packaging-mockup

name	packaging-mockup
description	Generate Amazon packaging mockup variations for split testing — Sunspot Packaging Test™, Open Package Test™, Back Packaging Test, Brand Name Test variations using Higgsfield Flux Kontext / Nano Banana. Use when running `packaging-mockup {ASIN} --variant={test-type}` or when running ProductPinion pre-defined templates that need packaging variations.

packaging-mockup — Packaging Variation Generator

Several ProductPinion pre-defined templates (Sunspot, Open Package, Back Packaging, Brand Name, Tilted Image) need packaging variations to test. This skill generates them via Higgsfield Flux Kontext or Nano Banana.

Prerequisites

• Higgsfield MCP
• Product reference image (current packaging)

Invocation

packaging-mockup {ASIN} --variant=sunspot       # solid color / darker package
packaging-mockup {ASIN} --variant=open          # product opened, contents visible
packaging-mockup {ASIN} --variant=back          # back of packaging visible
packaging-mockup {ASIN} --variant=brand-name    # alternate brand name on label
packaging-mockup {ASIN} --variant=tilted        # angled / tilted main image
packaging-mockup {ASIN} --variant=keyphrase     # visible keyphrase on label
packaging-mockup {ASIN} --all                   # generate all 6 variants

Output

/tmp/cro-content/{ASIN}-packaging-{variant}.png — one or more mockup images.

Variant Recipes

Sunspot Packaging Test™

Pull current product, generate variant with:

• Single dominant solid color (different from current)
• Darker overall tone
• Same logo placement
• Hides label complexity

Open Package Test™

• Product unboxed / opened
• Lid removed
• Contents partially visible
• Stops the scroll via curiosity

Back Packaging Test

• Back of packaging facing forward
• Shows ingredient list / spec / story
• Useful when back has differentiation copy

Brand Name Test

• Same packaging, alternate brand name on front
• Test 2-3 brand name candidates
• Use a separate ProductPinion poll to pick

Tilted Image Test

• Same packaging, angled/tilted vs straight-on
• Tests which presentation reads better at thumbnail

Keyphrase Test™

• Add a visible keyphrase to the label/front
• Examples: "0g Sugar" / "100% Natural" / "Made in USA"
• Tests whether claim-on-package improves CTR

Phase 1 — Generate

Higgsfield Flux Kontext for packaging modifications (best at structured edits to existing imagery). Nano Banana Pro fallback for cleaner standalone generation.

Phase 2 — Validate Compliance

Amazon main image rules apply if used as main:

• Pure white background
• 85%+ product fill
• No text added that violates compliance (varies by category — some allow on-package text, some don't)

Phase 3 — Output

# Packaging Mockup — {ASIN}

**Variant:** {type} | **Date:** {date}

## Current vs Variant

| | Current | Variant |
|---|---------|---------|
| Image | ![](current.png) | ![](variant.png) |

## Compliance Check

- White BG: ✅
- Product fill: 87%
- No added text: ✅

## Recommended Next Skill

- Test against current: `main-image-poll {ASIN} --concepts=current,variant`
- Or stage in main image pipeline: `main-image-pipeline {ASIN}`

Reference Files

• ProductPinion docs — pre-defined poll templates
• ~/.claude/skills/cro/main-image-best-practices.md — Amazon compliance rules

Quality Bar

• Variant is meaningfully different (not a near-duplicate)
• Compliance checked
• Side-by-side with current included

Auto-Triggers

• ProductPinion test plans that reference these template names
• User asks "generate {variant} mockup"

v0.1.12 — Read brain.image_strategy before generating

Before drafting any Higgsfield prompt, read brain.image_strategy from working memory:

• If null → pause and offer image-strategy first (~10 min, makes every future render bespoke). If user proceeds without, flag concept cards with "No category research used — generic aesthetic."
• If set + fresh (<90d) → inject prompt_adjustments.scene_keywords into prompt section 1, palette_keywords + mood_keywords into section 4, and do NOT include {anti_patterns_csv} near the end. Cite the strategy on each concept card.
• If stale (>90d) → flag and recommend image-strategy --refresh.

This brand-level strategy comes from skills/image-strategy.md's top-15-bestsellers analysis. Same adjustments apply across every render for this brand.

v0.1.13 — Visual verification gate + base64 embedding (NON-NEGOTIABLE)

Florence does NOT present an image she hasn't actually looked at, AND she does NOT use raw Higgsfield URLs in artifact HTML. Two blocking rules added in v0.1.13:

Verification gate

After Higgsfield returns each generated image, BEFORE adding to any artifact OR sending to Pinion:

1. Load the image into Florence's multimodal context — paste the URL in chat so Cowork's multimodal Claude loads it natively. Narrate: "Looking at #{N} before I include it."
2. Visually inspect against: aspect ratio (1:1 main+listing / 16:9 A+), product fidelity vs reference, technique landed (visibly), background appropriate, no clipart leak (anti-clipart rules), no text-on-main-image (TOS), no model faces.
3. If anything fails → re-prompt with specific fix + regenerate. Cap at 3 attempts per concept; surface honestly on attempt 4.
4. Only after ALL concepts pass → proceed to artifact emission.

The eye trumps the score. Even if the rubric said 100/100, if visual inspection finds a wrong product or off-aspect output, it fails the gate.

Base64 embedding

Higgsfield URLs are temporary AND Cowork's artifact iframe sandbox blocks external image loads in many builds. For every verified concept:

1. HTTP GET the Higgsfield URL → fetch image bytes
2. Detect MIME type from response headers (typically image/png)
3. Base64-encode the bytes
4. Substitute {{image-src}} (or {{winner-image-src}} for tests) with data:image/png;base64,<encoded> — NOT the raw Higgsfield URL

This makes the artifact self-contained — survives sandbox + URL expiry. ~1-3 MB per image is fine for Cowork.

The live product image ({{product-image-url}} in concepts.html hero strip / dossier head / cockpit product cards) stays as the live Amazon CDN URL — that's permanent and not affected.