[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6DK5TGVWAV06S0QWA6J24CV":3,"skill-same:01M6DK5TGVWAV06S0QWA6J24CV":77},{"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":28,"publishedAt":48,"linkOnly":49,"markdownContent":50,"markdownBody":51,"frontMatter":52,"refName":58,"commitSha":59,"skillDirectory":5,"forksCount":14,"repositoryHtmlUrl":60,"repositoryHomepageUrl":6,"ownerHtmlUrl":61,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":62,"install":63,"resources":76},"01M6DK5TGVWAV06S0QWA6J24CV","bindcraft",null,"End-to-end binder design using BindCraft hallucination. Use this skill when: (1) Designing protein binders with built-in AF2 validation, (2) Running production-quality binder campaigns, (3) Using different design protocols (fast, default, slow), (4) Need joint backbone and sequence optimization, (5) Want high experimental success rate.\nFor backbone-only generation, use rfdiffusion. For QC thresholds, use protein-qc. For tool selection guidance, use binder-design.","huamu668\u002Fclawhub-openclaw-medical","huamu668","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F202473531?v=4","01M6DK1V7D7VP9PJ06Z1VXVK13","huamu","01M6DK1VBYG1BX5FT154AFZHAS",0,"MIT","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fbindcraft\u002FSKILL.md",[18,23],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000015","视觉设计","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[5,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47],"selection","binder","sequence-design","structure-design","end","generation","campaigns","structure","methods","design","default","only","use","quality","tool","pipeline","Optimization","https","protocols","2026-11-10 10:12:30",false,"---\nname: bindcraft\ndescription: >\n  End-to-end binder design using BindCraft hallucination. Use this skill when:\n  (1) Designing protein binders with built-in AF2 validation,\n  (2) Running production-quality binder campaigns,\n  (3) Using different design protocols (fast, default, slow),\n  (4) Need joint backbone and sequence optimization,\n  (5) Want high experimental success rate.\n\n  For backbone-only generation, use rfdiffusion.\n  For QC thresholds, use protein-qc.\n  For tool selection guidance, use binder-design.\nlicense: MIT\ncategory: design-tools\ntags: [structure-design, sequence-design, binder, pipeline]\nproteinbase_slug: bindcraft\nproteinbase_url: https:\u002F\u002Fproteinbase.com\u002Fdesign-methods\u002Fbindcraft\nbiomodals_script: modal_bindcraft.py\n---\n\n# BindCraft Binder Design\n\n## Prerequisites\n\n| Requirement | Minimum | Recommended |\n|-------------|---------|-------------|\n| Python | 3.9+ | 3.10 |\n| CUDA | 11.7+ | 12.0+ |\n| GPU VRAM | 32GB | 48GB (L40S) |\n| RAM | 32GB | 64GB |\n\n## How to run\n\n> **First time?** See [Installation Guide](..\u002F..\u002Fdocs\u002Finstallation.md) to set up Modal and biomodals.\n\n### Option 1: Modal (recommended)\n```bash\ncd biomodals\nmodal run modal_bindcraft.py \\\n  --target-pdb target.pdb \\\n  --target-chain A \\\n  --binder-lengths 70-100 \\\n  --hotspots \"A45,A67,A89\" \\\n  --num-designs 50\n```\n\n**GPU**: L40S (48GB) | **Timeout**: 3600s default\n\n### Option 2: Local installation\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fmartinpacesa\u002FBindCraft.git\ncd BindCraft\npip install -r requirements.txt\n\npython bindcraft.py \\\n  --target target.pdb \\\n  --target_chains A \\\n  --binder_lengths 70-100 \\\n  --hotspots A45,A67,A89 \\\n  --num_designs 50\n```\n\n## Key parameters\n\n| Parameter | Default | Range | Description |\n|-----------|---------|-------|-------------|\n| `--target-pdb` | required | path | Target structure |\n| `--target-chain` | required | A-Z | Target chain(s) |\n| `--binder-lengths` | 70-100 | 40-150 | Length range |\n| `--hotspots` | None | residues | Target hotspots |\n| `--num-designs` | 50 | 1-500 | Number of designs |\n| `--protocol` | default | fast\u002Fdefault\u002Fslow | Quality vs speed |\n\n## Protocols\n\n| Protocol | Speed | Quality | Use Case |\n|----------|-------|---------|----------|\n| fast | Fast | Lower | Initial screening |\n| default | Medium | Good | Standard campaigns |\n| slow | Slow | High | Final production |\n\n## Output format\n\n```\noutput\u002F\n├── design_0\u002F\n│   ├── binder.pdb         # Final design\n│   ├── complex.pdb        # Binder + target\n│   ├── metrics.json       # QC scores\n│   └── trajectory\u002F        # Optimization trajectory\n├── design_1\u002F\n│   └── ...\n└── summary.csv            # All metrics\n```\n\n### Metrics Output\n```json\n{\n  \"plddt\": 0.89,\n  \"ptm\": 0.78,\n  \"iptm\": 0.62,\n  \"pae\": 8.5,\n  \"rmsd\": 1.2,\n  \"sequence\": \"MKTAYIAK...\"\n}\n```\n\n## Sample output\n\n### Successful run\n```\n$ modal run modal_bindcraft.py --target-pdb target.pdb --num-designs 50\n[INFO] Loading BindCraft model...\n[INFO] Target: target.pdb (chain A)\n[INFO] Hotspots: A45, A67, A89\n[INFO] Protocol: default\n[INFO] Generating 50 designs...\n\nDesign 1\u002F50:\n  Length: 78 AA\n  pLDDT: 0.89, ipTM: 0.62\n  Saved: output\u002Fdesign_0\u002F\n\nDesign 50\u002F50:\n  Length: 85 AA\n  pLDDT: 0.86, ipTM: 0.58\n  Saved: output\u002Fdesign_49\u002F\n\n[INFO] Campaign complete. Summary: output\u002Fsummary.csv\nPass rate: 32\u002F50 (64%) with ipTM > 0.5\n```\n\n**What good output looks like:**\n- pLDDT: > 0.85 for most designs\n- ipTM: > 0.5 for passing designs\n- Pass rate: 30-70% depending on target\n- Diverse sequences across designs\n\n## Decision tree\n\n```\nShould I use BindCraft?\n│\n├─ What type of design?\n│  ├─ Production-quality binders → BindCraft ✓\n│  ├─ High diversity exploration → RFdiffusion\n│  └─ All-atom precision → BoltzGen\n│\n├─ What matters most?\n│  ├─ Experimental success rate → BindCraft ✓\n│  ├─ Speed \u002F diversity → RFdiffusion + ProteinMPNN\n│  ├─ AF2 gradient optimization → ColabDesign\n│  └─ All-atom control → BoltzGen\n│\n└─ Compute resources?\n   ├─ Have L40S\u002FA100 → BindCraft ✓\n   └─ Only A10G → RFdiffusion + ProteinMPNN\n```\n\n## Typical performance\n\n| Campaign Size | Time (L40S) | Cost (Modal) | Notes |\n|---------------|-------------|--------------|-------|\n| 50 designs | 2-4h | ~$15 | Quick campaign |\n| 100 designs | 4-8h | ~$30 | Standard |\n| 200 designs | 8-16h | ~$60 | Large campaign |\n\n**Expected pass rate**: 30-70% with ipTM > 0.5 (target-dependent).\n\n---\n\n## Verify\n\n```bash\nfind output -name \"binder.pdb\" | wc -l  # Should match num_designs\n```\n\n---\n\n## Troubleshooting\n\n**Low ipTM scores**: Check hotspot selection, increase designs\n**Slow convergence**: Use fast protocol for screening\n**OOM errors**: Reduce num_models, use L40S GPU\n**Poor diversity**: Lower sampling_temp, run multiple seeds\n\n### Error interpretation\n\n| Error | Cause | Fix |\n|-------|-------|-----|\n| `RuntimeError: CUDA out of memory` | Large target or long binder | Use L40S\u002FA100, reduce binder length |\n| `ValueError: no hotspots` | Hotspots not found | Check residue numbering |\n| `TimeoutError` | Design taking too long | Use fast protocol |\n\n---\n\n**Next**: Rank by `ipsae` → experimental validation.\n","# BindCraft Binder Design\n\n## Prerequisites\n\n| Requirement | Minimum | Recommended |\n|-------------|---------|-------------|\n| Python | 3.9+ | 3.10 |\n| CUDA | 11.7+ | 12.0+ |\n| GPU VRAM | 32GB | 48GB (L40S) |\n| RAM | 32GB | 64GB |\n\n## How to run\n\n> **First time?** See [Installation Guide](..\u002F..\u002Fdocs\u002Finstallation.md) to set up Modal and biomodals.\n\n### Option 1: Modal (recommended)\n```bash\ncd biomodals\nmodal run modal_bindcraft.py \\\n  --target-pdb target.pdb \\\n  --target-chain A \\\n  --binder-lengths 70-100 \\\n  --hotspots \"A45,A67,A89\" \\\n  --num-designs 50\n```\n\n**GPU**: L40S (48GB) | **Timeout**: 3600s default\n\n### Option 2: Local installation\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fmartinpacesa\u002FBindCraft.git\ncd BindCraft\npip install -r requirements.txt\n\npython bindcraft.py \\\n  --target target.pdb \\\n  --target_chains A \\\n  --binder_lengths 70-100 \\\n  --hotspots A45,A67,A89 \\\n  --num_designs 50\n```\n\n## Key parameters\n\n| Parameter | Default | Range | Description |\n|-----------|---------|-------|-------------|\n| `--target-pdb` | required | path | Target structure |\n| `--target-chain` | required | A-Z | Target chain(s) |\n| `--binder-lengths` | 70-100 | 40-150 | Length range |\n| `--hotspots` | None | residues | Target hotspots |\n| `--num-designs` | 50 | 1-500 | Number of designs |\n| `--protocol` | default | fast\u002Fdefault\u002Fslow | Quality vs speed |\n\n## Protocols\n\n| Protocol | Speed | Quality | Use Case |\n|----------|-------|---------|----------|\n| fast | Fast | Lower | Initial screening |\n| default | Medium | Good | Standard campaigns |\n| slow | Slow | High | Final production |\n\n## Output format\n\n```\noutput\u002F\n├── design_0\u002F\n│   ├── binder.pdb         # Final design\n│   ├── complex.pdb        # Binder + target\n│   ├── metrics.json       # QC scores\n│   └── trajectory\u002F        # Optimization trajectory\n├── design_1\u002F\n│   └── ...\n└── summary.csv            # All metrics\n```\n\n### Metrics Output\n```json\n{\n  \"plddt\": 0.89,\n  \"ptm\": 0.78,\n  \"iptm\": 0.62,\n  \"pae\": 8.5,\n  \"rmsd\": 1.2,\n  \"sequence\": \"MKTAYIAK...\"\n}\n```\n\n## Sample output\n\n### Successful run\n```\n$ modal run modal_bindcraft.py --target-pdb target.pdb --num-designs 50\n[INFO] Loading BindCraft model...\n[INFO] Target: target.pdb (chain A)\n[INFO] Hotspots: A45, A67, A89\n[INFO] Protocol: default\n[INFO] Generating 50 designs...\n\nDesign 1\u002F50:\n  Length: 78 AA\n  pLDDT: 0.89, ipTM: 0.62\n  Saved: output\u002Fdesign_0\u002F\n\nDesign 50\u002F50:\n  Length: 85 AA\n  pLDDT: 0.86, ipTM: 0.58\n  Saved: output\u002Fdesign_49\u002F\n\n[INFO] Campaign complete. Summary: output\u002Fsummary.csv\nPass rate: 32\u002F50 (64%) with ipTM > 0.5\n```\n\n**What good output looks like:**\n- pLDDT: > 0.85 for most designs\n- ipTM: > 0.5 for passing designs\n- Pass rate: 30-70% depending on target\n- Diverse sequences across designs\n\n## Decision tree\n\n```\nShould I use BindCraft?\n│\n├─ What type of design?\n│  ├─ Production-quality binders → BindCraft ✓\n│  ├─ High diversity exploration → RFdiffusion\n│  └─ All-atom precision → BoltzGen\n│\n├─ What matters most?\n│  ├─ Experimental success rate → BindCraft ✓\n│  ├─ Speed \u002F diversity → RFdiffusion + ProteinMPNN\n│  ├─ AF2 gradient optimization → ColabDesign\n│  └─ All-atom control → BoltzGen\n│\n└─ Compute resources?\n   ├─ Have L40S\u002FA100 → BindCraft ✓\n   └─ Only A10G → RFdiffusion + ProteinMPNN\n```\n\n## Typical performance\n\n| Campaign Size | Time (L40S) | Cost (Modal) | Notes |\n|---------------|-------------|--------------|-------|\n| 50 designs | 2-4h | ~$15 | Quick campaign |\n| 100 designs | 4-8h | ~$30 | Standard |\n| 200 designs | 8-16h | ~$60 | Large campaign |\n\n**Expected pass rate**: 30-70% with ipTM > 0.5 (target-dependent).\n\n---\n\n## Verify\n\n```bash\nfind output -name \"binder.pdb\" | wc -l  # Should match num_designs\n```\n\n---\n\n## Troubleshooting\n\n**Low ipTM scores**: Check hotspot selection, increase designs\n**Slow convergence**: Use fast protocol for screening\n**OOM errors**: Reduce num_models, use L40S GPU\n**Poor diversity**: Lower sampling_temp, run multiple seeds\n\n### Error interpretation\n\n| Error | Cause | Fix |\n|-------|-------|-----|\n| `RuntimeError: CUDA out of memory` | Large target or long binder | Use L40S\u002FA100, reduce binder length |\n| `ValueError: no hotspots` | Hotspots not found | Check residue numbering |\n| `TimeoutError` | Design taking too long | Use fast protocol |\n\n---\n\n**Next**: Rank by `ipsae` → experimental validation.\n",{"name":5,"tags":53,"license":15,"category":54,"description":55,"proteinbase_url":56,"biomodals_script":57,"proteinbase_slug":5},[32,31,30,44],"design-tools","End-to-end binder design using BindCraft hallucination. Use this skill when: (1) Designing protein binders with built-in AF2 validation, (2) Running production-quality binder campaigns, (3) Using different design protocols (fast, default, slow), (4) Need joint backbone and sequence optimization, (5) Want high experimental success rate.\nFor backbone-only generation, use rfdiffusion. For QC thresholds, use protein-qc. For tool selection guidance, use binder-design.\n","https:\u002F\u002Fproteinbase.com\u002Fdesign-methods\u002Fbindcraft","modal_bindcraft.py","main","d3998984163a67a891ee8397e35adf79b14e872d","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical","https:\u002F\u002Fgithub.com\u002Fhuamu668","2026-03-12 11:49:42",{"source":8,"skillName":5,"commands":64,"prompt":74,"zipAvailable":75},[65,68,71],{"tool":66,"command":67},"npx","npx skills add huamu668\u002Fclawhub-openclaw-medical --skill bindcraft",{"tool":69,"command":70},"bunx","bunx skills add huamu668\u002Fclawhub-openclaw-medical --skill bindcraft",{"tool":72,"command":73},"pnpm","pnpm dlx skills add huamu668\u002Fclawhub-openclaw-medical --skill bindcraft","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fbindcraft\u002FSKILL.md",true,[],[78,98,126,146,160,181],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":48},"01M6DR3JGAEX5G5FR7TK3D8NS8","zarr-python","Chunked N-D arrays for cloud storage. Compressed arrays, parallel I\u002FO, S3\u002FGCS integration, NumPy\u002FDask\u002FXarray compatible, for large-scale scientific computing pipelines.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fzarr-python\u002FSKILL.md",[],[85,86,87,88,89,90,91,92,93,94,95,96,97],"python","zarr python","storage","scientific computing","numpy","arrays","pipelines","parallel","dask","cloud","gcs","s3","integration",{"skillNo":99,"name":100,"nameZh":6,"description":101,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":102,"categories":103,"keywords":107,"publishedAt":48},"01M6DR101MEDE5XT2ZZ5VBEAA5","vaex","Use this skill for processing and analyzing large tabular datasets (billions of rows) that exceed available RAM. Vaex excels at out-of-core DataFrame operations, lazy evaluation, fast aggregations, efficient visualization of big data, and machine learning on large datasets. Apply when users need to work with large CSV\u002FHDF5\u002FArrow\u002FParquet files, perform fast statistics on massive datasets, create visualizations of big data, or build ML pipelines that do not fit in memory.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fvaex\u002FSKILL.md",[104],{"categoryNo":105,"categoryName":106,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000013","分析与可视化",[100,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,41,124,125],"machine learning","in memory","learning","csv","visualization","create","core","ml","big data","memory","visualizations","fit","statistics","do","build","files","operations","work",{"skillNo":127,"name":128,"nameZh":6,"description":129,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":130,"categories":131,"keywords":132,"publishedAt":48},"01M6DQVMB0GBKF2W44FKXVD1P7","torch-geometric","Graph Neural Networks (PyG). Node\u002Fgraph classification, link prediction, GCN, GAT, GraphSAGE, heterogeneous graphs, molecular property prediction, for geometric deep learning.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Ftorch-geometric\u002FSKILL.md",[],[133,110,134,135,136,137,138,139,140,141,142,143,144,145],"torch geometric","node","Neural","link","gat","molecular","graphs","deep","networks","graph","neural networks","deep learning","classification",{"skillNo":147,"name":148,"nameZh":6,"description":149,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":150,"categories":151,"keywords":152,"publishedAt":48},"01M6DQH2GZPZKEQ140GRCJ9746","tiledbvcf","Efficient storage and retrieval of genomic variant data using TileDB. Scalable VCF\u002FBCF ingestion, incremental sample addition, compressed storage, parallel queries, and export capabilities for population genomics.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Ftiledbvcf\u002FSKILL.md",[],[148,153,154,155,156,157,158,92,159,87],"data","genomics","sample","queries","ingestion","export","capabilities",{"skillNo":161,"name":162,"nameZh":6,"description":163,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":164,"categories":165,"keywords":168,"publishedAt":48},"01M6DQD9G958EM6NHJ994D0SK1","solublempnn","Solubility-optimized protein sequence design using SolubleMPNN. Use this skill when: (1) Designing for E. coli expression, (2) Optimizing solubility of designed proteins, (3) Reducing aggregation propensity, (4) Need high-yield expression, (5) Avoiding inclusion body formation.\nFor standard design, use proteinmpnn. For ligand-aware design, use ligandmpnn.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fsolublempnn\u002FSKILL.md",[166,167],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":24,"categoryName":25,"taxonomyType":26,"taxonomyTypeDesc":27},[162,169,170,41,171,172,173,38,174,175,176,177,178,179,180,31],"inclusion","yield","modal","Tools","aggregation","proteins","solubility","proteinmpnn","ligandmpnn","ligand-aware","inverse-folding","expression",{"skillNo":182,"name":183,"nameZh":6,"description":184,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":185,"licenseFileUrl":6,"sourceHtmlUrl":186,"categories":187,"keywords":188,"publishedAt":48},"01M6DQ8W91A93MGGCEEW949HBC","scvelo","RNA velocity analysis with scVelo. Estimate cell state transitions from unspliced\u002Fspliced mRNA dynamics, infer trajectory directions, compute latent time, and identify driver genes in single-cell RNA-seq data. Complements Scanpy\u002FscVI-tools for trajectory inference.","BSD-3-Clause","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fscvelo\u002FSKILL.md",[],[183,189,190,191,192,193,194,195,196,197,198,172,153,199,200],"RNA","single-cell","infer","cell","transitions","time","compute","velocity","scanpy","analysis","state","inference"]