[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M791WZPSKXR82EY1MPMFA9CB":3,"skill-same:01M791WZPSKXR82EY1MPMFA9CB":71},{"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":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":33,"install":34,"resources":46},"01M791WZPSKXR82EY1MPMFA9CB","openfold",null,"Route OpenFold protein-structure prediction, data preparation, training, installation, and programmatic API tasks to focused repo skill guidance.","VectorSpaceLab\u002FAREX-Skill","VectorSpaceLab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F180637102?v=4","01M7538C5WPSJGWF7KHR4EFZ9F","01M7538C6R2GFP53NPXTCZNBRT",351,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fopenfold\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: openfold\ndescription: \"Route OpenFold protein-structure prediction, data preparation,\n  training, installation, and programmatic API tasks to focused repo skill\n  guidance.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# OpenFold Repo Skill\n\nUse this repo skill when a task is about OpenFold: installing or validating its runtime, planning protein-structure inference, preparing FASTA\u002FmmCIF\u002FMSA inputs, training or fine-tuning models, using OpenFold Python APIs, debugging optional CUDA\u002FTensorRT\u002FDeepSpeed dependencies, or deciding which OpenFold workflow fits a protein modeling request.\n\n## First Checks\n\n1. Read `references\u002Frepo-provenance.md` when deciding whether this skill matches a current OpenFold checkout or should be refreshed.\n2. Read `references\u002Fcapability-map.md` for a compact route map from user tasks to sub-skills and bundled helpers.\n3. Run `scripts\u002Fcheck_openfold_imports.py --json` in the target environment for a package import\u002Fconfig\u002Fextension check before deeper workflows.\n4. Use `references\u002Ftroubleshooting.md` for cross-cutting install\u002Fimport\u002Fbackend\u002Fdata-asset triage, then route to the nearest sub-skill for workflow-specific recovery.\n\n## Route by Task\n\n| User task | Read next | Why |\n| --- | --- | --- |\n| Install OpenFold, plan parameters\u002Fdatabases, diagnose import\u002Fbuild\u002Fbackend failures, or validate runtime readiness | `sub-skills\u002Finstallation-assets\u002F` | Owns environment setup, assets, optional dependencies, external binaries, and safe environment checks. |\n| Build or validate monomer, multimer, SoloSeq, precomputed-alignment, custom-template, long-sequence, or threading inference commands | `sub-skills\u002Finference\u002F` | Owns `run_pretrained_openfold.py`, `thread_sequence.py`, inference flags, outputs, relaxation, and inference troubleshooting. |\n| Prepare FASTA, MSA, mmCIF, precomputed alignment, alignment DB, duplicate-chain, cluster-file, or cache inputs | `sub-skills\u002Fdata-preparation\u002F` | Owns input file layouts, parser facts, alignment\u002Fcache validators, and data-prep troubleshooting. |\n| Train or fine-tune OpenFold, configure validation\u002Fdistillation, resume checkpoints, plan DeepSpeed, or construct distributed training commands | `sub-skills\u002Ftraining\u002F` | Owns `train_openfold.py`, training prerequisites, DeepSpeed\u002Fdistributed options, checkpoints, and training command builders. |\n| Use OpenFold from Python, inspect config presets, instantiate model APIs, import\u002Fconvert weights, or reason about acceleration internals | `sub-skills\u002Fmodel-apis\u002F` | Owns verified\u002Fsource-backed API signatures, config presets, weights\u002Fcheckpoints, output objects, metrics, and optional acceleration APIs. |\n\n## Installation Baseline\n\n- Public package\u002Fdistribution name: `openfold`.\n- Source metadata version used for this skill: `2.2.0`.\n- Documented production target: Linux, Python 3.10, PyTorch 2, CUDA-class GPU runtime for realistic inference\u002Ftraining.\n- For a local checkout, install only after selecting a compatible PyTorch\u002FCUDA runtime, then use `python -m pip install --no-build-isolation -e .` so `setup.py` can import the already-installed `torch` while building the OpenFold extension.\n- OpenFold setup imports `torch` at build time and builds `attn_core_inplace_cuda` as a CUDA extension when CUDA\u002FNVCC are available, otherwise as a CPU stub.\n- Full inference\u002Ftraining usually requires large external parameters, sequence databases, template mmCIFs, HMMER\u002FHHSuite\u002FKalign binaries, and GPU resources; do not run downloads, database searches, model inference, or training without explicit user approval.\n\nMinimal package check:\n\n```bash\npython - \u003C\u003C'PY'\nfrom openfold.config import model_config\nfrom openfold.data.parsers import parse_fasta\ncfg = model_config('model_1')\ntags, seqs = parse_fasta('>query\\nACD\\n')\nprint('openfold config\u002Fparser ok', cfg.model.template.enabled, tags, seqs)\nPY\n```\n\nFor fuller diagnostics, use:\n\n```bash\npython scripts\u002Fcheck_openfold_imports.py --json\n```\n\n## Shared Rules\n\n- Keep inference and training commands as dry-run plans until parameters, databases, hardware, external binaries, and output paths are confirmed.\n- If `run_pretrained_openfold.py`, `train_openfold.py`, or model imports fail with `ModuleNotFoundError: attn_core_inplace_cuda`, repair\u002Frebuild the OpenFold install before claiming model or CLI readiness.\n- Use precomputed alignments to avoid database searches only after validating the alignment directory or alignment DB layout in `sub-skills\u002Fdata-preparation\u002F`.\n- Route missing TensorRT, `cuda.cudart`, DeepSpeed, cuEquivariance, FlashAttention, OpenMM, PDBFixer, HMMER, HHSuite, or Kalign errors through `sub-skills\u002Finstallation-assets\u002F` first.\n- Do not assume CPU-only execution is enough for full OpenFold training or production inference; CPU-only is mainly useful for import\u002Fconfig\u002Fdata validation.\n\n## Bundled Root Files\n\n- `references\u002Fcapability-map.md` maps capabilities to sub-skill owners and safe helpers.\n- `references\u002Ftroubleshooting.md` covers cross-cutting install\u002Fimport\u002Fruntime failures.\n- `references\u002Frepo-routing-metadata.json` is structured metadata consumed by DisCo import tooling.\n- `scripts\u002Fcheck_openfold_imports.py` performs lightweight import\u002Fconfig\u002Fmodule availability checks without model execution.\n","# OpenFold Repo Skill\n\nUse this repo skill when a task is about OpenFold: installing or validating its runtime, planning protein-structure inference, preparing FASTA\u002FmmCIF\u002FMSA inputs, training or fine-tuning models, using OpenFold Python APIs, debugging optional CUDA\u002FTensorRT\u002FDeepSpeed dependencies, or deciding which OpenFold workflow fits a protein modeling request.\n\n## First Checks\n\n1. Read `references\u002Frepo-provenance.md` when deciding whether this skill matches a current OpenFold checkout or should be refreshed.\n2. Read `references\u002Fcapability-map.md` for a compact route map from user tasks to sub-skills and bundled helpers.\n3. Run `scripts\u002Fcheck_openfold_imports.py --json` in the target environment for a package import\u002Fconfig\u002Fextension check before deeper workflows.\n4. Use `references\u002Ftroubleshooting.md` for cross-cutting install\u002Fimport\u002Fbackend\u002Fdata-asset triage, then route to the nearest sub-skill for workflow-specific recovery.\n\n## Route by Task\n\n| User task | Read next | Why |\n| --- | --- | --- |\n| Install OpenFold, plan parameters\u002Fdatabases, diagnose import\u002Fbuild\u002Fbackend failures, or validate runtime readiness | `sub-skills\u002Finstallation-assets\u002F` | Owns environment setup, assets, optional dependencies, external binaries, and safe environment checks. |\n| Build or validate monomer, multimer, SoloSeq, precomputed-alignment, custom-template, long-sequence, or threading inference commands | `sub-skills\u002Finference\u002F` | Owns `run_pretrained_openfold.py`, `thread_sequence.py`, inference flags, outputs, relaxation, and inference troubleshooting. |\n| Prepare FASTA, MSA, mmCIF, precomputed alignment, alignment DB, duplicate-chain, cluster-file, or cache inputs | `sub-skills\u002Fdata-preparation\u002F` | Owns input file layouts, parser facts, alignment\u002Fcache validators, and data-prep troubleshooting. |\n| Train or fine-tune OpenFold, configure validation\u002Fdistillation, resume checkpoints, plan DeepSpeed, or construct distributed training commands | `sub-skills\u002Ftraining\u002F` | Owns `train_openfold.py`, training prerequisites, DeepSpeed\u002Fdistributed options, checkpoints, and training command builders. |\n| Use OpenFold from Python, inspect config presets, instantiate model APIs, import\u002Fconvert weights, or reason about acceleration internals | `sub-skills\u002Fmodel-apis\u002F` | Owns verified\u002Fsource-backed API signatures, config presets, weights\u002Fcheckpoints, output objects, metrics, and optional acceleration APIs. |\n\n## Installation Baseline\n\n- Public package\u002Fdistribution name: `openfold`.\n- Source metadata version used for this skill: `2.2.0`.\n- Documented production target: Linux, Python 3.10, PyTorch 2, CUDA-class GPU runtime for realistic inference\u002Ftraining.\n- For a local checkout, install only after selecting a compatible PyTorch\u002FCUDA runtime, then use `python -m pip install --no-build-isolation -e .` so `setup.py` can import the already-installed `torch` while building the OpenFold extension.\n- OpenFold setup imports `torch` at build time and builds `attn_core_inplace_cuda` as a CUDA extension when CUDA\u002FNVCC are available, otherwise as a CPU stub.\n- Full inference\u002Ftraining usually requires large external parameters, sequence databases, template mmCIFs, HMMER\u002FHHSuite\u002FKalign binaries, and GPU resources; do not run downloads, database searches, model inference, or training without explicit user approval.\n\nMinimal package check:\n\n```bash\npython - \u003C\u003C'PY'\nfrom openfold.config import model_config\nfrom openfold.data.parsers import parse_fasta\ncfg = model_config('model_1')\ntags, seqs = parse_fasta('>query\\nACD\\n')\nprint('openfold config\u002Fparser ok', cfg.model.template.enabled, tags, seqs)\nPY\n```\n\nFor fuller diagnostics, use:\n\n```bash\npython scripts\u002Fcheck_openfold_imports.py --json\n```\n\n## Shared Rules\n\n- Keep inference and training commands as dry-run plans until parameters, databases, hardware, external binaries, and output paths are confirmed.\n- If `run_pretrained_openfold.py`, `train_openfold.py`, or model imports fail with `ModuleNotFoundError: attn_core_inplace_cuda`, repair\u002Frebuild the OpenFold install before claiming model or CLI readiness.\n- Use precomputed alignments to avoid database searches only after validating the alignment directory or alignment DB layout in `sub-skills\u002Fdata-preparation\u002F`.\n- Route missing TensorRT, `cuda.cudart`, DeepSpeed, cuEquivariance, FlashAttention, OpenMM, PDBFixer, HMMER, HHSuite, or Kalign errors through `sub-skills\u002Finstallation-assets\u002F` first.\n- Do not assume CPU-only execution is enough for full OpenFold training or production inference; CPU-only is mainly useful for import\u002Fconfig\u002Fdata validation.\n\n## Bundled Root Files\n\n- `references\u002Fcapability-map.md` maps capabilities to sub-skill owners and safe helpers.\n- `references\u002Ftroubleshooting.md` covers cross-cutting install\u002Fimport\u002Fruntime failures.\n- `references\u002Frepo-routing-metadata.json` is structured metadata consumed by DisCo import tooling.\n- `scripts\u002Fcheck_openfold_imports.py` performs lightweight import\u002Fconfig\u002Fmodule availability checks without model execution.\n",{"name":5,"license":23,"metadata":24,"description":7,"disable-model-invocation":26},"Apache 2.0",{"disco-role":25},"operating",true,"main","ac3fe1afa80fb9a09775ecfb2b6cc3ba850a2db6","skills\u002Frepositories\u002Frepo-skills\u002Fopenfold",27,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab","2026-09-03 10:25:21",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":26},[36,39,42],{"tool":37,"command":38},"npx","npx skills add VectorSpaceLab\u002FAREX-Skill --skill openfold",{"tool":40,"command":41},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill openfold",{"tool":43,"command":44},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill openfold","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fopenfold\u002FSKILL.md",[47,53,57,61,65],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"references\u002Fcapability-map.md","REFERENCE","参考文档",5063,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fopenfold\u002Freferences\u002Fcapability-map.md",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"references\u002Frepo-provenance.md",2356,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fopenfold\u002Freferences\u002Frepo-provenance.md",{"path":58,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":59,"sourceHtmlUrl":60},"references\u002Frepo-routing-metadata.json",326,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fopenfold\u002Freferences\u002Frepo-routing-metadata.json",{"path":62,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":63,"sourceHtmlUrl":64},"references\u002Ftroubleshooting.md",5005,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fopenfold\u002Freferences\u002Ftroubleshooting.md",{"path":66,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":69,"sourceHtmlUrl":70},"scripts\u002Fcheck_openfold_imports.py","SCRIPT","脚本",13019,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fopenfold\u002Fscripts\u002Fcheck_openfold_imports.py",[72,79,86,93,100,107],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"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.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-triton-opt\u002FSKILL.md",[],[],{"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":14,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"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":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"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":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"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":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"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":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":111,"categories":112,"keywords":113,"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",[],[]]