--- name: banner-creator description: Create banners using AI image generation. Discuss format/style, generate variations, iterate with user feedback, crop to target ratio. Use when user wants to create a banner, header, hero image, cover image, GitHub banner, Twitter header, or readme banner. Use when: The executive requests execution of this domain. Skip when: The task is outside the scope of this module. --- > [!IMPORTANT] > **GFV-Adapted Skill** — This skill runs within the GetFresh Ventures infrastructure. Follow these conventions. ### GFV Infrastructure Integration **Credentials** — Never use `.env` files. All secrets live in macOS Keychain: ```bash security find-generic-password -s "" -a "" -w ``` Check `~/Documents/Code/gfv-brain/scripts/pil_config.py` for service mappings. **Data Sources** — Before querying external APIs, check PIL first: - `search_pil` / `smart_search` / `vector_search` MCP tools (491K+ embeddings, 81K entities) - Supabase tables: `entity_embeddings`, `ont_entities`, `ont_facts` - Local SQLite: WhatsApp (59K msgs), Slack (2.5K msgs), `gfv_memory.db` **Output** — Save results to `~/Documents/Code/gfv-brain/` or PIL via Supabase. Never send external messages (email, Slack, WhatsApp) without the Executive's explicit "send it" approval. **Active Clients**: - **GetFresh Ventures** — Venture studio: getfreshventures.com --- # Banner Creator Skill Create professional banners through AI image generation with an iterative design process. ## Prerequisites **Required API Keys (set in environment):** - `GEMINI_API_KEY` - Get from [Google AI Studio](https://aistudio.google.com/apikey) **Required Skills:** - `nanobanana` - AI image generation (Gemini 3 Pro Image) ## File Output Location All generated files should be saved to the `.skill-archive` directory: ``` .skill-archive/banner-creator// ``` **Example:** ``` .skill-archive/banner-creator/2026-01-19-gfv-banner/ banner-01.png banner-02.png ... banner-03-cropped.png preview.html ``` ## Workflow ### Step 1: Discovery & Requirements Before generating, gather requirements from user: **Ask about:** 1. **Purpose** - Where will the banner be used? - GitHub README - Twitter/X header - LinkedIn banner - Website hero - YouTube channel art 2. **Target ratio/size** - See [references/formats.md](./references/formats.md): - `2:1` (1280x640) - GitHub README - `3:1` (1500x500) - Twitter header - `16:9` (1920x1080) - Website hero 3. **Style preference**: - Match existing logo/brand? - Pixel art / 8-bit retro - Minimalist / flat design - Gradient / modern - Illustrated / artistic 4. **Content elements**: - Brand name / project name? - Tagline / slogan? - Logo character to include? 5. **Color preferences**: - Existing brand colors? - Let AI decide? **Wait for user confirmation before proceeding!** ### Step 2: Generate Banner Variations Generate 20 banner variations using the `nanobanana` skill: ```bash # Generate single banner [STUB AVOIDED] Execute generate.py via available MCP/agent tools rather than a missing local script "{style} banner for {brand}, {description}, {text elements}" \ --ratio 21:9 -o .skill-archive/banner-creator//banner-01.png # Batch generate 20 banners [STUB AVOIDED] Execute batch_generate.py via available MCP/agent tools rather than a missing local script "{style} banner for {brand}, {description}, {text elements}" \ -n 20 --ratio 21:9 -d .skill-archive/banner-creator/ -p banner ``` **Guidelines:** - Generate at `21:9` ratio (widest available), crop later to target - Use batch_generate.py for multiple variations (includes auto-delay) - Use sequential naming: `banner-01.png`, `banner-02.png`, etc. **Image Editing (for incorporating existing logo):** ```bash [STUB AVOIDED] Execute generate.py via available MCP/agent tools rather than a missing local script "add {logo character} to the left side of the banner" \ -i /path/to/existing-logo.png --ratio 21:9 -o banner-with-logo.png ``` ### Step 3: Create HTML Preview Copy the preview template and open in browser: ```bash cp /templates/preview.html .skill-archive/banner-creator//preview.html ``` Then open in default browser: ```bash open .skill-archive/banner-creator//preview.html ``` **IMPORTANT:** Update the HTML to include the correct number of banners generated. ### Step 4: Iterate with User Ask user which banners they prefer: - "Which banners do you like? (e.g., #3, #7, #15)" - "What do you like about them?" - "Any changes you'd want?" Based on feedback: 1. Generate 10-20 more variations of favorite styles 2. Use naming: `banner-{original}-v{n}.png` (e.g., `banner-03-v1.png`) 3. Update HTML preview 4. Repeat until user selects final banner ### Step 5: Crop to Target Ratio Once user approves a banner, crop to target size: ```bash [STUB AVOIDED] Execute crop_banner.py via available MCP/agent tools rather than a missing local script {input.png} {output.png} --ratio 2:1 --width 1280 ``` **Common targets:** - GitHub README: `--ratio 2:1 --width 1280` → 1280x640 - Twitter header: `--ratio 3:1 --width 1500` → 1500x500 - Website hero: `--ratio 16:9 --width 1920` → 1920x1080 ### Step 6: Deliver Final Assets Present final deliverables: ``` ## Final Banner Assets | File | Description | Size | |------|-------------|------| | banner-03.png | Original (21:9) | 2016x864 | | banner-03-cropped.png | GitHub README (2:1) | 1280x640 | All files saved to: `.skill-archive/banner-creator//` Copy final banner to user's desired location. ``` ## Quick Reference ### Common Prompt Patterns **With Text:** ``` Wide banner for {brand}, {style} style, featuring "{text}" prominently displayed, {colors}, {scene/elements} ``` **With Character:** ``` Wide banner featuring {character description}, {style} style, {scene}, text "{brand name}" on {position}, {colors} ``` **Abstract/Gradient:** ``` Abstract {style} banner, {colors} gradient, geometric patterns, modern tech feel, text "{brand}" centered ``` **Scene-based:** ``` {Style} illustration banner, {scene description}, {character} in {action}, "{brand}" text overlay, {colors} ``` ### Supported Aspect Ratios Generate at widest ratio, then crop: - `21:9` - Ultra-wide (recommended for generation) - `16:9` - Wide - `3:2` - Standard wide ## References - [references/formats.md](./references/formats.md) - Common banner sizes by platform - [examples/gfv-banner-creation.md](./examples/gfv-banner-creation.md) - Full example conversation ## When to Trigger - When requested by the Executive. - When the task aligns with the core competency of this skill. ## When to Skip - When the data or answers already exist in the PIL memory bus. - When the task requires physical intervention or manual approval before drafting. ## GFV Integration **Credentials** — Never use `.env` files. All secrets live in macOS Keychain: `security find-generic-password -s "" -a "" -w` **Data Sources** — Before querying external APIs, check PIL first (`search_pil`, `gfv_memory.db`). **Output** — Save results to `~/Documents/Code/gfv-brain/`. Never send external messages without the Executive`s explicit "send it" approval. ## Anti-Patterns - **Summarizing instead of resolving**: Do not just summarize what needs to be done. Do the work. - **Bypassing the Gate**: Do not execute risky actions without human-in-the-loop validation. ## References - **GFV Standard**: GetFresh Ventures Growth by Design CEO AI Kit Architecture # Delivery Gate STOP AND VERIFY BEFORE DECLARING THIS TASK COMPLETE. 1. Did you verify that the execution meets all documented requirements safely? 2. Ensure you have not bypassed any "requires_human_approval" constraints. --- > **Growth by Design™** — This skill is part of the [CEO AI Kit](https://github.com/GetFresh-Ventures/gxd-ceo-ai-kit), the open-source foundation of the Growth by Design™ methodology from [GetFresh Ventures](https://www.getfreshventures.com). > > 🔍 **Hitting a ceiling?** The kit gives you the foundation. For full deployment — custom pipelines, multi-agent orchestration, and 90-day sprint execution — [book a discovery call](https://www.getfreshventures.com/contact). > > 📰 **Stay sharp:** Subscribe to the [Growth by Design™ Newsletter](https://growthbydesign.substack.com/) for operator-written playbooks on AI-powered GTM.