# Ecommerce Product Imagery

> Provider-independent ecommerce product imagery production for marketplace listings, product pages, hero shots, lifestyle scenes, comparison charts, infographics, variant/packaging images, virtual try-on or placement mockups, ads, and AI-assisted product visuals. Use when an agent must plan, generate, edit, review, localize, package, or QA truthful commercial product images while managing SKU cont…

- **Type:** Skill
- **Install:** `agentstack add skill-calesthio-generative-media-skills-ecommerce-product-imagery`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [calesthio](https://agentstack.voostack.com/s/calesthio)
- **Installs:** 0
- **Category:** [Developer Tools](https://agentstack.voostack.com/c/developer-tools)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [calesthio](https://github.com/calesthio)
- **Source:** https://github.com/calesthio/generative-media-skills/tree/main/skills/production/content-formats/ecommerce-product-imagery
- **Website:** https://github.com/calesthio/OpenMontage

## Install

```sh
agentstack add skill-calesthio-generative-media-skills-ecommerce-product-imagery
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Ecommerce Product Imagery

Use this skill to produce commercially usable product imagery, not just attractive pictures. The governing rule is: the image must help a buyer understand the actual purchasable product without inventing features, quantities, sizes, colors, materials, packaging, certifications, results, endorsements, or included accessories.

## Operating principles

Separate three categories before creating anything:

1. **Documented product fact**: confirmed by physical sample, client PDP, packaging, SKU sheet, spec sheet, or approved brand guide.
2. **Creative direction**: lighting, setting, camera, styling, props, layout, crop, visual hierarchy, and mood.
3. **Regulated or platform-sensitive claim**: performance claims, health/beauty outcomes, environmental claims, comparative superiority, pricing, discounts, guarantees, safety, awards, certifications, shipping promises, and "before/after" implications.

Do not let category 2 change category 1. Do not render category 3 unless the client supplies substantiation and approves the exact wording and visual implication.

Label your own assertions while working:

- **Documented fact**: cite the source artifact or page.
- **Production heuristic**: explain that it is a practical rule of thumb.
- **Volatile platform fact**: include the date checked and tell the user it must be re-checked before upload or ad launch.
- **Legal/compliance risk**: do not give legal advice; escalate to client, counsel, or marketplace review.

## First response workflow

For any ecommerce image request, collect or infer the minimum viable brief:

- Product identity: brand, product name, SKU/ASIN/GTIN, variant names, bundle quantity, packaging version, included accessories.
- Channel: Amazon, Shopify/PDP, Google Merchant Center/Shopping, Meta ads, TikTok ads, retail media, email, marketplace carousel, A+ content, social organic, print, or internal concepting.
- Deliverables: count, orientation, aspect ratio, transparent/white background, lifestyle, infographic, comparison, size chart, localized versions, ad versions, thumbnails.
- Source material: packshots, packaging flats, label art, CAD/3D files, brand guide, prior approved imagery, physical sample photos, model/property releases.
- Truth constraints: dimensions, colors, material finish, ingredients, claims, certifications, warranty, target market, restricted categories.
- Use of generation/editing: full generation, background replacement, compositing, cleanup, upscaling, virtual placement, virtual try-on, synthetic models, or concept-only.
- Approval path: who approves product accuracy, claims, legal/compliance, marketplace upload, and final art.

If the user asks for production before this information exists, proceed only with a clearly marked concept or placeholder plan. Do not create final commercial imagery when product truth cannot be verified.

## Product truth and SKU continuity

Treat the SKU as the continuity anchor. Every image in a set must depict the same sellable item unless the deliverable intentionally compares variants.

Maintain a small product truth table:

| Field | Required source | QA check |
|---|---|---|
| SKU/variant | SKU sheet, PDP, package, order system | Same variant name and color across all images |
| Color/material | sample photo under neutral light, brand spec, material swatch | No hue shift, texture swap, or invented finish |
| Scale/dimensions | dimensions sheet, model/hand reference, known object | Product does not appear misleadingly large or small |
| Packaging | approved label art, packaging flat, sample photo | Correct front/back/side, barcode/claims if visible |
| Included items | SKU contents or bundle list | No extra accessories, props, refills, chargers, stands, or ingredients implied as included |
| Claims | approved claims deck, substantiation, legal approval | Exact wording and visual implication approved |
| Compliance marks | certification documentation | Marks are real, current, and used according to owner rules |

Never invent:

- logos, label text, nutrition/drug/supplement facts, certifications, seals, badges, QR codes, barcodes, serial numbers, warning labels, or regulatory markings;
- product performance outcomes such as "whitens teeth," "cures acne," "non-toxic," "100% biodegradable," "clinically proven," or "FDA approved" without explicit substantiation;
- variant colors or bundle quantities;
- endorsements, awards, press mentions, reviews, ratings, or marketplace badges.

If an image model cannot reproduce text/logo/label details reliably, use compositing: generate or photograph the scene with a blank/neutral product surface, then overlay approved vector/raster label art in a conventional editor or design tool. QA the final pixels at 100% and at marketplace thumbnail size.

## Reference and capture strategy

Prefer real references over prompt descriptions. Request:

- neutral front, back, side, top, bottom packshots;
- one angled hero reference;
- close-ups of distinctive materials, seams, ports, ingredients, texture, stitching, finish, cap, applicator, or hardware;
- label or packaging art as flat files, not only perspective photos;
- a ruler, hand, model, or known object when scale matters;
- color target or neutral-light reference when color accuracy matters;
- separate references for every variant and packaging version.

For AI-assisted workflows, use references narrowly:

- **Identity reference**: preserves shape, silhouette, label placement, distinctive features.
- **Material reference**: preserves finish, texture, transparency, fabric, metal, glass, plastic, liquid, powder, or food surface.
- **Scene reference**: guides environment and lighting only.
- **Layout reference**: guides infographic/chart composition.

Do not use a competitor's image as a style or composition reference in a way that copies distinctive trade dress, layout, proprietary packaging, or campaign art. Use broad category language instead.

## Shot types and production guidance

### Marketplace primary image

Goal: a clear sellable product identifier.

- Use a neutral or pure white background when required by the destination.
- Show only what is included with the purchase.
- Avoid props, lifestyle context, decorative overlays, badges, ratings, or unsupported text unless the specific marketplace and category permit them.
- Keep product geometry natural; avoid wide-angle distortion.
- Leave margins for marketplace cropping while keeping the product large enough to recognize.
- For mirrored/glossy materials, preserve real reflections without changing product color or shape.

Volatile platform note verified 2026-07-11: Amazon product image requirements are category- and locale-sensitive; Seller Central image rules include main-image restrictions and minimum technical requirements. Re-check the active Seller Central help page for the target marketplace before upload: https://sellercentral.amazon.com/help/hub/reference/external/G1881

### Secondary/gallery images

Use secondary images to answer buyer questions:

- scale in hand, on body, in room, on desk, inside bag, or next to measured objects;
- material/texture macro;
- open/closed states;
- included components laid out;
- packaging front/back;
- benefit diagram with substantiated claims;
- how-to-use sequence;
- variant comparison;
- compatibility chart;
- before/after only when substantiated, truthful, and allowed by the platform/category.

### Lifestyle images

Lifestyle scenes may be aspirational, but product facts remain literal.

- Match plausible use context, climate, demographics, and geography.
- Keep the product readable, not hidden by props.
- Do not imply unapproved use cases, medical outcomes, safety performance, environmental benefits, or social status guarantees.
- Ensure props are not mistaken as included items. Add captioning or layout clarity when needed.
- For food, cosmetics, supplements, children's products, medical devices, and safety products, escalate riskier scenes for compliance review.

### Hero shots and ads

Hero imagery may exaggerate drama, not facts.

- Use lighting, composition, motion cues, scale of set, and background to create premium feel.
- Do not enlarge the product relative to the buyer's likely expectation unless the image is clearly symbolic.
- If using splashes, particles, ingredients, smoke, flames, ice, flowers, or scientific graphics, ensure they communicate approved attributes only.
- For paid ads, re-check platform creative and claims policies at launch time; ad review standards change.

Volatile platform notes verified 2026-07-11:

- Google Merchant Center image and product data requirements are enforced through attributes such as `image_link`; check current rules before feed submission: https://support.google.com/merchants/answer/6324350
- Google Merchant Center currently requires generative-AI product images to retain embedded AI-source metadata such as IPTC `DigitalSourceType`; do not strip provenance metadata when exporting or optimizing feed images. Re-check before submission: https://support.google.com/merchants/answer/6324350
- Meta's Advertising Standards cover misleading claims, personal attributes, before/after and other sensitive content; check current standards before ad launch: https://transparency.meta.com/policies/ad-standards/
- TikTok advertising policies cover misleading, exaggerated, and prohibited claims and landing-page consistency; check current rules for region/category: https://ads.tiktok.com/help/article/tiktok-advertising-policies-ad-creatives-landing-page

### Infographics, comparison charts, and callout images

Use infographics to clarify, not to decorate.

- Anchor every claim to an approved source.
- Use exact units, dimensions, materials, compatibility, counts, and version names.
- Avoid "best," "#1," "clinically proven," "non-toxic," "eco-friendly," "safe," "guaranteed," and similar claims unless substantiated and approved.
- If comparing with competitors, require client/legal approval for the comparison basis, date, methodology, and jurisdictions.
- Make chart axes, icons, and callouts legible at mobile thumbnail scale.
- Preserve accessibility: do not rely on color alone; use text labels and contrast.

### Variant, bundle, and packaging images

- Create a variant matrix before rendering: variant name, color, finish, size, scent/flavor, packaging art, quantity, and any unique claims.
- Keep camera angle, lighting, crop, and scale consistent across variants unless intentionally showing a functional difference.
- Do not "average" variants in prompts. Render or composite each variant from its own reference.
- For bundles, show exactly the quantity and components included. If the image includes non-included props, visually separate them.
- For localized packaging, do not translate legal labels, warnings, ingredients, or regulated information without client-provided localized art.

### Virtual try-on, synthetic models, and placement mockups

Use these as fit/context aids, not proof of universal fit.

- Require model consent/releases for real identifiable people.
- For synthetic people, avoid implying endorsement by a real person or protected-class targeting.
- For apparel/accessories, preserve actual pattern scale, seam placement, hardware, opacity, and drape as far as the source material allows.
- For furniture/home placement, verify dimensions and show room scale honestly.
- Mark approximate visualization when exact fit, color, or scale cannot be guaranteed.
- Escalate high-risk uses: children, medical/health products, body-image transformations, beauty outcomes, biometric/face edits, regulated uniforms, culturally sensitive garments, or private property.

## AI-generated and AI-edited image boundaries

Before using generation, classify the task:

- **Safe for generation**: background, lighting mood, generic props, abstract ingredient atmosphere, empty rooms, concept boards, non-final exploratory ads.
- **Use generation plus compositing**: product in lifestyle scene, hero image with accurate label, packaging variants, infographic backgrounds, ad layouts.
- **Use real photography/3D or client assets first**: main marketplace image, exact color/material, regulated packaging, dense label text, legally significant marks, technical dimensions.
- **Do not generate without escalation**: claims requiring evidence, identifiable real people without release, celebrity/public figure likeness, private property/trade dress, medical/legal/financial promises, competitor comparisons, platform badges/reviews.

Generation prompts should preserve product truth by separating fixed facts from creative variables:

```text
Fixed product facts:
- [SKU name, color, finish, exact included components, dimensions, packaging version]
- Preserve silhouette, label placement, material, and scale from references.

Creative direction:
- [camera angle, lens feel, lighting, surface, background, styling]

Negative constraints:
- No extra accessories, no changed logo, no invented text, no badges, no claims, no altered color, no deformation, no additional variants.

Post-production plan:
- Composite approved label/logo from source file; QA against reference; export platform crops.
```

Record model/tool, version if available, date, prompt, references, seed/settings if available, edits, and human approvals. This audit trail is useful for reproducibility, rights review, and correcting defects.

## Claims, compliance, and escalation

Use this decision rule: if an image would cause a reasonable buyer to believe something about performance, safety, health, ingredients, sustainability, certification, pricing, scarcity, compatibility, or endorsement, treat it as a claim.

Escalate to client/counsel/compliance before final use when imagery includes:

- health, medical, wellness, supplement, beauty, weight-loss, body-transformation, child-safety, financial, insurance, legal, or environmental claims;
- before/after images or outcome simulations;
- comparative claims against competitors;
- certification marks, government agency references, seals, awards, reviews, marketplace badges, "Amazon's Choice," star ratings, press logos, or influencer endorsements;
- "free," discount, price, guarantee, warranty, delivery, return, inventory, or limited-time claims;
- user-generated content, testimonials, influencer/affiliate content, or material connections that may require disclosure;
- altered model bodies, skin, results, or fit;
- restricted categories or regional legal differences.

Official FTC references verified 2026-07-11:

- Advertising claims require a reasonable basis and substantiation before dissemination: https://www.ftc.gov/business-guidance/advertising-marketing/advertising-claims
- Endorsements and testimonials must be truthful and disclose material connections where required: https://www.ftc.gov/business-guidance/advertising-marketing/endorsements-influencers-reviews
- Environmental marketing claims require care and substantiation; check current Green Guides/status before using green claims: https://www.ftc.gov/business-guidance/advertising-marketing/environmental-marketing

Do not provide legal advice. Say what needs approval and why.

## Rights, releases, and provenance

For final commercial imagery, track provenance:

- who supplied each asset;
- license/ownership status;
- model/property release status;
- AI tool and settings;
- post-production edits;
- approval timestamp and approver.

Require explicit approval or counsel review before using:

- recognizable people, tattoos, uniforms, private homes, distinctive architecture, vehicles, license plates, pets with owners, artworks, branded props, third-party logos, screenshots, app UIs, maps, copyrighted packaging, celebrity lookalikes, or public-figure likenesses;
- AI-generated people that resemble real individuals;
- stock assets outside license terms;
- marketplace badges, platform logos, or partner marks.

Helpful rights/provenance refer

…

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [calesthio](https://github.com/calesthio)
- **Source:** [calesthio/generative-media-skills](https://github.com/calesthio/generative-media-skills)
- **License:** MIT
- **Homepage:** https://github.com/calesthio/OpenMontage

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-calesthio-generative-media-skills-ecommerce-product-imagery
- Seller: https://agentstack.voostack.com/s/calesthio
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.