[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77R3G0T65J9GQ778CDPBSAS":3,"skill-same:01M77R3G0T65J9GQ778CDPBSAS":54},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":31,"commitSha":32,"skillDirectory":33,"forksCount":34,"repositoryHtmlUrl":35,"repositoryHomepageUrl":36,"ownerHtmlUrl":37,"ownerBlogUrl":6,"ownerDescription":38,"updatedAt":39,"install":40,"resources":53},"01M77R3G0T65J9GQ778CDPBSAS","ai-image",null,"Full AI image creation workflow — intent classification, prompt enhancement, multi-direction generation via fal.ai, and error recovery. Triggers on \"generate image\", \"ai image\", \"create image\", \"make an image\", \"image generation\", \"\u002Fai-image\", or any image creation request.","Lord1Egypt\u002Fawesome-skill-forge","Lord1Egypt","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F99746359?v=4","01M7592YDNT0XK9WJ0T2MQVPDN","Mohamed Mounir Ali AL Keshita","01M75GRFNHC8VZNWEB8DKDE433",3,"MIT","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fp\u002Fphy-ai-image\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ai-image\ndescription: Full AI image creation workflow — intent classification, prompt enhancement, multi-direction generation via fal.ai, and error recovery. Triggers on \"generate image\", \"ai image\", \"create image\", \"make an image\", \"image generation\", \"\u002Fai-image\", or any image creation request.\nhomepage: https:\u002F\u002Fcanlah.ai\nmetadata: {\"openclaw\": {\"emoji\": \"🎨\", \"os\": [\"darwin\", \"linux\"]}}\n---\n\n# AI Image Creation Workflow\n\nComplete image creation pipeline: classify intent, enhance prompt, generate via fal.ai, recover from errors.\n\nFor brand-specific visuals, use a dedicated brand image skill if available.\nFor model selection help, reference `\u002Ffal-model-selector`.\n\n---\n\n## Phase 1: Intent Classification\n\nWhen a user mentions image creation, FIRST classify their intent into one of six categories:\n\n### A. EXPLORING — No clear idea yet\n**Signals:** \"help me think of something\", \"any inspiration\", \"not sure what to make\"\n\n**Action:**\n1. Ask about use case (social media? product? marketing? personal?)\n2. Search `references\u002Fprompt-library-curated.json` for relevant examples\n3. Show 3-5 options with descriptions\n4. Let user pick, THEN proceed to generation\n\n### B. BRIEF IDEA — Intent but underspecified\n**Signals:** Description under ~30 words, lacks visual details (composition, lighting, color, texture, perspective)\n**Examples:** \"portrait photo\", \"tech logo\", \"gazebo in garden\"\n\n**Action:**\n1. Enhance the prompt using the appropriate style template from `references\u002Fprompt-templates.md`\n2. Show the enhanced prompt with brief creative rationale\n3. Wait for user confirmation before generating\n\n### C. DETAILED PROMPT — User knows what they want\n**Signals:** Multi-sentence prompt with specific visual details, style references, or technical terms\n**Examples:** Prompt with lens specs, lighting setup, material descriptions\n\n**Action:**\n1. Generate directly — do NOT over-process\n2. Only suggest minor tweaks if obvious improvements spotted\n\n### D. EDIT\u002FMODIFY — Change an existing image\n**Signals:** User provides\u002Freferences an image AND describes a specific change\n**Examples:** \"add text to this\", \"change the background\", \"remove the person\"\n\n**Action:**\n1. Keep prompt minimal and edit-focused\n2. Upload reference image if local (use fal.ai upload or public URL)\n3. Prompt describes ONLY the edit, not the entire image\n4. Use edit-capable models: `fal-ai\u002Fflux-2\u002Fedit`, `fal-ai\u002Fgpt-image-1.5\u002Fedit`, `fal-ai\u002Fbytedance\u002Fseedream\u002Fv4.5\u002Fedit`\n\n### E. BATCH REQUEST — Multiple variations\n**Signals:** \"4 directions\", \"multiple versions\", \"a set of assets\"\n\n**Action:**\n1. Plan the variants first, present as a table\n2. ASK user which direction(s) to try: \"Pick a number, or I can generate all N\"\n3. NEVER auto-generate all variants without explicit choice\n4. Generate AFTER user responds\n5. Max 4 parallel per batch\n\n### F. CREATIVE + EXTENSIONS — Multi-step workflow\n**Signals:** \"design a logo and make mockups\", \"create X and apply to Y\"\n\n**Action:**\n1. Plan 3-5 design directions, present to user, ASK which to try\n2. Generate ONLY the chosen direction(s)\n3. Show result, get approval\n4. THEN plan and generate extensions\u002Fderivatives\n5. NEVER jump from plan to generating everything\n\n---\n\n## Phase 2: Prompt Enhancement\n\nWhen the user provides a brief idea (Intent B), enhance using the appropriate style template.\n\n### Style Selection\n\n| Style | When to Use | Template |\n|-------|-------------|----------|\n| **Realistic** | Product photography, architecture, portraits, any photorealistic need | See `references\u002Fprompt-templates.md` Section 1 |\n| **Anime** | 2D illustration, manga style, character art | See `references\u002Fprompt-templates.md` Section 2 |\n| **Illustration** | Concept art, digital painting, editorial illustration | See `references\u002Fprompt-templates.md` Section 3 |\n\n### Enhancement Protocol\n\n1. Read the appropriate template from `references\u002Fprompt-templates.md`\n2. Apply the template's methodology to transform the user's brief input\n3. Output the enhanced prompt in two parts:\n   - **Narrative Specification**: Dense descriptive paragraph\n   - **Technical Metadata**: Visual style, key elements, lighting, composition\n4. Show to user, explain creative choices briefly\n5. Proceed only after confirmation\n\n### Key Principles (from \"Blueprint Method\")\n- **Technical precision over feeling**: Not \"cinematic\" but \"three-point lighting with warm key, cool fill, 85mm f\u002F1.4\"\n- **Quantifiable spatial logic**: Foreground\u002Fmidground\u002Fbackground, lens specs, lighting setup\n- **Material physics**: How materials interact with light — \"brushed aluminum catching overhead softbox\", not just \"shiny metal\"\n- **Cohesive narrative**: Prompt reads like a director's brief, not a tag list\n\n---\n\n## Phase 3: Generation via fal.ai\n\n### Default Models by Task\n\n| Task | Model | Price | Notes |\n|------|-------|-------|-------|\n| **General image** | `fal-ai\u002Fflux-2-pro` | $0.03\u002FMP | Zero-config, production-ready |\n| **Product photo** | `fal-ai\u002Fimage-apps-v2\u002Fproduct-photography` | - | Studio lighting presets |\n| **Text in image** | `fal-ai\u002Fideogram\u002Fv3` | $0.03-0.09 | Best typography |\n| **Photo-real portrait** | `fal-ai\u002Fnano-banana-pro` | ~$0.10 | High quality portraits |\n| **Scene editing** | `fal-ai\u002Fbytedance\u002Fseedream\u002Fv4.5\u002Fedit` | - | Best for scene changes |\n| **General editing** | `fal-ai\u002Fflux-2\u002Fedit` | $0.012\u002FMP | Multi-image support |\n| **Google quality** | `fal-ai\u002Fgemini-3-pro-image-preview` | $0.15-0.30 | Up to 4K |\n\n### Generation Code Pattern\n\n```python\nimport fal_client\nimport os\n\n# Set your FAL API key via environment variable\nfal_client.api_key = os.environ.get(\"FAL_API_KEY\")\n\n# Text to image\nresult = fal_client.subscribe(\n    \"fal-ai\u002Fflux-2-pro\",\n    arguments={\n        \"prompt\": \"YOUR ENHANCED PROMPT HERE\",\n        \"image_size\": \"landscape_16_9\",  # or square, portrait_4_3, etc.\n        \"num_images\": 1,\n    }\n)\nimage_url = result[\"images\"][0][\"url\"]\nprint(f\"Generated: {image_url}\")\n\n# Download locally\nimport urllib.request\nurllib.request.urlretrieve(image_url, os.path.expanduser(\"~\u002FPictures\u002Fai-image\u002Foutput.jpg\"))\n```\n\n### Multi-Direction Parallel Generation\n\nFor batch requests (Intent E), generate up to 4 in parallel:\n\n```python\nimport fal_client\nimport concurrent.futures\nimport os\n\nfal_client.api_key = os.environ.get(\"FAL_API_KEY\")\n\nprompts = [\n    \"Direction 1: ...\",\n    \"Direction 2: ...\",\n    \"Direction 3: ...\",\n]\n\ndef generate(prompt):\n    return fal_client.subscribe(\n        \"fal-ai\u002Fflux-2-pro\",\n        arguments={\"prompt\": prompt, \"image_size\": \"landscape_16_9\"}\n    )\n\nwith concurrent.futures.ThreadPoolExecutor(max_workers=4) as pool:\n    results = list(pool.map(generate, prompts))\n\nfor i, r in enumerate(results):\n    print(f\"Direction {i+1}: {r['images'][0]['url']}\")\n```\n\n### Image Size Options (fal.ai)\n\n| Size Name | Dimensions | Use Case |\n|-----------|------------|----------|\n| `square` | 1024x1024 | Social media, profile |\n| `square_hd` | 1536x1536 | High-res square |\n| `portrait_4_3` | 768x1024 | Portrait |\n| `portrait_16_9` | 576x1024 | Stories, vertical |\n| `landscape_4_3` | 1024x768 | Standard landscape |\n| `landscape_16_9` | 1024x576 | Banner, cinematic |\n\n### Reference Image Usage\n\n```python\n# Edit with reference\nresult = fal_client.subscribe(\n    \"fal-ai\u002Fflux-2\u002Fedit\",\n    arguments={\n        \"prompt\": \"Change the background to a sunset beach\",\n        \"image_url\": \"https:\u002F\u002Fexample.com\u002Foriginal.jpg\",\n    }\n)\n\n# Product photography with reference\nresult = fal_client.subscribe(\n    \"fal-ai\u002Fimage-apps-v2\u002Fproduct-photography\",\n    arguments={\n        \"product_image_url\": \"https:\u002F\u002Fexample.com\u002Fproduct.jpg\",\n        \"aspect_ratio\": \"1:1\",\n    }\n)\n```\n\n---\n\n## Phase 4: Presenting Results\n\n### Before generating:\n- When enhancing prompts, briefly explain creative direction\n- When planning variants, describe each direction distinctively\n\n### After generating:\n- Show the image URL and local file path\n- Do NOT describe or imagine what the image looks like (you cannot see it)\n- Keep it brief\n- Suggest next steps: \"Want to try a different direction?\" or \"Ready to create extensions?\"\n\n### Format:\n```\n**Direction 1: Modern Minimal**\nImage: [url]\nSaved to: ~\u002FPictures\u002Fai-image\u002F...\n\n**Direction 2: Dramatic Lighting**\nImage: [url]\nSaved to: ~\u002FPictures\u002Fai-image\u002F...\n```\n\n---\n\n## Phase 5: Error Recovery\n\nWhen generation fails, diagnose and guide:\n\n| Error Type | Detection | Recovery |\n|-----------|-----------|----------|\n| **Safety\u002Fcontent** | \"safety\", \"policy\", \"flagged\" | Auto-rephrase prompt, offer cleaned version |\n| **Rate limit** | 429, \"rate limit\" | Wait 30s and retry once |\n| **Model unavailable** | \"not found\", \"inactive\" | Suggest alternative from `\u002Ffal-model-selector` |\n| **Invalid size\u002Fratio** | \"not supported\" | Show valid options |\n| **Network error** | \"fetch failed\", \"timeout\" | Retry once automatically |\n| **Credit\u002Fbilling** | 402, \"insufficient\" | Warn user, suggest cheaper model |\n\n### Auto-retry logic:\n- Network errors: retry once after 5s\n- Rate limits: retry once after 30s\n- All others: show error + guidance, don't retry\n\n---\n\n## Prompt Library\n\nA curated collection of high-quality prompt templates is available at `references\u002Fprompt-library-curated.json`.\n\nSearch it when user is exploring (Intent A) or needs inspiration.\n\n### How to search:\n```python\nimport json\nimport os\n\nwith open(os.path.expanduser(\"~\u002F.claude\u002Fskills\u002Fai-image\u002Freferences\u002Fprompt-library-curated.json\")) as f:\n    library = json.load(f)\n\n# Search by keyword\nkeyword = \"product\"\nmatches = [p for p in library if keyword.lower() in p[\"prompt\"].lower() or keyword.lower() in \" \".join(p[\"categories\"]).lower()]\n```\n\n### Categories available:\n- Product (product photography, packaging, advertising)\n- 3D (icons, infographics, typography)\n- Photograph (editorial, fashion, landscape)\n- Food (food photography, restaurant)\n- App (UI\u002FUX mockups)\n\n---\n\n## Decision Tree Summary\n\n```\nUser mentions image creation\n    |\n    v\nClassify intent (A-F)\n    |\n    +-- A (Exploring) --> Search prompt library --> Show options --> User picks --> Enhance --> Generate\n    +-- B (Brief idea) --> Enhance prompt --> Show to user --> Confirm --> Generate\n    +-- C (Detailed) --> Generate directly\n    +-- D (Edit) --> Upload ref image --> Minimal edit prompt --> Generate with edit model\n    +-- E (Batch) --> Plan variants --> User picks --> Generate in parallel (max 4)\n    +-- F (Multi-step) --> Plan directions --> User picks --> Generate --> Approve --> Extensions\n    |\n    v\nPresent results (URL + local path, no description)\n    |\n    v\nSuggest next steps\n```\n\n---\n\n## Integration with Other Skills\n\n| Need | Skill |\n|------|-------|\n| Brand-specific visuals | Use your brand image skill |\n| Model selection help | `\u002Ffal-model-selector` |\n| Video generation | `\u002Ffal-model-selector` (video section) |\n| Full model database | `\u002Ffal-ai-models` |\n\n---\n\n## Author\n\n**[Canlah AI](https:\u002F\u002Fcanlah.ai)** — Run performance marketing without breaking your brand.\n\n- GitHub: [github.com\u002FPHY041](https:\u002F\u002Fgithub.com\u002FPHY041)\n- All Skills: [clawhub.ai\u002FPHY041](https:\u002F\u002Fclawhub.ai\u002FPHY041)\n","# AI Image Creation Workflow\n\nComplete image creation pipeline: classify intent, enhance prompt, generate via fal.ai, recover from errors.\n\nFor brand-specific visuals, use a dedicated brand image skill if available.\nFor model selection help, reference `\u002Ffal-model-selector`.\n\n---\n\n## Phase 1: Intent Classification\n\nWhen a user mentions image creation, FIRST classify their intent into one of six categories:\n\n### A. EXPLORING — No clear idea yet\n**Signals:** \"help me think of something\", \"any inspiration\", \"not sure what to make\"\n\n**Action:**\n1. Ask about use case (social media? product? marketing? personal?)\n2. Search `references\u002Fprompt-library-curated.json` for relevant examples\n3. Show 3-5 options with descriptions\n4. Let user pick, THEN proceed to generation\n\n### B. BRIEF IDEA — Intent but underspecified\n**Signals:** Description under ~30 words, lacks visual details (composition, lighting, color, texture, perspective)\n**Examples:** \"portrait photo\", \"tech logo\", \"gazebo in garden\"\n\n**Action:**\n1. Enhance the prompt using the appropriate style template from `references\u002Fprompt-templates.md`\n2. Show the enhanced prompt with brief creative rationale\n3. Wait for user confirmation before generating\n\n### C. DETAILED PROMPT — User knows what they want\n**Signals:** Multi-sentence prompt with specific visual details, style references, or technical terms\n**Examples:** Prompt with lens specs, lighting setup, material descriptions\n\n**Action:**\n1. Generate directly — do NOT over-process\n2. Only suggest minor tweaks if obvious improvements spotted\n\n### D. EDIT\u002FMODIFY — Change an existing image\n**Signals:** User provides\u002Freferences an image AND describes a specific change\n**Examples:** \"add text to this\", \"change the background\", \"remove the person\"\n\n**Action:**\n1. Keep prompt minimal and edit-focused\n2. Upload reference image if local (use fal.ai upload or public URL)\n3. Prompt describes ONLY the edit, not the entire image\n4. Use edit-capable models: `fal-ai\u002Fflux-2\u002Fedit`, `fal-ai\u002Fgpt-image-1.5\u002Fedit`, `fal-ai\u002Fbytedance\u002Fseedream\u002Fv4.5\u002Fedit`\n\n### E. BATCH REQUEST — Multiple variations\n**Signals:** \"4 directions\", \"multiple versions\", \"a set of assets\"\n\n**Action:**\n1. Plan the variants first, present as a table\n2. ASK user which direction(s) to try: \"Pick a number, or I can generate all N\"\n3. NEVER auto-generate all variants without explicit choice\n4. Generate AFTER user responds\n5. Max 4 parallel per batch\n\n### F. CREATIVE + EXTENSIONS — Multi-step workflow\n**Signals:** \"design a logo and make mockups\", \"create X and apply to Y\"\n\n**Action:**\n1. Plan 3-5 design directions, present to user, ASK which to try\n2. Generate ONLY the chosen direction(s)\n3. Show result, get approval\n4. THEN plan and generate extensions\u002Fderivatives\n5. NEVER jump from plan to generating everything\n\n---\n\n## Phase 2: Prompt Enhancement\n\nWhen the user provides a brief idea (Intent B), enhance using the appropriate style template.\n\n### Style Selection\n\n| Style | When to Use | Template |\n|-------|-------------|----------|\n| **Realistic** | Product photography, architecture, portraits, any photorealistic need | See `references\u002Fprompt-templates.md` Section 1 |\n| **Anime** | 2D illustration, manga style, character art | See `references\u002Fprompt-templates.md` Section 2 |\n| **Illustration** | Concept art, digital painting, editorial illustration | See `references\u002Fprompt-templates.md` Section 3 |\n\n### Enhancement Protocol\n\n1. Read the appropriate template from `references\u002Fprompt-templates.md`\n2. Apply the template's methodology to transform the user's brief input\n3. Output the enhanced prompt in two parts:\n   - **Narrative Specification**: Dense descriptive paragraph\n   - **Technical Metadata**: Visual style, key elements, lighting, composition\n4. Show to user, explain creative choices briefly\n5. Proceed only after confirmation\n\n### Key Principles (from \"Blueprint Method\")\n- **Technical precision over feeling**: Not \"cinematic\" but \"three-point lighting with warm key, cool fill, 85mm f\u002F1.4\"\n- **Quantifiable spatial logic**: Foreground\u002Fmidground\u002Fbackground, lens specs, lighting setup\n- **Material physics**: How materials interact with light — \"brushed aluminum catching overhead softbox\", not just \"shiny metal\"\n- **Cohesive narrative**: Prompt reads like a director's brief, not a tag list\n\n---\n\n## Phase 3: Generation via fal.ai\n\n### Default Models by Task\n\n| Task | Model | Price | Notes |\n|------|-------|-------|-------|\n| **General image** | `fal-ai\u002Fflux-2-pro` | $0.03\u002FMP | Zero-config, production-ready |\n| **Product photo** | `fal-ai\u002Fimage-apps-v2\u002Fproduct-photography` | - | Studio lighting presets |\n| **Text in image** | `fal-ai\u002Fideogram\u002Fv3` | $0.03-0.09 | Best typography |\n| **Photo-real portrait** | `fal-ai\u002Fnano-banana-pro` | ~$0.10 | High quality portraits |\n| **Scene editing** | `fal-ai\u002Fbytedance\u002Fseedream\u002Fv4.5\u002Fedit` | - | Best for scene changes |\n| **General editing** | `fal-ai\u002Fflux-2\u002Fedit` | $0.012\u002FMP | Multi-image support |\n| **Google quality** | `fal-ai\u002Fgemini-3-pro-image-preview` | $0.15-0.30 | Up to 4K |\n\n### Generation Code Pattern\n\n```python\nimport fal_client\nimport os\n\n# Set your FAL API key via environment variable\nfal_client.api_key = os.environ.get(\"FAL_API_KEY\")\n\n# Text to image\nresult = fal_client.subscribe(\n    \"fal-ai\u002Fflux-2-pro\",\n    arguments={\n        \"prompt\": \"YOUR ENHANCED PROMPT HERE\",\n        \"image_size\": \"landscape_16_9\",  # or square, portrait_4_3, etc.\n        \"num_images\": 1,\n    }\n)\nimage_url = result[\"images\"][0][\"url\"]\nprint(f\"Generated: {image_url}\")\n\n# Download locally\nimport urllib.request\nurllib.request.urlretrieve(image_url, os.path.expanduser(\"~\u002FPictures\u002Fai-image\u002Foutput.jpg\"))\n```\n\n### Multi-Direction Parallel Generation\n\nFor batch requests (Intent E), generate up to 4 in parallel:\n\n```python\nimport fal_client\nimport concurrent.futures\nimport os\n\nfal_client.api_key = os.environ.get(\"FAL_API_KEY\")\n\nprompts = [\n    \"Direction 1: ...\",\n    \"Direction 2: ...\",\n    \"Direction 3: ...\",\n]\n\ndef generate(prompt):\n    return fal_client.subscribe(\n        \"fal-ai\u002Fflux-2-pro\",\n        arguments={\"prompt\": prompt, \"image_size\": \"landscape_16_9\"}\n    )\n\nwith concurrent.futures.ThreadPoolExecutor(max_workers=4) as pool:\n    results = list(pool.map(generate, prompts))\n\nfor i, r in enumerate(results):\n    print(f\"Direction {i+1}: {r['images'][0]['url']}\")\n```\n\n### Image Size Options (fal.ai)\n\n| Size Name | Dimensions | Use Case |\n|-----------|------------|----------|\n| `square` | 1024x1024 | Social media, profile |\n| `square_hd` | 1536x1536 | High-res square |\n| `portrait_4_3` | 768x1024 | Portrait |\n| `portrait_16_9` | 576x1024 | Stories, vertical |\n| `landscape_4_3` | 1024x768 | Standard landscape |\n| `landscape_16_9` | 1024x576 | Banner, cinematic |\n\n### Reference Image Usage\n\n```python\n# Edit with reference\nresult = fal_client.subscribe(\n    \"fal-ai\u002Fflux-2\u002Fedit\",\n    arguments={\n        \"prompt\": \"Change the background to a sunset beach\",\n        \"image_url\": \"https:\u002F\u002Fexample.com\u002Foriginal.jpg\",\n    }\n)\n\n# Product photography with reference\nresult = fal_client.subscribe(\n    \"fal-ai\u002Fimage-apps-v2\u002Fproduct-photography\",\n    arguments={\n        \"product_image_url\": \"https:\u002F\u002Fexample.com\u002Fproduct.jpg\",\n        \"aspect_ratio\": \"1:1\",\n    }\n)\n```\n\n---\n\n## Phase 4: Presenting Results\n\n### Before generating:\n- When enhancing prompts, briefly explain creative direction\n- When planning variants, describe each direction distinctively\n\n### After generating:\n- Show the image URL and local file path\n- Do NOT describe or imagine what the image looks like (you cannot see it)\n- Keep it brief\n- Suggest next steps: \"Want to try a different direction?\" or \"Ready to create extensions?\"\n\n### Format:\n```\n**Direction 1: Modern Minimal**\nImage: [url]\nSaved to: ~\u002FPictures\u002Fai-image\u002F...\n\n**Direction 2: Dramatic Lighting**\nImage: [url]\nSaved to: ~\u002FPictures\u002Fai-image\u002F...\n```\n\n---\n\n## Phase 5: Error Recovery\n\nWhen generation fails, diagnose and guide:\n\n| Error Type | Detection | Recovery |\n|-----------|-----------|----------|\n| **Safety\u002Fcontent** | \"safety\", \"policy\", \"flagged\" | Auto-rephrase prompt, offer cleaned version |\n| **Rate limit** | 429, \"rate limit\" | Wait 30s and retry once |\n| **Model unavailable** | \"not found\", \"inactive\" | Suggest alternative from `\u002Ffal-model-selector` |\n| **Invalid size\u002Fratio** | \"not supported\" | Show valid options |\n| **Network error** | \"fetch failed\", \"timeout\" | Retry once automatically |\n| **Credit\u002Fbilling** | 402, \"insufficient\" | Warn user, suggest cheaper model |\n\n### Auto-retry logic:\n- Network errors: retry once after 5s\n- Rate limits: retry once after 30s\n- All others: show error + guidance, don't retry\n\n---\n\n## Prompt Library\n\nA curated collection of high-quality prompt templates is available at `references\u002Fprompt-library-curated.json`.\n\nSearch it when user is exploring (Intent A) or needs inspiration.\n\n### How to search:\n```python\nimport json\nimport os\n\nwith open(os.path.expanduser(\"~\u002F.claude\u002Fskills\u002Fai-image\u002Freferences\u002Fprompt-library-curated.json\")) as f:\n    library = json.load(f)\n\n# Search by keyword\nkeyword = \"product\"\nmatches = [p for p in library if keyword.lower() in p[\"prompt\"].lower() or keyword.lower() in \" \".join(p[\"categories\"]).lower()]\n```\n\n### Categories available:\n- Product (product photography, packaging, advertising)\n- 3D (icons, infographics, typography)\n- Photograph (editorial, fashion, landscape)\n- Food (food photography, restaurant)\n- App (UI\u002FUX mockups)\n\n---\n\n## Decision Tree Summary\n\n```\nUser mentions image creation\n    |\n    v\nClassify intent (A-F)\n    |\n    +-- A (Exploring) --> Search prompt library --> Show options --> User picks --> Enhance --> Generate\n    +-- B (Brief idea) --> Enhance prompt --> Show to user --> Confirm --> Generate\n    +-- C (Detailed) --> Generate directly\n    +-- D (Edit) --> Upload ref image --> Minimal edit prompt --> Generate with edit model\n    +-- E (Batch) --> Plan variants --> User picks --> Generate in parallel (max 4)\n    +-- F (Multi-step) --> Plan directions --> User picks --> Generate --> Approve --> Extensions\n    |\n    v\nPresent results (URL + local path, no description)\n    |\n    v\nSuggest next steps\n```\n\n---\n\n## Integration with Other Skills\n\n| Need | Skill |\n|------|-------|\n| Brand-specific visuals | Use your brand image skill |\n| Model selection help | `\u002Ffal-model-selector` |\n| Video generation | `\u002Ffal-model-selector` (video section) |\n| Full model database | `\u002Ffal-ai-models` |\n\n---\n\n## Author\n\n**[Canlah AI](https:\u002F\u002Fcanlah.ai)** — Run performance marketing without breaking your brand.\n\n- GitHub: [github.com\u002FPHY041](https:\u002F\u002Fgithub.com\u002FPHY041)\n- All Skills: [clawhub.ai\u002FPHY041](https:\u002F\u002Fclawhub.ai\u002FPHY041)\n",{"name":5,"homepage":24,"metadata":25,"description":7},"https:\u002F\u002Fcanlah.ai",{"openclaw":26},{"os":27,"emoji":30},[28,29],"darwin","linux","🎨","master","4aa004da461d820f4ea5499e58eb8d196cf99016","community\u002Fclawhub\u002Fp\u002Fphy-ai-image",0,"https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge","https:\u002F\u002Fskill-explorer-rho.vercel.app","https:\u002F\u002Fgithub.com\u002FLord1Egypt","Unbreakable. Unstoppable. Unmatched⚡.\r\n Speaker of machine tongues 🤖  | World-builder of decentralized dreams 🌐 | Defender of the digital realm🛡️.","2026-06-10 08:21:01",{"source":8,"skillName":5,"commands":41,"prompt":51,"zipAvailable":52},[42,45,48],{"tool":43,"command":44},"npx","npx skills add Lord1Egypt\u002Fawesome-skill-forge --skill ai-image",{"tool":46,"command":47},"bunx","bunx skills add Lord1Egypt\u002Fawesome-skill-forge --skill ai-image",{"tool":49,"command":50},"pnpm","pnpm dlx skills add Lord1Egypt\u002Fawesome-skill-forge --skill ai-image","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fp\u002Fphy-ai-image\u002FSKILL.md",true,[],[55,62,69,76,83,90],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M785WKJB99G3STR1NF9NRTJW","youtube-content","YouTube transcripts to summaries, threads, blogs.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fyoutube-content\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M785WEFHB0G2S9YA6EPR3H0E","xurl","X\u002FTwitter via xurl CLI: post, search, DM, media, v2 API.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fxurl\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M785W98EDN4EFJXJPMQEHCRC","touchdesigner-mcp","Control a running TouchDesigner instance via twozero MCP — create operators, set parameters, wire connections, execute Python, build real-time visuals. 36 native tools.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftouchdesigner-mcp\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M785W5S9FP3QBH2V191E025P","test-driven-development","TDD: enforce RED-GREEN-REFACTOR, tests before code.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M785W2AS08NSSH9YATCV6FJB","teams-meeting-pipeline","Operate the Teams meeting summary pipeline via Hermes CLI — summarize meetings, inspect pipeline status, replay jobs, manage Microsoft Graph subscriptions.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fteams-meeting-pipeline\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":19},"01M785VXKW4T34BTP1SCHW1R7F","systematic-debugging","4-phase root cause debugging: understand bugs before fixing.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fsystematic-debugging\u002FSKILL.md",[],[]]