visual-prompting
Skill0 starsChanged 7 months agoMITView on GitHub
Generate and iterate ecommerce-ready image/video prompts for product photography, website banners, Meta ads, and short-form video. Use when users ask for better prompts, competitor-style visuals, asset specs, or conversion-focused creative iteration for Shopify and paid ads
In this repository13 other files it gives its agents.
- deploy-pipelineSkill
- ecommerce-builderSkill
- hackathon-copilotSkill
- know-meSkill
- memory-managementSkill
- obsidian-memory-syncSkill
- polymarket-multi-agentSkill
- polymarket-vps-opsSkill
- researcherSkill
- self-healingSkill
- self-improvementSkill
- strategy-preflightSkill
- trade-journalSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Use the 8-layer prompt architecture below
- Distinctive: line 52
Keep product geometry/materials consistent with references
- Very specific: line 56
Video generation: Default to fal-ai/kling-video/v3/pro/image-to-video
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Visual Prompting
Production-grade prompts and creative iterations for ecommerce assets.
Workflow
- Classify the visual request (product photo, website asset, static ad, video ad, social, brand conceptual)
- Build a concise creative brief
- Use the 8-layer prompt architecture below
- Generate 1-3 variants first, then run Evaluate -> Diagnose -> Adjust
- Save winning prompts and record wins/gotchas in memory
Phase Routing
- Product photos -> Phase 3 (studio/lifestyle shots)
- Website hero/banner assets -> Phase 4 (composition for text overlay)
- Meta static ads -> Phase 5 (bold, readable at small size)
- Video ads / image-to-video -> Phase 6 (use fal.ai kling-video endpoint)
- Prompt debugging -> Phase 2 + Phase 7 (iterate on failures)
8-Layer Prompt Template
[Subject]: {product/scene description}
[Style]: {photography style - editorial, lifestyle, flat lay, studio}
[Lighting]: {natural, studio softbox, golden hour, dramatic}
[Composition]: {rule of thirds, centered, negative space for text}
[Color]: {palette/mood - warm, cool, brand colors}
[Camera]: {lens - 85mm portrait, 35mm wide, macro}
[Context]: {background/environment}
[Constraints]: {no text, no hands, specific dimensions, artifact prevention}Platform Dimensions Quick Reference
| Asset | Dimensions | Notes |
| Shopify hero | 1920x800 | Leave center 600px for headline overlay |
| Product photo | 1:1 (2048x2048) | White/transparent bg for main, lifestyle for gallery |
| Meta feed ad | 1:1 (1080x1080) | Bold, readable at small size |
| Meta story ad | 9:16 (1080x1920) | Product in bottom 2/3, top 1/3 for headline |
| Collection banner | 1600x400 | Subtle, don't compete with product grid |
| Email hero | 600x300 | Keep under 100KB |
Rules for Output Quality
- Keep product geometry/materials consistent with references
- Require usable composition for placement (negative space for headline/CTA)
- Prefer platform-native dimensions
- Include artifact-prevention constraints for FLUX endpoints
- Video generation: Default to
fal-ai/kling-video/v3/pro/image-to-video
Failure Modes & Fixes
| Problem | Fix |
| AI generates text/watermarks | Add "no text, no watermarks, no logos" to prompt |
| Product geometry distorted | Be explicit about shape: "cylindrical bottle", "rectangular box" |
| Wrong material/finish | Specify exactly: "matte black anodized aluminum" not "black metal" |
| Composition unusable (no CTA space) | Specify "large negative space on [left/right/top] for text overlay" |
| Inconsistent across variants | Use same seed + keep style/lighting/camera params identical |
| Video endpoint returns stills | Check endpoint version, try different motion parameters |
Integration Points
- ecommerce-builder: Generate assets for all store phases (hero, PDP, collection banners)
- researcher: Research competitor visual styles before generating prompts
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.