[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7FC3B0S99QPYZB9Q4CMAT91":3,"skill-same:01M7FC3B0S99QPYZB9Q4CMAT91":67},{"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":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":34,"updatedAt":35,"install":36,"resources":48},"01M7FC3B0S99QPYZB9Q4CMAT91","image-generation",null,"Plan and troubleshoot Sana image generation workflows across Diffusers, native pipelines, Sprint, ControlNet, high-resolution tiling, quantization, and Gradio launch modes.","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsana\u002Fsub-skills\u002Fimage-generation\u002FSKILL.md",[],[],"2026-11-16 10:05:34",false,"---\nname: image-generation\ndescription: \"Plan and troubleshoot Sana image generation workflows across\n  Diffusers, native pipelines, Sprint, ControlNet, high-resolution tiling,\n  quantization, and Gradio launch modes.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# Sana Image Generation\n\nUse this sub-skill when a user wants to generate or troubleshoot still images\nwith Sana, Sana-1.5, SANA-Sprint, Sana ControlNet HED, Diffusers Sana pipelines,\n4-bit\u002F8-bit quantized inference, 2K\u002F4K high-resolution generation, prompt-file\nbatch inference, or Gradio image demos.\n\nThis sub-skill is for planning and operating image inference workflows. It does\nnot own training\u002Fdata conversion, benchmark metrics, checkpoint conversion,\nSGLang\u002FComfyUI deployment, video generation, world models, or streaming.\n\n## Fast Route\n\n1. **Choose the execution surface.**\n   - For common text-to-image use in Python, prefer Diffusers\n     `SanaPipeline`, `SanaPAGPipeline`, or `SanaSprintPipeline`; read\n     [diffusers-and-quantization.md](references\u002Fdiffusers-and-quantization.md).\n   - For repository-native `.pth` checkpoints, config YAMLs, batch prompt\n     files, or ControlNet HED JSON inputs, read\n     [native-image-workflows.md](references\u002Fnative-image-workflows.md).\n   - For model IDs, native checkpoint\u002Fconfig pairing, resolution, precision,\n     and 2K\u002F4K selection, read\n     [model-config-selection.md](references\u002Fmodel-config-selection.md).\n   - For CUDA, Hugging Face, dtype, OOM, prompt file, ControlNet, xformers,\n     flash-attn, or Gradio failures, read\n     [troubleshooting.md](references\u002Ftroubleshooting.md).\n2. **Use a safe helper before running expensive generation.**\n   - `scripts\u002Fplan_sana_image_command.py` prints a native command or Diffusers\n     snippet and warnings without loading models or starting servers.\n   - `scripts\u002Fvalidate_controlnet_request.py` checks a ControlNet JSON and\n     reference images\u002Fcontrol maps without running the model or annotator.\n3. **Validate outputs after the real run.** Check that the output directory\n   exists, at least one `.jpg`\u002F`.png` file was written, images open with PIL,\n   and sizes match the requested or aspect-ratio-binned resolution.\n\n## Decision Checklist\n\n- **Need lowest friction?** Use Diffusers with a `*_diffusers` model id and a\n  matching `variant`\u002F`torch_dtype` pair.\n- **Need PAG guidance?** Use `SanaPAGPipeline` or native `pag_scale > 1` with a\n  linear-attention config; otherwise guidance falls back to classifier-free.\n- **Need one\u002Ffew-step speed?** Use SANA-Sprint with `SanaSprintPipeline`, the\n  Sprint `.pth` checkpoint\u002Fconfig pair, `sampling_algo=scm`, and usually two\n  steps.\n- **Need edge\u002Fsketch conditioning?** Use Sana ControlNet HED. Validate that each\n  JSON item has a prompt and exactly one of `ref_image_path` or\n  `ref_controlmap_path`; HED preprocessing requires a CUDA annotator checkpoint.\n- **Need 2K\u002F4K?** Use the 2K\u002F4K bf16 model IDs\u002Fconfigs, enable VAE tiling for\n  4K, and plan for high VRAM. Do not claim CPU fallback for generation.\n- **Need low VRAM?** Prefer 4-bit SVDQuant\u002FNunchaku for image inference when\n  the Nunchaku engine is installed; use 8-bit bitsandbytes for Diffusers\n  component quantization when quality and dependency trade-offs are acceptable.\n\n## Safe Command Planning Examples\n\nPrint a Diffusers PAG snippet:\n\n```bash\npython scripts\u002Fplan_sana_image_command.py \\\n  --mode diffusers \\\n  --diffusers-pipeline pag \\\n  --model-family sana15 \\\n  --prompt 'a cyberpunk cat with a neon sign that says \"Sana\"' \\\n  --seed 42\n```\n\nPrint a native batch command for a prompt text file:\n\n```bash\npython scripts\u002Fplan_sana_image_command.py \\\n  --mode native \\\n  --native-workflow sana \\\n  --resolution 1024 \\\n  --prompt-file prompts.txt \\\n  --work-dir output\u002Fsana-run \\\n  --sample-nums 10\n```\n\nCheck a ControlNet request before running generation:\n\n```bash\npython scripts\u002Fvalidate_controlnet_request.py \\\n  --json-file controlnet_request.json \\\n  --config configs\u002Fsana_controlnet_config\u002FSana_1600M_1024px_controlnet_bf16.yaml\n```\n\n## Boundaries\n\n- Route training, image\u002Ftext data layout, WebDataset\u002FWIDS conversion,\n  LoRA\u002FDreamBooth, FSDP\u002FDDP, or Sprint training to `training-data-configs`.\n- Route metrics, checkpoint conversion\u002Fexport, SVDQuant conversion, SGLang,\n  ComfyUI, or broad deployment to `evaluation-conversion-deployment`.\n- Route video, SANA-WM, LongSANA, or streaming V2V to `video-world-streaming`.\n\n## Evidence Labels\n\nThis sub-skill was distilled from the Sana README, installation\u002Fmodel-zoo\u002Fimage\ninference\u002FSprint\u002FControlNet\u002F4-bit\u002F8-bit docs; native app pipeline classes; image\ninference scripts; ControlNet utility code; and config families under\n`configs\u002Fsana_config`, `configs\u002Fsana1-5_config`, `configs\u002Fsana_sprint_config`,\n`configs\u002Fsana_controlnet_config`, and `configs\u002Fsana_app_config`. These are\nprovenance labels only; the operating guidance above is self-contained.\n","# Sana Image Generation\n\nUse this sub-skill when a user wants to generate or troubleshoot still images\nwith Sana, Sana-1.5, SANA-Sprint, Sana ControlNet HED, Diffusers Sana pipelines,\n4-bit\u002F8-bit quantized inference, 2K\u002F4K high-resolution generation, prompt-file\nbatch inference, or Gradio image demos.\n\nThis sub-skill is for planning and operating image inference workflows. It does\nnot own training\u002Fdata conversion, benchmark metrics, checkpoint conversion,\nSGLang\u002FComfyUI deployment, video generation, world models, or streaming.\n\n## Fast Route\n\n1. **Choose the execution surface.**\n   - For common text-to-image use in Python, prefer Diffusers\n     `SanaPipeline`, `SanaPAGPipeline`, or `SanaSprintPipeline`; read\n     [diffusers-and-quantization.md](references\u002Fdiffusers-and-quantization.md).\n   - For repository-native `.pth` checkpoints, config YAMLs, batch prompt\n     files, or ControlNet HED JSON inputs, read\n     [native-image-workflows.md](references\u002Fnative-image-workflows.md).\n   - For model IDs, native checkpoint\u002Fconfig pairing, resolution, precision,\n     and 2K\u002F4K selection, read\n     [model-config-selection.md](references\u002Fmodel-config-selection.md).\n   - For CUDA, Hugging Face, dtype, OOM, prompt file, ControlNet, xformers,\n     flash-attn, or Gradio failures, read\n     [troubleshooting.md](references\u002Ftroubleshooting.md).\n2. **Use a safe helper before running expensive generation.**\n   - `scripts\u002Fplan_sana_image_command.py` prints a native command or Diffusers\n     snippet and warnings without loading models or starting servers.\n   - `scripts\u002Fvalidate_controlnet_request.py` checks a ControlNet JSON and\n     reference images\u002Fcontrol maps without running the model or annotator.\n3. **Validate outputs after the real run.** Check that the output directory\n   exists, at least one `.jpg`\u002F`.png` file was written, images open with PIL,\n   and sizes match the requested or aspect-ratio-binned resolution.\n\n## Decision Checklist\n\n- **Need lowest friction?** Use Diffusers with a `*_diffusers` model id and a\n  matching `variant`\u002F`torch_dtype` pair.\n- **Need PAG guidance?** Use `SanaPAGPipeline` or native `pag_scale > 1` with a\n  linear-attention config; otherwise guidance falls back to classifier-free.\n- **Need one\u002Ffew-step speed?** Use SANA-Sprint with `SanaSprintPipeline`, the\n  Sprint `.pth` checkpoint\u002Fconfig pair, `sampling_algo=scm`, and usually two\n  steps.\n- **Need edge\u002Fsketch conditioning?** Use Sana ControlNet HED. Validate that each\n  JSON item has a prompt and exactly one of `ref_image_path` or\n  `ref_controlmap_path`; HED preprocessing requires a CUDA annotator checkpoint.\n- **Need 2K\u002F4K?** Use the 2K\u002F4K bf16 model IDs\u002Fconfigs, enable VAE tiling for\n  4K, and plan for high VRAM. Do not claim CPU fallback for generation.\n- **Need low VRAM?** Prefer 4-bit SVDQuant\u002FNunchaku for image inference when\n  the Nunchaku engine is installed; use 8-bit bitsandbytes for Diffusers\n  component quantization when quality and dependency trade-offs are acceptable.\n\n## Safe Command Planning Examples\n\nPrint a Diffusers PAG snippet:\n\n```bash\npython scripts\u002Fplan_sana_image_command.py \\\n  --mode diffusers \\\n  --diffusers-pipeline pag \\\n  --model-family sana15 \\\n  --prompt 'a cyberpunk cat with a neon sign that says \"Sana\"' \\\n  --seed 42\n```\n\nPrint a native batch command for a prompt text file:\n\n```bash\npython scripts\u002Fplan_sana_image_command.py \\\n  --mode native \\\n  --native-workflow sana \\\n  --resolution 1024 \\\n  --prompt-file prompts.txt \\\n  --work-dir output\u002Fsana-run \\\n  --sample-nums 10\n```\n\nCheck a ControlNet request before running generation:\n\n```bash\npython scripts\u002Fvalidate_controlnet_request.py \\\n  --json-file controlnet_request.json \\\n  --config configs\u002Fsana_controlnet_config\u002FSana_1600M_1024px_controlnet_bf16.yaml\n```\n\n## Boundaries\n\n- Route training, image\u002Ftext data layout, WebDataset\u002FWIDS conversion,\n  LoRA\u002FDreamBooth, FSDP\u002FDDP, or Sprint training to `training-data-configs`.\n- Route metrics, checkpoint conversion\u002Fexport, SVDQuant conversion, SGLang,\n  ComfyUI, or broad deployment to `evaluation-conversion-deployment`.\n- Route video, SANA-WM, LongSANA, or streaming V2V to `video-world-streaming`.\n\n## Evidence Labels\n\nThis sub-skill was distilled from the Sana README, installation\u002Fmodel-zoo\u002Fimage\ninference\u002FSprint\u002FControlNet\u002F4-bit\u002F8-bit docs; native app pipeline classes; image\ninference scripts; ControlNet utility code; and config families under\n`configs\u002Fsana_config`, `configs\u002Fsana1-5_config`, `configs\u002Fsana_sprint_config`,\n`configs\u002Fsana_controlnet_config`, and `configs\u002Fsana_app_config`. These are\nprovenance labels only; the operating guidance above is self-contained.\n",{"name":5,"license":24,"metadata":25,"description":7,"disable-model-invocation":27},"Apache 2.0",{"disco-role":26},"operating",true,"main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","06-前端与设计\u002Fsana\u002Fsub-skills\u002Fimage-generation",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":27},[38,41,44],{"tool":39,"command":40},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill image-generation",{"tool":42,"command":43},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill image-generation",{"tool":45,"command":46},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill image-generation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsana\u002Fsub-skills\u002Fimage-generation\u002FSKILL.md",[49,55,59,63],{"path":50,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":53,"sourceHtmlUrl":54},"references\u002Fdiffusers-and-quantization.md","REFERENCE","参考文档",8010,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsana\u002Fsub-skills\u002Fimage-generation\u002Freferences\u002Fdiffusers-and-quantization.md",{"path":56,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":57,"sourceHtmlUrl":58},"references\u002Fmodel-config-selection.md",9961,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsana\u002Fsub-skills\u002Fimage-generation\u002Freferences\u002Fmodel-config-selection.md",{"path":60,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":61,"sourceHtmlUrl":62},"references\u002Fnative-image-workflows.md",11428,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsana\u002Fsub-skills\u002Fimage-generation\u002Freferences\u002Fnative-image-workflows.md",{"path":64,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":65,"sourceHtmlUrl":66},"references\u002Ftroubleshooting.md",9306,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsana\u002Fsub-skills\u002Fimage-generation\u002Freferences\u002Ftroubleshooting.md",[68,76,81,88,95,102],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],{"skillNo":77,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":19},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]