52 main-image tactics
A tactic library for Amazon Image 1 concepts, including when each tactic fits and when it should be avoided.
A downloadable workflow for brand owners and operators who want sharper Amazon hero-image tests: 52 main-image tactics, reusable AI prompts, and a ProductPinion-ready test brief structure.
No email wall. No repo spelunking. Download the zip, install the folder, and run it on your next Image 1 test.
Built for operators who need useful creative directions, clean prompts, and a test structure before a designer touches the final file.
The download is built for real operators: enough structure to stop guessing, enough guardrails to keep the test clean, and no requirement to connect to ProductPinion before the creative thinking is useful.
A tactic library for Amazon Image 1 concepts, including when each tactic fits and when it should be avoided.
Reusable JSON prompts and customer-facing wrappers that preserve product identity instead of turning every image into generic concept art.
A practical ProductPinion sequence for Search Simulation, Image Split Test, Ranked Test, and winner confirmation.
Unzip the folder, keep the reference files next to the main skill file, then place the full folder in your local agent skills directory.
Common install paths:
~/.claude/skills/productpinion-main-image-lab/
~/.codex/skills/productpinion-main-image-lab/
Validate after install:
cd ~/.codex/skills/productpinion-main-image-lab
python3 scripts/validate_tactics.py
Expected:
Validated 52 ProductPinion tactics.
After installing, attach a product image and paste this prompt into your AI workspace.
Use ProductPinion Main Image Lab on this Amazon product image.
Pick the 5 best-fit main-image tactics, explain why they fit, create one reusable JSON image prompt for each tactic, and prepare a ProductPinion test brief comparing the original image against the strongest challenger concepts.
Do not invent claims, certifications, awards, pack counts, ingredients, or variants. Preserve the product identity and call out anything that would need human cleanup before Amazon use.
If image generation is available, generate one separate concept image per selected tactic. Keep each variant isolated so ProductPinion can test the tactic cleanly.
The point is not making prettier pictures. The point is isolating visual hypotheses so you can learn what actually earns the click.
Run a ProductPinion Search Simulation against 3-6 real competitors to see what your current hero image is losing against.
Create 3-5 hero-image variants that each test a different visual hypothesis, not five versions of the same idea.
Use Image Split Test for original versus challenger, Ranked Test for ordering 3+ concepts, then re-test the winner in search context.
Get the full folder: tactic library, prompts, workflow files, validation script, and installation README.