[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78SMM5HZ07CWAP728NETMAQ":3,"skill-same:01M78SMM5HZ07CWAP728NETMAQ":78},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":32,"install":33,"resources":45},"01M78SMM5HZ07CWAP728NETMAQ","inference-and-demo",null,"Runs AnyDoor image generation, demo, and prediction workflows.","VectorSpaceLab\u002FAREX-Skill","VectorSpaceLab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F180637102?v=4","01M7538C5WPSJGWF7KHR4EFZ9F","01M7538C6R2GFP53NPXTCZNBRT",351,"MIT","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: inference-and-demo\ndescription: \"Runs AnyDoor image generation, demo, and prediction workflows.\"\nmetadata:\n  disco-role: operating\ndisable-model-invocation: true\nlicense: MIT\n---\n\n# Inference and Demo\n\nUse this sub-skill when the task is about generating a customized image,\nlaunching the local Gradio demo, or understanding the prediction interface.\n\n## What this sub-skill owns\n\n- Single-image AnyDoor inference.\n- VITON-HD-style batch inference.\n- Mask and image validation before generation.\n- Local Gradio launch behavior and interactive options.\n- Cog \u002F Replicate-style prediction input and output contracts.\n\n## What this sub-skill does not own\n\n- Environment installation and checkpoint patching.\n- Dataset format design or training data preparation.\n- Weight conversion and training launch details.\n\nRoute those to the sibling sub-skills.\n\n## Read first\n\n- `..\u002F..\u002Freferences\u002Foverview.md`\n- `..\u002F..\u002Freferences\u002Fcheckpoints-and-configs.md`\n- `..\u002F..\u002Freferences\u002Ftroubleshooting.md`\n- `references\u002Finference-workflows.md`\n- `references\u002Fmask-and-image-formats.md`\n- `references\u002Fdemo-and-cog-deployment.md`\n- `references\u002Ftroubleshooting.md`\n\n## Use these scripts\n\n- `scripts\u002Fvalidate_inference_inputs.py` to check image and mask pairs.\n- `scripts\u002Frun_inference_checked.sh` to validate a repo root before running the\n  source inference entry point.\n- `scripts\u002Flaunch_gradio_checked.sh` to validate the same prerequisites before a\n  local demo launch.\n\n## Trigger phrases\n\nThis is usually the right branch when the user says things like:\n\n- “Generate from this reference object.”\n- “How do I run inference on a single image?”\n- “I need VITON-HD results.”\n- “Launch the local demo.”\n- “What does the Cog predictor expect?”\n- “Why is my mask or crop wrong?”\n- “Can the demo refine the reference mask?”\n\n## Input contract\n\nThe main generation route expects:\n\n- a reference image,\n- a reference mask or alpha channel,\n- a target\u002Fbackground image,\n- a target mask,\n- and the patched AnyDoor \u002F DINOv2 checkpoints.\n\nThe preprocessing contract is mask-aware:\n\n- masks must be non-empty,\n- masks must match the image size you intend to use,\n- and masks should be binary by the time they reach the generation step.\n\n## Output contract\n\nA successful generation route should explain:\n\n- where the customized image is saved,\n- whether the result is a single image or a composite visualization,\n- which guidance, strength, and step settings were used,\n- and whether the output comes from the standard or shape-control path.\n\n## Default workflow order\n\n1. Run the environment preflight from `setup-and-checkpoints` if you have not\n   already done so.\n2. Validate the input images and masks.\n3. Confirm the checkpoints are patched and accessible.\n4. Run the generation or demo launch command.\n5. If the result is wrong, inspect the mask and crop assumptions before changing\n   model settings.\n\n## Common decisions\n\n- **Reference image has alpha**: derive the reference mask from alpha when that\n  is the cleanest representation.\n- **Separate reference mask exists**: prefer explicit validation over guessing.\n- **Interactive mask refinement**: treat it as optional and explain the weight\n  requirement.\n- **Shape control**: explain that it changes how the target mask influences the\n  collage path.\n- **Cog deployment**: document the download\u002Fcache assumption so the user knows\n  when offline use will fail.\n\n## Troubleshooting focus\n\nThis branch owns symptoms such as:\n\n- empty or non-binary masks,\n- wrong crop placement,\n- target\u002Freference mask dimension mismatches,\n- checkpoint placeholders that were never patched,\n- demo launch issues,\n- Cog network download failures,\n- and generation failures that are really data-shape problems.\n\n## Handy reminders\n\n- The runtime defaults in the source scripts use 512x512 generation space.\n- Guidance scale and DDIM steps are part of the user-facing API; document them.\n- The local demo is intentionally coarse-mask sensitive.\n- A successful import check does not prove the generation path is ready.\n\n## Cross-links\n\nWhen the issue is actually about setup or checkpoints, route back to\n`setup-and-checkpoints`.\nWhen the issue is about dataset layout, route to `data-and-training`.\n\n## Quality bar\n\nA future agent should be able to answer generation and demo questions from this\nbranch alone, plus the bundled references and validation scripts, without\nreopening the source repository.\n","# Inference and Demo\n\nUse this sub-skill when the task is about generating a customized image,\nlaunching the local Gradio demo, or understanding the prediction interface.\n\n## What this sub-skill owns\n\n- Single-image AnyDoor inference.\n- VITON-HD-style batch inference.\n- Mask and image validation before generation.\n- Local Gradio launch behavior and interactive options.\n- Cog \u002F Replicate-style prediction input and output contracts.\n\n## What this sub-skill does not own\n\n- Environment installation and checkpoint patching.\n- Dataset format design or training data preparation.\n- Weight conversion and training launch details.\n\nRoute those to the sibling sub-skills.\n\n## Read first\n\n- `..\u002F..\u002Freferences\u002Foverview.md`\n- `..\u002F..\u002Freferences\u002Fcheckpoints-and-configs.md`\n- `..\u002F..\u002Freferences\u002Ftroubleshooting.md`\n- `references\u002Finference-workflows.md`\n- `references\u002Fmask-and-image-formats.md`\n- `references\u002Fdemo-and-cog-deployment.md`\n- `references\u002Ftroubleshooting.md`\n\n## Use these scripts\n\n- `scripts\u002Fvalidate_inference_inputs.py` to check image and mask pairs.\n- `scripts\u002Frun_inference_checked.sh` to validate a repo root before running the\n  source inference entry point.\n- `scripts\u002Flaunch_gradio_checked.sh` to validate the same prerequisites before a\n  local demo launch.\n\n## Trigger phrases\n\nThis is usually the right branch when the user says things like:\n\n- “Generate from this reference object.”\n- “How do I run inference on a single image?”\n- “I need VITON-HD results.”\n- “Launch the local demo.”\n- “What does the Cog predictor expect?”\n- “Why is my mask or crop wrong?”\n- “Can the demo refine the reference mask?”\n\n## Input contract\n\nThe main generation route expects:\n\n- a reference image,\n- a reference mask or alpha channel,\n- a target\u002Fbackground image,\n- a target mask,\n- and the patched AnyDoor \u002F DINOv2 checkpoints.\n\nThe preprocessing contract is mask-aware:\n\n- masks must be non-empty,\n- masks must match the image size you intend to use,\n- and masks should be binary by the time they reach the generation step.\n\n## Output contract\n\nA successful generation route should explain:\n\n- where the customized image is saved,\n- whether the result is a single image or a composite visualization,\n- which guidance, strength, and step settings were used,\n- and whether the output comes from the standard or shape-control path.\n\n## Default workflow order\n\n1. Run the environment preflight from `setup-and-checkpoints` if you have not\n   already done so.\n2. Validate the input images and masks.\n3. Confirm the checkpoints are patched and accessible.\n4. Run the generation or demo launch command.\n5. If the result is wrong, inspect the mask and crop assumptions before changing\n   model settings.\n\n## Common decisions\n\n- **Reference image has alpha**: derive the reference mask from alpha when that\n  is the cleanest representation.\n- **Separate reference mask exists**: prefer explicit validation over guessing.\n- **Interactive mask refinement**: treat it as optional and explain the weight\n  requirement.\n- **Shape control**: explain that it changes how the target mask influences the\n  collage path.\n- **Cog deployment**: document the download\u002Fcache assumption so the user knows\n  when offline use will fail.\n\n## Troubleshooting focus\n\nThis branch owns symptoms such as:\n\n- empty or non-binary masks,\n- wrong crop placement,\n- target\u002Freference mask dimension mismatches,\n- checkpoint placeholders that were never patched,\n- demo launch issues,\n- Cog network download failures,\n- and generation failures that are really data-shape problems.\n\n## Handy reminders\n\n- The runtime defaults in the source scripts use 512x512 generation space.\n- Guidance scale and DDIM steps are part of the user-facing API; document them.\n- The local demo is intentionally coarse-mask sensitive.\n- A successful import check does not prove the generation path is ready.\n\n## Cross-links\n\nWhen the issue is actually about setup or checkpoints, route back to\n`setup-and-checkpoints`.\nWhen the issue is about dataset layout, route to `data-and-training`.\n\n## Quality bar\n\nA future agent should be able to answer generation and demo questions from this\nbranch alone, plus the bundled references and validation scripts, without\nreopening the source repository.\n",{"name":5,"license":14,"metadata":23,"description":7,"disable-model-invocation":25},{"disco-role":24},"operating",true,"main","ac3fe1afa80fb9a09775ecfb2b6cc3ba850a2db6","skills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo",27,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab","2026-09-03 10:25:21",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":25},[35,38,41],{"tool":36,"command":37},"npx","npx skills add VectorSpaceLab\u002FAREX-Skill --skill inference-and-demo",{"tool":39,"command":40},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill inference-and-demo",{"tool":42,"command":43},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill inference-and-demo","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002FSKILL.md",[46,52,56,60,64,70,74],{"path":47,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":51},"references\u002Fdemo-and-cog-deployment.md","REFERENCE","参考文档",1672,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002Freferences\u002Fdemo-and-cog-deployment.md",{"path":53,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":54,"sourceHtmlUrl":55},"references\u002Finference-workflows.md",2228,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002Freferences\u002Finference-workflows.md",{"path":57,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002Fmask-and-image-formats.md",1929,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002Freferences\u002Fmask-and-image-formats.md",{"path":61,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":62,"sourceHtmlUrl":63},"references\u002Ftroubleshooting.md",1412,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002Freferences\u002Ftroubleshooting.md",{"path":65,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":68,"sourceHtmlUrl":69},"scripts\u002Flaunch_gradio_checked.sh","SCRIPT","脚本",976,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002Fscripts\u002Flaunch_gradio_checked.sh",{"path":71,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":72,"sourceHtmlUrl":73},"scripts\u002Frun_inference_checked.sh",970,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002Fscripts\u002Frun_inference_checked.sh",{"path":75,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":76,"sourceHtmlUrl":77},"scripts\u002Fvalidate_inference_inputs.py",4554,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fany-door\u002Fsub-skills\u002Finference-and-demo\u002Fscripts\u002Fvalidate_inference_inputs.py",[79,87,94,101,108,115],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":18},"01M799NYSH1R3RVAHSM2D88Y84","passnet-triton-opt","Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block\u002Fgrid\u002Fwarp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.","Apache-2.0","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-triton-opt\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":18},"01M799NS5APEC0N1FZ7XZVW6BN","passnet-solve","END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-solve\u002FSKILL.md",[],[],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":18},"01M799NNY9BKSM21V5R7QVTBJX","passnet-skill","PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under .\u002Fpass_dir\u002F, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-skill\u002FSKILL.md",[],[],{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":18},"01M799NHH6NP7Z1RK77KGXAN6S","passnet-pattern-fusion","Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()\u002Freplacement_args()\u002Freplacement_func() files, when a pass \"failed to match\", or when deciding how to fuse multiple kernels\u002Fops.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-pattern-fusion\u002FSKILL.md",[],[],{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":18},"01M799ND2J4WNC2PGAZ0A06NDF","passnet-orchestrate","Round\u002Fbatch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep\u002Frevert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-orchestrate\u002FSKILL.md",[],[],{"skillNo":116,"name":117,"nameZh":6,"description":118,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":119,"categories":120,"keywords":121,"publishedAt":18},"01M799N8SBTMY02ZDG2AZF5ZWE","passnet-feedback","Fast, deterministic feedback for PassNet work: pre-flight pattern\u002Fmatch verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-feedback\u002FSKILL.md",[],[]]