode:** hacks that are restatements of each other. Each slide must survive alone — if two hacks collapse into one idea, merge them and go shorter.
---
## Framework D: Rant Callout (4–8 slides)
The "someone had to say it" carousel. Personality-led and polarizing by design — the structure exists to keep the heat *fair* so it reads as conviction, not bitterness.
| Slide | Job | Pattern |
|---|---|---|
| 1 — Provocative claim | An unpopular opinion or direct accusation about a common practice. | "Unpopular opinion — [common practice]" |
| 2 to N−2 — Escalate the argument | Sensory, specific detail. Show the offense; don't gesture at it abstractly. Each slide raises the stakes. | "[Escalating, specific complaint]" |
| N−1 — Fairness pivot | "Don't get me wrong…" — clarify what you're *not* attacking. Anti-laziness, not anti-tool; anti-practice, not anti-person. | "The problem isn't [X], it's [Y]" |
| N — Close | Firm, personality-forward sign-off. Signed rants read as owned opinions; anonymous ones read as potshots. | "[Firm sign-off] — signed [name/persona]" |
**The fairness pivot is what makes it work.** Without it you're yelling; with it you're drawing a precise line, and precise lines get quoted. It also pre-empts the top hostile comment.
*SaaS example:* "Your AI-generated LinkedIn posts are costing you customers." → the specifics (same em-dash cadence, same 'game-changer' vocabulary, zero lived detail) → "Don't get me wrong — I use AI daily. The problem isn't the tool, it's publishing the first draft." → "Write like you talk. — [Name]"
**Failure modes:** skipping the pivot (reads unhinged, invites pile-ons), or ranting about something your own product/content visibly does (the comments will find it).
---
## Framework E: Demo Walkthrough (5–11 slides)
The "watch it work" carousel for product or workflow content. Show the finished result first, then the steps — the outcome earns attention for the process.
| Slide | Job | Pattern |
|---|---|---|
| 1 — Product + outcome | Brand/product plus the finished result. Lead with what they get. | "[Product] + [outcome], finished result" |
| 2 — Problem it solves | The specific frustration this replaces. | "[The specific pain this replaces]" |
| 3 — Process overview | The numbered step list, shown *before* the detail. Seeing the whole path lowers mid-swipe drop-off — readers commit when they know how long the road is. | "[N] steps to [outcome] in [timeframe]" |
| 4 to N−1 — Step-by-step | One real UI screenshot per step with a short caption. Real screenshots; mockups read as vaporware. | "Step [n] — [screenshot] — [caption]" |
| N — Result + positioning | The full-size output plus one brand-philosophy line. | "[Final output] + [positioning line]" |
**Overview-before-detail is the key move** (same reason recipe sites list ingredients before instructions): a reader who's seen "4 steps, 10 minutes" on slide 3 will finish; a reader dropped straight into step 1 with no map bails at step 2.
*SaaS example:* "From CSV to live dashboard in 4 steps" → the pain (weekly manual reporting) → the 4-step overview → one screenshot per step → the finished dashboard + "Built for marketers, not analysts."
**Failure mode:** screenshots with too much UI. Crop to the action, zoom the relevant control, annotate with one arrow — each screenshot should make sense at thumbnail size.
---
## Platform Notes
**Instagram:** 1080×1350 (4:5) slides; up to 20 slides but the frameworks' ranges above reflect where completion rates hold. The caption carries the searchable text — restate the hook and add context, don't just repeat slide 1. First slide doubles as the grid thumbnail.
**LinkedIn (document posts):** Upload as PDF. Square or 4:5. LinkedIn shows a "N pages" badge — Value-Stack's exact-count cover plays especially well here. B2B tolerates denser text per slide than Instagram, but the one-idea-per-slide rule still holds. The post text above the document is a second hook — write it as its own post, not a throwaway "see attached."
**TikTok photo mode:** Same frameworks work; pacing skews faster — trim each framework toward its minimum slide count.
## Production Checklist
- [ ] Slide 1 works as a standalone feed post (test: would it stop the scroll with no carousel behind it?)
- [ ] One idea per slide — if a slide needs two sentences of setup, split it
- [ ] Same visual template on every interior slide
- [ ] Any count promised on the cover is exactly delivered (no filler, no rounding)
- [ ] Claims and numbers are real and yours — no invented stats or borrowed screenshots
- [ ] Text is legible at thumbnail size (minimum ~28pt at 1080px wide)
- [ ] Final slide has one CTA, not three
- [ ] Caption/post text written as its own hook, not a repeat of slide 1
## Measuring What Worked
Judge carousels on **saves and completion**, not likes — saves signal reference value (the algorithmic win for carousels), and where per-slide analytics exist (Instagram professional insights, LinkedIn document analytics, or third-party tools), drop-off position tells you which slide broke the swipe chain. When a carousel underperforms, the fix is usually slide 1 (didn't stop the scroll) or the framework choice (list content forced into a rant structure), not the middle slides. Track per-framework performance and double down on the two that fit your audience — see [reverse-engineering.md](reverse-engineering.md) for the analysis workflow.
@@ -133,6 +133,8 @@ Tweet 7: [Your key takeaway]
Caption: [Expand on the topic, add context, include CTA]
```
This is the generic shape. For five full slide-by-slide narrative architectures (Value-Stack, Problem-Proof, Hack List, Rant Callout, Demo Walkthrough) with selection guidance and per-slide copy slots, see [carousel-frameworks.md](carousel-frameworks.md).
### The Reel Script
```
Hook (0-2 sec): [Pattern interrupt or bold claim]