Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate high-quality images and assets for components (hero sections, features, transformations) using Replicate's FLUX 1.1 Pro model.
.claude/skills/aiskillstore-generate-assets/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -71% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 18% | 0% |
You are the Asset Creator. You generate high-quality images and visual assets for website components using AI.
public/assets/images/ or appropriate subdirectories.⚠️ IMPORTANT: Before using this skill, ensure REPLICATE_API_TOKEN is set in your environment variables (.env, .env.local, or .env.production).
and Dev Requirement: pnpm add sharp -D
Script: .claude/skills/generate-assets/scripts/generate-asset.ts
Usage:
bashpnpm run script .claude/skills/generate-assets/scripts/generate-asset.ts "<prompt>" "[aspect-ratio]" "[filename]" "[folder]" "[asset-type]" "[transparent]"
Parameters:
prompt: Base description of the image to generate. Will be automatically enhanced with modern UI elements, gradients, patterns, and theme-aware styling.aspect-ratio: Optional aspect ratio. If omitted and file exists, uses existing image dimensions. Otherwise defaults by asset type.filename: Name of the file without extension (default: auto-generated from prompt).folder: Subfolder in public/assets/images/ (default: root of public/assets/images/).asset-type: Optional type: "hero", "feature", "transformation", "foreground", "other" (default: "other"). Affects prompt enhancement and default aspect ratio.transparent: Optional "true" or "false" (default: "false"). If true, removes background for transparent assets.Examples:
bash# Hero section with automatic enhancement pnpm run script .claude/skills/generate-assets/scripts/generate-asset.ts "modern dashboard interface" "16:9" "hero-dashboard" "hero" "hero" "false" # Feature showcase (will auto-detect size if file exists) pnpm run script .claude/skills/generate-assets/scripts/generate-asset.ts "AI content generation" "" "feature-ai" "features" "feature" "false" # Foreground asset with transparent background pnpm run script .claude/skills/generate-assets/scripts/generate-asset.ts "decorative stars" "" "stars" "foreground" "foreground" "true"
When asked to "Generate an asset for X" or "Create an image for the hero section":
REPLICATE_API_TOKEN is set. If not, inform the user.16:9 or 21:91:1 or 4:316:9The script automatically enhances prompts with:
black-forest-labs/flux-1.1-probria/remove-background (for transparent assets)public/assets/images/[folder]/[filename].webpFor advanced configuration and model details, see reference.md.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 24,670 | 2,017 | -92% | 1 | 1 | 0% | 4,893 | 1,724 | -65% | 0 | 0 | — |
case-01 | fail→pass | 29,873 | 3,094 | -90% | 1 | 1 | 0% | 6,180 | 1,776 | -71% | 0 | 0 | — |
case-03 | fail→pass | 23,479 | 2,429 | -90% | 1 | 1 | 0% | 4,984 | 1,838 | -63% | 0 | 0 | — |
case-04 | pass→pass | 7,171 | 3,113 | -57% | 1 | 1 | 0% | 1,075 | 1,913 | +78% | 0 | 0 | — |
case-05 | fail→pass | 13,686 | 2,744 | -80% | 1 | 1 | 0% | 2,063 | 1,785 | -13% | 0 | 0 | — |
case-06 | pass→pass | 32,928 | 2,571 | -92% | 1 | 1 | 0% | 6,162 | 1,747 | -72% | 0 | 0 | — |
case-07 | fail→pass | 9,298 | 1,899 | -80% | 1 | 1 | 0% | 1,414 | 1,674 | +18% | 0 | 0 | — |
case-08 | pass→pass | 8,123 | 1,895 | -77% | 1 | 1 | 0% | 1,380 | 1,671 | +21% | 0 | 0 | — |
case-09 | pass→pass | 8,505 | 1,741 | -80% | 1 | 1 | 0% | 1,312 | 1,624 | +24% | 0 | 0 | — |
case-10 | fail→pass | 9,577 | 1,379 | -86% | 1 | 1 | 0% | 1,483 | 1,622 | +9% | 0 | 0 | — |
case-11 | fail→pass | 2,018 | 1,433 | -29% | 1 | 1 | 0% | 289 | 1,583 | +448% | 0 | 0 | — |
case-12 | fail→pass | 11,298 | 1,896 | -83% | 1 | 1 | 0% | 1,806 | 1,641 | -9% | 0 | 0 | — |
case-13 | fail→pass | 4,471 | 1,753 | -61% | 1 | 1 | 0% | 759 | 1,655 | +118% | 0 | 0 | — |
case-14 | pass→pass | 5,809 | 1,140 | -80% | 1 | 1 | 0% | 810 | 1,549 | +91% | 0 | 0 | — |
case-15 | fail→pass | 6,777 | 1,979 | -71% | 1 | 1 | 0% | 1,131 | 1,710 | +51% | 0 | 0 | — |
case-16 | fail→pass | 8,593 | 3,658 | -57% | 1 | 1 | 0% | 1,358 | 2,080 | +53% | 0 | 0 | — |
case-17 | pass→pass | 9,478 | 2,980 | -69% | 1 | 1 | 0% | 1,471 | 1,898 | +29% | 0 | 0 | — |
case-18 | fail→pass | 6,723 | 1,326 | -80% | 1 | 1 | 0% | 992 | 1,559 | +57% | 0 | 0 | — |
case-19 | pass→pass | 3,560 | 1,142 | -68% | 1 | 1 | 0% | 450 | 1,527 | +239% | 0 | 0 | — |
case-20 | pass→pass | 3,207 | 3,364 | +5% | 1 | 1 | 0% | 560 | 1,982 | +254% | 0 | 0 | — |
case-21 | pass→pass | 4,282 | 4,148 | -3% | 1 | 1 | 0% | 749 | 2,061 | +175% | 0 | 0 | — |
case-22 | pass→pass | 2,635 | 2,585 | -2% | 1 | 1 | 0% | 385 | 1,786 | +364% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +55 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.