opondo300/born-builders-agent-library/instagram-carousel-builder

Claude Code and agent skill by opondo300/born-builders-agent-library

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

End-to-end process for building 1080×1080 Instagram carousels — from brief through desire audit, copy, implementation, and Playwright verification. Covers 7-slide progressive revelation architecture, design system, component patterns, and per-slide persuasion checklist. Invoke before starting any carousel build.

Install

npx vskill@latest install opondo300/born-builders-agent-library/instagram-carousel-builder

SKILL.md

Instagram Carousel Builder

Reference carousels (working examples):

  • website/src/app/slides/slop-*/ — Anti-AI Slop Protocol (7 slides)
  • website/src/app/slides/stress-*/ — Business Advisor Stress-Test (7 slides)
  • website/src/app/slides/map-*/ — Buyer Psychology Map (7 slides)

Phase 1 — Brief

Before writing a single word, answer these five questions. If any are unanswered, the carousel will underperform regardless of execution quality.

  1. Topic: What skill/framework/protocol does this carousel deliver?
  2. Unique mechanism: WHY does this work differently from the obvious alternative? Name the mechanism — it must be specific enough to label (e.g. "The uncanny valley of text", "Five lenses. One ranked output.", "8 questions from real calls").
  3. Comment keyword: One word the viewer will comment to trigger ManyChat delivery (e.g. SLOP, STRESS, MAP). All-caps. This word becomes the CTA on slide 07 and appears as a footer keyword on slides 01–06.
  4. Primary human drive: Which of the 8 core drives is this primarily activating? Most carousels hit Learn + Defend. Identify both.
  5. Hopkins private fear: What is the private, embarrassing cost the viewer carries but does not say out loud? Not the intellectual problem — the silent personal consequence. This goes on slide 01.

Phase 2 — Desire Audit (run before writing copy)

Answer all six. If you cannot, the copy will be intellectually clear but emotionally inert.

  1. Which drive? Learn / Defend / Status / Acquire / Bond / Create / Experience / Transcend
  2. Wanting vs. liking? Are you building anticipation (wanting system — dopamine, scarcity, incomplete loops) or describing satisfaction (liking system — outcomes, testimonials)? Carousels should be 80% wanting, 20% liking.
  3. Loss framing: What specific loss does inaction produce? Frame in terms of what they lose, not what they gain.
  4. Identity Triangle:
  • Current self: who they are now (problem state)
  • Aspirational self: who they become with the skill
  • Rejected self: who they become if they skip it — name this explicitly on slides 01 or 06
  1. Tribal signal: What does using this skill signal about the viewer to their peers? ("I write at a human level" / "I test my assumptions before I commit")
  2. Somatic marker: What should the visual + copy make the body feel before the mind decides? (urgency / relief / recognition / respect)

Phase 3 — 7-Slide Architecture

Progressive revelation — each slide ends on an incomplete thought. The reader must swipe to resolve. Only slide 07 resolves (with a question, not a statement).

Slide 01 — Hook (Confrontational claim + Hopkins private fear)

Purpose: Activate the Defend drive. Name the intellectual problem AND the private, embarrassing cost in the same slide.

Structure:

  • Big headline (100–112px Instrument Serif, dark): confrontational claim
  • Gold squiggle divider
  • Body mono (14px JetBrains Mono): intellectual problem + Hopkins private fear line
  • Visual proof card at bottom (white card with flagged example, comparison table, or "describes everyone" demo)

Hopkins private fear rule: The private fear is not the intellectual problem. It is what the viewer carries silently. Not "your writing has AI words in it" — "You'll never know who stopped reading. Or what you lost."

Copy checklist:

  • [ ] First sentence = the point (no warmup)
  • [ ] Hopkins private fear: named explicitly
  • [ ] Rejected self: implied or stated
  • [ ] Ends without resolution — the visual card raises the problem but does not solve it

Slide 02 — Villain escalation (Data + forward pull)

Purpose: Prove the problem is worse than they thought. End with a line that pulls them to slide 03.

Structure:

  • Context label (9px JetBrains Mono gold, uppercase): e.g. "THE RESEARCH"
  • Hero stat (240–360px Instrument Serif, dark): the number that shocks (e.g. "48×")
  • Gold squiggle divider
  • Stat label (15px JetBrains Mono): what the number means
  • Attribution line (10px JetBrains Mono, faint): source + year
  • Forward pull (13px JetBrains Mono italic, faint): incomplete thought pulling to slide 03

Forward pull rule: The last line must be incomplete. It points at what is coming without resolving it.

  • Correct: "That's one word. There are 47 more."
  • Correct: "The 8 questions do."
  • Wrong: "Which is why the protocol exists." (resolves — no reason to swipe)

Slide 03 — Unique mechanism (THE MECHANISM)

Purpose: Name WHY this works differently. Not what it contains — the underlying reason it produces results the alternative cannot.

Structure:

  • Context label (9px JetBrains Mono gold uppercase): "THE MECHANISM"
  • Headline split (86–90px Instrument Serif): dark line + gold line
  • Gold squiggle divider
  • Body (3 JetBrains Mono paragraphs, 13–14px): mechanism explained in plain language — cause then effect then consequence
  • Attribution (10px faint): if sourced from research

Mechanism writing rule: The mechanism is not a feature list. It is the WHY. "Five frameworks each attack a different failure mode" is a mechanism. "Contains 5 frameworks" is a feature.

Body paragraph structure:

  1. The underlying reality (what is actually happening)
  2. The chain of events it triggers (cause → effect)
  3. The consequence (italic Instrument Serif for weight — this is the line that lands)

Slide 04 — Specific components (What is inside)

Purpose: Make the mechanism concrete. Show the parts.

Structure options:

  • Numbered list (bold Montserrat names + JetBrains Mono descriptions)
  • Two-column grid with gold numbered labels
  • Hero number (200–240px Instrument Serif) + component grid

Rule: Every item must be specific enough to be actionable. "Resistance Beliefs" is specific. "Belief mapping" is not.


Slide 05 — Output sample (What it produces)

Purpose: Show the result. The viewer should be able to see themselves using this.

Structure:

  • Headline split: "What the skill / produces." (dark + gold)
  • White card labeled "Sample Output" with 3–5 realistic output rows
  • Output rows use alternating background for scannability

Output sample rule: The sample must look like something they could actually use tomorrow. Generic examples ("identifies key issues") are worthless. Specific examples ("Resistance belief: 'I've tried automation before. It didn't work for my team.'") create desire.


Slide 06 — Stakes (What skipping it costs + personal exposure)

Purpose: Activate loss aversion. Name the business cost AND the personal cost (credibility, reputation, the fact that their name is attached).

Structure:

  • Headline split: "What skipping it / costs." (dark + gold)
  • Gold squiggle divider
  • 3 red-left-border white cards: each names a specific cost scenario
  • Footer stat (faint JetBrains Mono): the closing line that names the personal exposure

Personal exposure rule (Google/CEB B2B research): Business costs alone underperform. The viewer needs to see their personal stake. "Your name is on the call you made with confidence." "You're the one who said this was solid."

Footer stat rule: Must name credibility, reputation, or personal exposure — not just money or time.


Slide 07 — CTA (Comment keyword + offer + question)

Purpose: Trigger the comment. The question at the end activates the algorithm (comment intent) and completes the wanting loop.

Structure:

  • Context label (faint JetBrains Mono): "Comment"
  • Keyword (140–220px Montserrat 900, gold): the comment trigger word
  • Dark squiggle divider (opacity 0.2)
  • Offer body (36–40px Instrument Serif, dark): what they receive when they comment — 2–3 sentences, concrete
  • Italic question (12–13px JetBrains Mono italic, faint): the closing question
  • Bottom left: [YOUR BRAND] + yourdomain.com
  • Bottom right: CLAUDE SKILL · PASTE → USE

Offer body rule: Say exactly what they receive. "We'll send the full protocol. 47 trigger words. 13 structural tells. The 16-point humanization checklist. Paste into Claude before anything ships."

Question rule:

  • Must be a genuine question about their specific situation
  • Must NOT be "What do you think?" or "Have you experienced this?" (hollow engagement bait)
  • Must connect directly to the carousel unique mechanism
  • Examples:
  • SLOP: "What's the word you keep catching yourself using?"
  • STRESS: "What's the decision you're about to commit to?"
  • MAP: "What's the thing your buyers say that you still don't fully understand?"

Phase 4 — Design System

Canvas
Width: 1080px
Height: 1080px
Background: #EDE8DF (cream)
Padding: 68px 72px 60px
Color palette
Dark (primary text): #0C1419
Gold (accent): #C8A24F
Gold dark (labels): #7A5C1E
Red (villain/wrong): rgba(200,50,40,X) — 0.3 border / 0.6 label / 0.85 text
White (cards): #FFFFFF
Faint text: rgba(12,20,25,0.35–0.48)
Very faint: rgba(12,20,25,0.15–0.28)
Typography
Display headline: Instrument Serif, Georgia, serif — weight 400, lineHeight 0.92–0.95
Dominant keyword: Montserrat, sans-serif — weight 900 (CTA keywords, footer keyword)
Body / labels: JetBrains Mono, monospace — weight 400–500
Decorative SVG elements

Decorative arc (every slide, bottom-right):

jsx
<svg width="200" height="200" viewBox="0 0 200 200" fill="none"
style={{ position: "absolute", bottom: "-70px", right: "-70px", opacity: 0.20 }}>
<circle cx="200" cy="200" r="170" stroke="#C8A24F" strokeWidth="2.5"/>
<circle cx="200" cy="200" r="140" stroke="#C8A24F" strokeWidth="2"/>
</svg>

Gold squiggle divider (after headline, before body):

jsx
<svg width="68" height="10" viewBox="0 0 68 10" fill="none"
style={{ display: "block", marginBottom: "28px" }}>
<path
d="M0 5 C4.5 1, 9 1, 13.5 5 C18 9, 22.5 9, 27 5 C31.5 1, 36 1, 40.5 5 C45 9, 49.5 9, 54 5 C58.5 1, 63 1, 67.5 5"
stroke="#C8A24F" strokeWidth="3.5" strokeLinecap="round" fill="none"/>
</svg>

Dark squiggle (slide 07 only, under keyword): same path, stroke rgba(12,20,25,0.12), strokeWidth 3.

Standard header (every slide)
jsx
<div style={{ display: "flex", justifyContent: "space-between",
alignItems: "center", marginBottom: "52px" }}>
<span style={{ fontFamily: "'JetBrains Mono', monospace", fontSize: "10px",
letterSpacing: "0.22em", textTransform: "uppercase",
color: "rgba(12,20,25,0.35)" }}>
{SERIES_TITLE}
</span>
<span style={{ fontFamily: "'JetBrains Mono', monospace", fontSize: "10px",
letterSpacing: "0.18em", color: "rgba(12,20,25,0.2)" }}>
{NN} — 07
</span>
</div>
Standard footer (slides 01–06)
jsx
<div style={{ display: "flex", justifyContent: "space-between",
alignItems: "flex-end" }}>
<div style={{ display: "flex", flexDirection: "column", gap: "8px" }}>
<span style={{ fontFamily: "'JetBrains Mono', monospace", fontSize: "9px",
letterSpacing: "0.24em", textTransform: "uppercase",
color: "rgba(12,20,25,0.28)" }}>
Comment to receive the skill
</span>
<span style={{ fontFamily: "'Montserrat', sans-serif", fontWeight: 900,
fontSize: "28px", letterSpacing: "-0.02em", color: "#0C1419", lineHeight: 1,
borderBottom: "1.5px solid #C8A24F", paddingBottom: "4px",
display: "inline-block" }}>
{KEYWORD}
</span>
</div>
<span style={{ fontFamily: "'Montserrat', sans-serif", fontWeight: 900,
fontSize: "10px", letterSpacing: "0.04em", color: "rgba(12,20,25,0.15)" }}>
BUILD N BLOOM
</span>
</div>
White card (comparison / audit / sample output)
jsx
<div style={{ background: "#FFFFFF", border: "1px solid rgba(12,20,25,0.08)",
borderTop: "2px solid #C8A24F", borderRadius: "6px", padding: "22px 20px" }}>

This is the start of the file. Read the full SKILL.md on GitHub.

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/opondo300/born-builders-agent-library/instagram-carousel-builder/badge)](https://verified-skill.com/skills/opondo300/born-builders-agent-library/instagram-carousel-builder)

Something wrong with this skill or its page? Tell us.