| name | ads-generate |
| description | AI image generation for paid ad creatives. Reads campaign-brief.md and brand-profile.json to produce platform-sized ad images using Google Gemini 3 Pro Image / Imagen 4 via the google-genai SDK. Triggers on: generate ads, create images, make ad creatives, generate visuals, create ad images, generate campaign images, make the images, generate from brief. Use when this capability is needed. |
Ads Generate: AI Ad Image Generator
Generates platform-sized ad creative images from your campaign brief and
brand profile. Uses Google Gemini / Imagen 4 via scripts/generate_image.py.
Quick Reference
| Command | What it does |
|---|
/ads generate | Generate all images from campaign-brief.md (default quality: pro) |
/ads generate --platform meta | Generate Meta assets only |
/ads generate --quality high | Force Imagen 4 Ultra for photoreal product shots |
/ads generate --quality fast | High-volume variants with Gemini 2.5 Flash Image |
/ads generate --prompt "text" --ratio 9:16 | Standalone generation without brief |
Environment Setup
Required before running:
If GOOGLE_API_KEY is not set, this skill will display setup instructions
and stop. It will never fail silently.
Optional alternative providers (OpenAI, Stability AI, Replicate) are documented
in ~/.claude/skills/ads/references/image-providers.md. Switch via
ADS_IMAGE_PROVIDER env var.
Quality Presets
The --quality flag controls which Google model is used:
| Preset | Model | Cost/img | Use |
|---|
pro (default) | Gemini 3 Pro Image Preview | $0.134 | Hero creatives, key visuals |
high | Imagen 4 Ultra | $0.06 | Photoreal product shots |
flash | Gemini 3.1 Flash Image Preview | $0.067 | Multi-reference editing |
fast | Gemini 2.5 Flash Image | $0.039 | Volume / A-B variants |
Process
Step 1: Verify Prerequisites
Check that GOOGLE_API_KEY is set. If not, display setup instructions:
export GOOGLE_API_KEY="your-key"
# Get a key at https://aistudio.google.com/app/apikey
Step 2: Locate Source Files
Check for:
campaign-brief.md → primary source for prompts and dimensionsbrand-profile.json → brand color/style injection (optional but recommended)
If campaign-brief.md is found: Use ## Image Generation Briefs section as the
generation job list.
If no campaign-brief.md: Enter standalone mode (Step 2b).
Step 2b: Standalone Mode
Ask the user:
- Generation prompt (what should the image show?)
- Target platform (to set correct dimensions)
- Quality preset (pro/high/flash/fast, default: pro)
- Output filename (optional)
Then skip to Step 5.
Step 3: Read Provider Config
Load ~/.claude/skills/ads/references/image-providers.md to confirm:
- Current model pricing (show user the cost estimate)
- Rate limits for current tier
- Aspect ratio support per model
Step 4: Read Platform Specs
For each platform in the campaign brief, load the relevant spec reference:
~/.claude/skills/ads/references/meta-creative-specs.md~/.claude/skills/ads/references/google-creative-specs.md~/.claude/skills/ads/references/tiktok-creative-specs.md~/.claude/skills/ads/references/linkedin-creative-specs.md~/.claude/skills/ads/references/youtube-creative-specs.md~/.claude/skills/ads/references/microsoft-creative-specs.md
Step 5: Select Quality Per Asset
Map each brief entry to the right --quality preset based on its purpose:
| Asset type | Preset | Why |
|---|
| Hero / key visual | pro | Maximum reasoning + text rendering |
| Product packshot | high | Imagen 4 Ultra is best-in-class for photoreal products |
| Brand-reference edit | flash | Only Gemini 3.x accepts reference images |
| Many variants of same concept | fast | 3–4× cheaper per image |
Step 6: Spawn Visual Designer Agent
Spawn the visual-designer agent using the Task tool with context: fork,
passing the selected quality preset and brand profile.
The agent will:
- Parse the image generation briefs from campaign-brief.md
- Inject brand colors and mood from brand-profile.json into each prompt
- Invoke
python ~/.claude/skills/ads/scripts/generate_image.py per asset
- Save to
./ad-assets/[platform]/[concept]/ directory structure
- Write
generation-manifest.json
Example call the agent runs:
python ~/.claude/skills/ads/scripts/generate_image.py \
"Minimalist SaaS dashboard screenshot, dark UI with teal accents, clean data viz" \
--ratio 16:9 \
--quality pro \
--output ./ad-assets/linkedin/concept-1/hero-1920x1080.png \
--json
Step 7: Validate with Format Adapter
After the visual-designer completes, spawn the format-adapter agent
with context: fork to validate dimensions and report missing formats.
Step 8: Quality Gate
Use Claude vision to assess each generated image against the brief (score 1 to 10
on brand alignment, composition, platform fit). If any image scores below 6,
regenerate once with an adjusted prompt.
Quality Gate Rubric:
- 9-10: Professional quality, brand-aligned, platform-optimized, no issues
- 7-8: Good quality, minor composition or brand alignment improvements possible
- 5-6: Acceptable but needs regeneration. Text readability issues, poor composition, or brand mismatch
- Below 5: Reject. Regenerate with adjusted prompt
Step 9: Aggregate Costs
Sum costs from each generation's JSON output (script emits quality and
model per image). Compute total using the pricing table in
image-providers.md and include in generation-manifest.json.
Step 10: Report Results
Present a summary:
Generation complete:
Generated assets:
✓ ./ad-assets/meta/concept-1/feed-1080x1350.png [quality: pro, $0.134]
✓ ./ad-assets/tiktok/concept-1/vertical-1080x1920.png [quality: fast, $0.039]
✗ ./ad-assets/google/concept-1/landscape-1200x628.png [error: SAFETY]
Format validation: See format-report.md
Cost: $[N] total creative spend (X images)
Next steps:
1. Review assets in ./ad-assets/
2. Check format-report.md for any missing formats
3. Upload to your ad platform managers
Cost Transparency
Before generating, estimate and show the cost:
- Count the number of image briefs in campaign-brief.md
- Multiply by the per-image cost of the chosen preset
- If total >$1.00, ask for confirmation before proceeding
- Suggest dropping non-hero assets to
--quality fast to reduce spend
Standalone Mode (No campaign-brief.md)
When running without a campaign brief:
Platform target → dimensions used:
meta-feed → 1080×1350 (4:5)
meta-reels → 1080×1920 (9:16)
tiktok → 1080×1920 (9:16)
google-pmax → 1200×628 (1.91:1)
linkedin → 1080×1080 (1:1)
youtube → 1280×720 (16:9)
youtube-short → 1080×1920 (9:16)
Invoke the script directly:
python ~/.claude/skills/ads/scripts/generate_image.py \
"<prompt>" --ratio <ratio> --quality <preset> --output <file>
Reference Files
~/.claude/skills/ads/references/image-providers.md: model config, pricing, limits, presets~/.claude/skills/ads/references/[platform]-creative-specs.md: per-platform specs~/.claude/skills/ads/references/brand-dna-template.md: brand injection schema
Source: bendervintimilla/skills-AI-marketing-tool — distributed by TomeVault.
