[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-microsoft-msresearch-bioemu":3,"mdc--sii3zw-key":34,"related-repo-microsoft-msresearch-bioemu":738,"related-org-microsoft-msresearch-bioemu":778},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"msresearch-bioemu","generate protein conformational ensembles with BioEmu","Microsoft Research's BioEmu — generates protein conformational ensembles from amino acid sequence on a local GPU. Use when users mention BioEmu by name, want to predict protein dynamics or flexibility from sequence, or are setting up \u002F running \u002F troubleshooting BioEmu locally.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"microsoft","Microsoft","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fmicrosoft.png",[12,16,19,22],{"name":13,"slug":14,"type":15},"Deep Learning","deep-learning","tag",{"name":17,"slug":18,"type":15},"Life Sciences","life-sciences",{"name":20,"slug":21,"type":15},"Bioinformatics","bioinformatics",{"name":9,"slug":8,"type":15},34,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fvibe-kit","2026-07-07T06:53:46.003518","MIT",7,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"Microsoft Research Vibe Kit is a customized vibe coding environment to enable faster prototyping and integration of Microsoft Research innovations using GitHub Copilot inside Visual Studio Code. ","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fvibe-kit\u002Ftree\u002FHEAD\u002Fskills\u002Fmsresearch-bioemu","---\nname: msresearch-bioemu\ndescription: Microsoft Research's BioEmu — generates protein conformational ensembles from amino acid sequence on a local GPU. Use when users mention BioEmu by name, want to predict protein dynamics or flexibility from sequence, or are setting up \u002F running \u002F troubleshooting BioEmu locally.\nlicense: MIT\n---\n\n## Scope\n\n- Run BioEmu locally on a CUDA-capable NVIDIA GPU to sample protein conformational ensembles from amino acid sequence.\n- Two supported entry points:\n  1. **CLI** — `pip install bioemu[cuda]` and `python -m bioemu.sample` directly from the upstream package.\n  2. **Reference app** — the React + Flask app in `assets\u002Freference-app\u002F`. Run its bundled `score\u002F` Flask server locally on `:5001`, point the proxy backend at it via `.env`, and use the React frontend for visualization.\n- Analyze trajectories with MDTraj (RMSD, RMSF, Rg, secondary structure) and visualize with Molstar in the reference app.\n\n## Prerequisites\n\n- Linux (or **WSL2 on Windows 11** with NVIDIA GPU passthrough — install WSL per [Microsoft's guide](https:\u002F\u002Flearn.microsoft.com\u002Fwindows\u002Fwsl\u002Finstall) and follow [NVIDIA's CUDA-on-WSL guide](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Fwsl-user-guide\u002Findex.html) for driver setup) with a CUDA-capable NVIDIA GPU. CPU works only for ~10-residue toy sequences; everything else is unusably slow.\n- Python 3.10+, ~5 GB free disk for cached weights (AlphaFold2 weights ~3.5 GB + BioEmu checkpoint + working space) at `~\u002F.cache\u002Fcolabfold\u002F` (CLI) or `\u002Fapp\u002Fcolabfold_cache` (Docker).\n- Path B only: Docker with GPU support (Docker Desktop on Windows includes this; Linux needs `nvidia-container-toolkit`), or system Python for a bare-metal `score\u002F` run; Node.js 18+ for the frontend.\n\n## Workflow\n\n1. Load `docs\u002Fabout-bioemu.md` when users ask what BioEmu is, how it works, or need scientific background.\n2. Follow `docs\u002Fquick-start.md` to get a first ensemble locally — Path A (CLI smoke test) before Path B (reference app).\n3. For code examples (Python sampling API, MDTraj analysis, AlphaFold comparison, output file formats), route to `docs\u002Fapplication-patterns.md`.\n4. Route errors to `docs\u002Ftroubleshooting.md`.\n\n### Operating rules\n\n- **Windows users:** BioEmu is Linux-only. On Windows, all commands must run inside a WSL2 distro (Ubuntu recommended) with NVIDIA drivers installed on the **Windows host** (not inside WSL). Native Windows Python and PowerShell are not supported. Point Windows users at [Microsoft's WSL install guide](https:\u002F\u002Flearn.microsoft.com\u002Fwindows\u002Fwsl\u002Finstall) and [NVIDIA's CUDA-on-WSL guide](https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Fwsl-user-guide\u002Findex.html) before running any quick-start step.\n- **GPU probe first:** Before suggesting any non-trivial sampling, run `nvidia-smi` and confirm a CUDA-capable GPU is visible. If none is present, warn the user that BioEmu is unusable beyond ~10-residue toy sequences and stop — do not proceed to install.\n- **Weight download:** First call downloads ~3.5 GB of AlphaFold2 + BioEmu weights. Tell the user this will be slow and must not be interrupted. Subsequent runs reuse the cache.\n- **Install times:** `pip install bioemu[cuda]`, `pip install -r requirements.txt`, and `npm install --legacy-peer-deps` are all slow on first run. Do not interrupt installs.\n- **Three terminals (Path B):** `score\u002F` server, proxy backend, and frontend each run in their own terminal and must stay running. Never run other commands in a terminal hosting a live server.\n- **Local credentials:** For Path B the `.env` keeps its `AZURE_BIOEMU_*` variable names because the proxy reads those exact names — but `AZURE_BIOEMU_ENDPOINT` should point at `http:\u002F\u002Flocalhost:5001\u002Fscore` and `AZURE_BIOEMU_KEY` can be any non-empty string (the local `score\u002F` server doesn't enforce auth). Explain this when guiding users; never ask them to paste secrets into chat.\n- **Honest scope:** This is *local inference*, not *fully offline*. ColabFold MSA generation still hits an external MMseqs2 server on first use of a new sequence. Do not claim air-gapped operation.\n- **Execute, don't display:** When terminal execution is available, run quick-start commands directly rather than printing bash blocks for the user to copy.\n- **Always offer the next step:** After loading any explainer-style doc (e.g. `docs\u002Fabout-bioemu.md`), end your response with a concrete offer to advance the user along the Learning Path. Default phrasing: *\"Want to try BioEmu? I can walk you through running it locally — three commands gets you a first ensemble on your GPU, and from there we can wire up the full reference app UI if you want it.\"* Adapt wording to context, but never end an explainer response without a concrete next-step offer.\n\n## Routing\n\n| Doc | When to load |\n|---|---|\n| `docs\u002Fabout-bioemu.md` | User asks what BioEmu is, how it works, performance metrics, limitations, or the scientific FAQ |\n| `docs\u002Fquick-start.md` | User wants to run BioEmu locally (CLI or full reference app) |\n| `docs\u002Fapplication-patterns.md` | User wants code examples for sampling, MDTraj analysis, output file formats, or AlphaFold comparison |\n| `docs\u002Ftroubleshooting.md` | User hits an error, missing GPU, weight download stall, port conflict, or MSA timeout |\n\n## Learning Path\n\n1. `docs\u002Fabout-bioemu.md` — Understand what BioEmu is and why it matters\n2. `docs\u002Fquick-start.md` Path A — Three-command CLI smoke test on your GPU\n3. `docs\u002Fquick-start.md` Path B — Run the full reference app against a local `score\u002F` server\n4. `docs\u002Fapplication-patterns.md` — Build your own sampling and analysis workflows\n\n## Reference Links\n\n- Science paper: https:\u002F\u002Fwww.science.org\u002Fdoi\u002F10.1126\u002Fscience.adv9817\n- GitHub: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fbioemu\n- Hugging Face: https:\u002F\u002Fhuggingface.co\u002Fmicrosoft\u002Fbioemu\n\n## Assets\n\n- `assets\u002Freference-app\u002F` — Self-contained React + Flask + `score\u002F` Docker app. This is Path B; no separate clone needed.\n",{"data":35,"body":36},{"name":4,"description":6,"license":26},{"type":37,"children":38},"root",[39,48,142,148,224,230,284,291,516,522,614,620,670,676,712,718],{"type":40,"tag":41,"props":42,"children":44},"element","h2",{"id":43},"scope",[45],{"type":46,"value":47},"text","Scope",{"type":40,"tag":49,"props":50,"children":51},"ul",{},[52,58,137],{"type":40,"tag":53,"props":54,"children":55},"li",{},[56],{"type":46,"value":57},"Run BioEmu locally on a CUDA-capable NVIDIA GPU to sample protein conformational ensembles from amino acid sequence.",{"type":40,"tag":53,"props":59,"children":60},{},[61,63],{"type":46,"value":62},"Two supported entry points:\n",{"type":40,"tag":64,"props":65,"children":66},"ol",{},[67,95],{"type":40,"tag":53,"props":68,"children":69},{},[70,76,78,85,87,93],{"type":40,"tag":71,"props":72,"children":73},"strong",{},[74],{"type":46,"value":75},"CLI",{"type":46,"value":77}," — ",{"type":40,"tag":79,"props":80,"children":82},"code",{"className":81},[],[83],{"type":46,"value":84},"pip install bioemu[cuda]",{"type":46,"value":86}," and ",{"type":40,"tag":79,"props":88,"children":90},{"className":89},[],[91],{"type":46,"value":92},"python -m bioemu.sample",{"type":46,"value":94}," directly from the upstream package.",{"type":40,"tag":53,"props":96,"children":97},{},[98,103,105,111,113,119,121,127,129,135],{"type":40,"tag":71,"props":99,"children":100},{},[101],{"type":46,"value":102},"Reference app",{"type":46,"value":104}," — the React + Flask app in ",{"type":40,"tag":79,"props":106,"children":108},{"className":107},[],[109],{"type":46,"value":110},"assets\u002Freference-app\u002F",{"type":46,"value":112},". Run its bundled ",{"type":40,"tag":79,"props":114,"children":116},{"className":115},[],[117],{"type":46,"value":118},"score\u002F",{"type":46,"value":120}," Flask server locally on ",{"type":40,"tag":79,"props":122,"children":124},{"className":123},[],[125],{"type":46,"value":126},":5001",{"type":46,"value":128},", point the proxy backend at it via ",{"type":40,"tag":79,"props":130,"children":132},{"className":131},[],[133],{"type":46,"value":134},".env",{"type":46,"value":136},", and use the React frontend for visualization.",{"type":40,"tag":53,"props":138,"children":139},{},[140],{"type":46,"value":141},"Analyze trajectories with MDTraj (RMSD, RMSF, Rg, secondary structure) and visualize with Molstar in the reference app.",{"type":40,"tag":41,"props":143,"children":145},{"id":144},"prerequisites",[146],{"type":46,"value":147},"Prerequisites",{"type":40,"tag":49,"props":149,"children":150},{},[151,183,204],{"type":40,"tag":53,"props":152,"children":153},{},[154,156,161,163,172,174,181],{"type":46,"value":155},"Linux (or ",{"type":40,"tag":71,"props":157,"children":158},{},[159],{"type":46,"value":160},"WSL2 on Windows 11",{"type":46,"value":162}," with NVIDIA GPU passthrough — install WSL per ",{"type":40,"tag":164,"props":165,"children":169},"a",{"href":166,"rel":167},"https:\u002F\u002Flearn.microsoft.com\u002Fwindows\u002Fwsl\u002Finstall",[168],"nofollow",[170],{"type":46,"value":171},"Microsoft's guide",{"type":46,"value":173}," and follow ",{"type":40,"tag":164,"props":175,"children":178},{"href":176,"rel":177},"https:\u002F\u002Fdocs.nvidia.com\u002Fcuda\u002Fwsl-user-guide\u002Findex.html",[168],[179],{"type":46,"value":180},"NVIDIA's CUDA-on-WSL guide",{"type":46,"value":182}," for driver setup) with a CUDA-capable NVIDIA GPU. CPU works only for ~10-residue toy sequences; everything else is unusably slow.",{"type":40,"tag":53,"props":184,"children":185},{},[186,188,194,196,202],{"type":46,"value":187},"Python 3.10+, ~5 GB free disk for cached weights (AlphaFold2 weights ~3.5 GB + BioEmu checkpoint + working space) at ",{"type":40,"tag":79,"props":189,"children":191},{"className":190},[],[192],{"type":46,"value":193},"~\u002F.cache\u002Fcolabfold\u002F",{"type":46,"value":195}," (CLI) or ",{"type":40,"tag":79,"props":197,"children":199},{"className":198},[],[200],{"type":46,"value":201},"\u002Fapp\u002Fcolabfold_cache",{"type":46,"value":203}," (Docker).",{"type":40,"tag":53,"props":205,"children":206},{},[207,209,215,217,222],{"type":46,"value":208},"Path B only: Docker with GPU support (Docker Desktop on Windows includes this; Linux needs ",{"type":40,"tag":79,"props":210,"children":212},{"className":211},[],[213],{"type":46,"value":214},"nvidia-container-toolkit",{"type":46,"value":216},"), or system Python for a bare-metal ",{"type":40,"tag":79,"props":218,"children":220},{"className":219},[],[221],{"type":46,"value":118},{"type":46,"value":223}," run; Node.js 18+ for the frontend.",{"type":40,"tag":41,"props":225,"children":227},{"id":226},"workflow",[228],{"type":46,"value":229},"Workflow",{"type":40,"tag":64,"props":231,"children":232},{},[233,246,259,272],{"type":40,"tag":53,"props":234,"children":235},{},[236,238,244],{"type":46,"value":237},"Load ",{"type":40,"tag":79,"props":239,"children":241},{"className":240},[],[242],{"type":46,"value":243},"docs\u002Fabout-bioemu.md",{"type":46,"value":245}," when users ask what BioEmu is, how it works, or need scientific background.",{"type":40,"tag":53,"props":247,"children":248},{},[249,251,257],{"type":46,"value":250},"Follow ",{"type":40,"tag":79,"props":252,"children":254},{"className":253},[],[255],{"type":46,"value":256},"docs\u002Fquick-start.md",{"type":46,"value":258}," to get a first ensemble locally — Path A (CLI smoke test) before Path B (reference app).",{"type":40,"tag":53,"props":260,"children":261},{},[262,264,270],{"type":46,"value":263},"For code examples (Python sampling API, MDTraj analysis, AlphaFold comparison, output file formats), route to ",{"type":40,"tag":79,"props":265,"children":267},{"className":266},[],[268],{"type":46,"value":269},"docs\u002Fapplication-patterns.md",{"type":46,"value":271},".",{"type":40,"tag":53,"props":273,"children":274},{},[275,277,283],{"type":46,"value":276},"Route errors to ",{"type":40,"tag":79,"props":278,"children":280},{"className":279},[],[281],{"type":46,"value":282},"docs\u002Ftroubleshooting.md",{"type":46,"value":271},{"type":40,"tag":285,"props":286,"children":288},"h3",{"id":287},"operating-rules",[289],{"type":46,"value":290},"Operating rules",{"type":40,"tag":49,"props":292,"children":293},{},[294,325,343,353,386,402,457,482,492],{"type":40,"tag":53,"props":295,"children":296},{},[297,302,304,309,311,317,318,323],{"type":40,"tag":71,"props":298,"children":299},{},[300],{"type":46,"value":301},"Windows users:",{"type":46,"value":303}," BioEmu is Linux-only. On Windows, all commands must run inside a WSL2 distro (Ubuntu recommended) with NVIDIA drivers installed on the ",{"type":40,"tag":71,"props":305,"children":306},{},[307],{"type":46,"value":308},"Windows host",{"type":46,"value":310}," (not inside WSL). Native Windows Python and PowerShell are not supported. Point Windows users at ",{"type":40,"tag":164,"props":312,"children":314},{"href":166,"rel":313},[168],[315],{"type":46,"value":316},"Microsoft's WSL install guide",{"type":46,"value":86},{"type":40,"tag":164,"props":319,"children":321},{"href":176,"rel":320},[168],[322],{"type":46,"value":180},{"type":46,"value":324}," before running any quick-start step.",{"type":40,"tag":53,"props":326,"children":327},{},[328,333,335,341],{"type":40,"tag":71,"props":329,"children":330},{},[331],{"type":46,"value":332},"GPU probe first:",{"type":46,"value":334}," Before suggesting any non-trivial sampling, run ",{"type":40,"tag":79,"props":336,"children":338},{"className":337},[],[339],{"type":46,"value":340},"nvidia-smi",{"type":46,"value":342}," and confirm a CUDA-capable GPU is visible. If none is present, warn the user that BioEmu is unusable beyond ~10-residue toy sequences and stop — do not proceed to install.",{"type":40,"tag":53,"props":344,"children":345},{},[346,351],{"type":40,"tag":71,"props":347,"children":348},{},[349],{"type":46,"value":350},"Weight download:",{"type":46,"value":352}," First call downloads ~3.5 GB of AlphaFold2 + BioEmu weights. Tell the user this will be slow and must not be interrupted. Subsequent runs reuse the cache.",{"type":40,"tag":53,"props":354,"children":355},{},[356,361,363,368,370,376,378,384],{"type":40,"tag":71,"props":357,"children":358},{},[359],{"type":46,"value":360},"Install times:",{"type":46,"value":362}," ",{"type":40,"tag":79,"props":364,"children":366},{"className":365},[],[367],{"type":46,"value":84},{"type":46,"value":369},", ",{"type":40,"tag":79,"props":371,"children":373},{"className":372},[],[374],{"type":46,"value":375},"pip install -r requirements.txt",{"type":46,"value":377},", and ",{"type":40,"tag":79,"props":379,"children":381},{"className":380},[],[382],{"type":46,"value":383},"npm install --legacy-peer-deps",{"type":46,"value":385}," are all slow on first run. Do not interrupt installs.",{"type":40,"tag":53,"props":387,"children":388},{},[389,394,395,400],{"type":40,"tag":71,"props":390,"children":391},{},[392],{"type":46,"value":393},"Three terminals (Path B):",{"type":46,"value":362},{"type":40,"tag":79,"props":396,"children":398},{"className":397},[],[399],{"type":46,"value":118},{"type":46,"value":401}," server, proxy backend, and frontend each run in their own terminal and must stay running. Never run other commands in a terminal hosting a live server.",{"type":40,"tag":53,"props":403,"children":404},{},[405,410,412,417,419,425,427,433,435,441,442,448,450,455],{"type":40,"tag":71,"props":406,"children":407},{},[408],{"type":46,"value":409},"Local credentials:",{"type":46,"value":411}," For Path B the ",{"type":40,"tag":79,"props":413,"children":415},{"className":414},[],[416],{"type":46,"value":134},{"type":46,"value":418}," keeps its ",{"type":40,"tag":79,"props":420,"children":422},{"className":421},[],[423],{"type":46,"value":424},"AZURE_BIOEMU_*",{"type":46,"value":426}," variable names because the proxy reads those exact names — but ",{"type":40,"tag":79,"props":428,"children":430},{"className":429},[],[431],{"type":46,"value":432},"AZURE_BIOEMU_ENDPOINT",{"type":46,"value":434}," should point at ",{"type":40,"tag":79,"props":436,"children":438},{"className":437},[],[439],{"type":46,"value":440},"http:\u002F\u002Flocalhost:5001\u002Fscore",{"type":46,"value":86},{"type":40,"tag":79,"props":443,"children":445},{"className":444},[],[446],{"type":46,"value":447},"AZURE_BIOEMU_KEY",{"type":46,"value":449}," can be any non-empty string (the local ",{"type":40,"tag":79,"props":451,"children":453},{"className":452},[],[454],{"type":46,"value":118},{"type":46,"value":456}," server doesn't enforce auth). Explain this when guiding users; never ask them to paste secrets into chat.",{"type":40,"tag":53,"props":458,"children":459},{},[460,465,467,473,475,480],{"type":40,"tag":71,"props":461,"children":462},{},[463],{"type":46,"value":464},"Honest scope:",{"type":46,"value":466}," This is ",{"type":40,"tag":468,"props":469,"children":470},"em",{},[471],{"type":46,"value":472},"local inference",{"type":46,"value":474},", not ",{"type":40,"tag":468,"props":476,"children":477},{},[478],{"type":46,"value":479},"fully offline",{"type":46,"value":481},". ColabFold MSA generation still hits an external MMseqs2 server on first use of a new sequence. Do not claim air-gapped operation.",{"type":40,"tag":53,"props":483,"children":484},{},[485,490],{"type":40,"tag":71,"props":486,"children":487},{},[488],{"type":46,"value":489},"Execute, don't display:",{"type":46,"value":491}," When terminal execution is available, run quick-start commands directly rather than printing bash blocks for the user to copy.",{"type":40,"tag":53,"props":493,"children":494},{},[495,500,502,507,509,514],{"type":40,"tag":71,"props":496,"children":497},{},[498],{"type":46,"value":499},"Always offer the next step:",{"type":46,"value":501}," After loading any explainer-style doc (e.g. ",{"type":40,"tag":79,"props":503,"children":505},{"className":504},[],[506],{"type":46,"value":243},{"type":46,"value":508},"), end your response with a concrete offer to advance the user along the Learning Path. Default phrasing: ",{"type":40,"tag":468,"props":510,"children":511},{},[512],{"type":46,"value":513},"\"Want to try BioEmu? I can walk you through running it locally — three commands gets you a first ensemble on your GPU, and from there we can wire up the full reference app UI if you want it.\"",{"type":46,"value":515}," Adapt wording to context, but never end an explainer response without a concrete next-step offer.",{"type":40,"tag":41,"props":517,"children":519},{"id":518},"routing",[520],{"type":46,"value":521},"Routing",{"type":40,"tag":523,"props":524,"children":525},"table",{},[526,545],{"type":40,"tag":527,"props":528,"children":529},"thead",{},[530],{"type":40,"tag":531,"props":532,"children":533},"tr",{},[534,540],{"type":40,"tag":535,"props":536,"children":537},"th",{},[538],{"type":46,"value":539},"Doc",{"type":40,"tag":535,"props":541,"children":542},{},[543],{"type":46,"value":544},"When to load",{"type":40,"tag":546,"props":547,"children":548},"tbody",{},[549,566,582,598],{"type":40,"tag":531,"props":550,"children":551},{},[552,561],{"type":40,"tag":553,"props":554,"children":555},"td",{},[556],{"type":40,"tag":79,"props":557,"children":559},{"className":558},[],[560],{"type":46,"value":243},{"type":40,"tag":553,"props":562,"children":563},{},[564],{"type":46,"value":565},"User asks what BioEmu is, how it works, performance metrics, limitations, or the scientific FAQ",{"type":40,"tag":531,"props":567,"children":568},{},[569,577],{"type":40,"tag":553,"props":570,"children":571},{},[572],{"type":40,"tag":79,"props":573,"children":575},{"className":574},[],[576],{"type":46,"value":256},{"type":40,"tag":553,"props":578,"children":579},{},[580],{"type":46,"value":581},"User wants to run BioEmu locally (CLI or full reference app)",{"type":40,"tag":531,"props":583,"children":584},{},[585,593],{"type":40,"tag":553,"props":586,"children":587},{},[588],{"type":40,"tag":79,"props":589,"children":591},{"className":590},[],[592],{"type":46,"value":269},{"type":40,"tag":553,"props":594,"children":595},{},[596],{"type":46,"value":597},"User wants code examples for sampling, MDTraj analysis, output file formats, or AlphaFold comparison",{"type":40,"tag":531,"props":599,"children":600},{},[601,609],{"type":40,"tag":553,"props":602,"children":603},{},[604],{"type":40,"tag":79,"props":605,"children":607},{"className":606},[],[608],{"type":46,"value":282},{"type":40,"tag":553,"props":610,"children":611},{},[612],{"type":46,"value":613},"User hits an error, missing GPU, weight download stall, port conflict, or MSA timeout",{"type":40,"tag":41,"props":615,"children":617},{"id":616},"learning-path",[618],{"type":46,"value":619},"Learning Path",{"type":40,"tag":64,"props":621,"children":622},{},[623,633,643,660],{"type":40,"tag":53,"props":624,"children":625},{},[626,631],{"type":40,"tag":79,"props":627,"children":629},{"className":628},[],[630],{"type":46,"value":243},{"type":46,"value":632}," — Understand what BioEmu is and why it matters",{"type":40,"tag":53,"props":634,"children":635},{},[636,641],{"type":40,"tag":79,"props":637,"children":639},{"className":638},[],[640],{"type":46,"value":256},{"type":46,"value":642}," Path A — Three-command CLI smoke test on your GPU",{"type":40,"tag":53,"props":644,"children":645},{},[646,651,653,658],{"type":40,"tag":79,"props":647,"children":649},{"className":648},[],[650],{"type":46,"value":256},{"type":46,"value":652}," Path B — Run the full reference app against a local ",{"type":40,"tag":79,"props":654,"children":656},{"className":655},[],[657],{"type":46,"value":118},{"type":46,"value":659}," server",{"type":40,"tag":53,"props":661,"children":662},{},[663,668],{"type":40,"tag":79,"props":664,"children":666},{"className":665},[],[667],{"type":46,"value":269},{"type":46,"value":669}," — Build your own sampling and analysis workflows",{"type":40,"tag":41,"props":671,"children":673},{"id":672},"reference-links",[674],{"type":46,"value":675},"Reference Links",{"type":40,"tag":49,"props":677,"children":678},{},[679,690,701],{"type":40,"tag":53,"props":680,"children":681},{},[682,684],{"type":46,"value":683},"Science paper: ",{"type":40,"tag":164,"props":685,"children":688},{"href":686,"rel":687},"https:\u002F\u002Fwww.science.org\u002Fdoi\u002F10.1126\u002Fscience.adv9817",[168],[689],{"type":46,"value":686},{"type":40,"tag":53,"props":691,"children":692},{},[693,695],{"type":46,"value":694},"GitHub: ",{"type":40,"tag":164,"props":696,"children":699},{"href":697,"rel":698},"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fbioemu",[168],[700],{"type":46,"value":697},{"type":40,"tag":53,"props":702,"children":703},{},[704,706],{"type":46,"value":705},"Hugging Face: ",{"type":40,"tag":164,"props":707,"children":710},{"href":708,"rel":709},"https:\u002F\u002Fhuggingface.co\u002Fmicrosoft\u002Fbioemu",[168],[711],{"type":46,"value":708},{"type":40,"tag":41,"props":713,"children":715},{"id":714},"assets",[716],{"type":46,"value":717},"Assets",{"type":40,"tag":49,"props":719,"children":720},{},[721],{"type":40,"tag":53,"props":722,"children":723},{},[724,729,731,736],{"type":40,"tag":79,"props":725,"children":727},{"className":726},[],[728],{"type":46,"value":110},{"type":46,"value":730}," — Self-contained React + Flask + ",{"type":40,"tag":79,"props":732,"children":734},{"className":733},[],[735],{"type":46,"value":118},{"type":46,"value":737}," Docker app. This is Path B; no separate clone needed.",{"items":739,"total":777},[740,757,764],{"slug":741,"name":741,"fn":742,"description":743,"org":744,"tags":745,"stars":23,"repoUrl":24,"updatedAt":756},"msresearch-aurora","forecast weather with Aurora foundation model","Aurora — Microsoft's foundation model for the Earth system. Use this skill for both inference (regional weather\u002Fclimate forecasting, Norway example, ERA5\u002FCDS data, wind\u002Ftemperature\u002Fpressure mapping) and fine-tuning Aurora on new variables, regions, or domains (training, evaluation, checkpoint extension, gradient\u002Floss debugging). Triggers on Aurora, Microsoft Aurora, foundation models for Earth systems, ERA5\u002FCDS\u002FNetCDF weather data, weather forecast validation, and adapting Aurora to new use cases.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[746,749,750,753],{"name":747,"slug":748,"type":15},"LLM","llm",{"name":9,"slug":8,"type":15},{"name":751,"slug":752,"type":15},"Physics","physics",{"name":754,"slug":755,"type":15},"Research","research","2026-07-03T16:32:28.677219",{"slug":4,"name":4,"fn":5,"description":6,"org":758,"tags":759,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[760,761,762,763],{"name":20,"slug":21,"type":15},{"name":13,"slug":14,"type":15},{"name":17,"slug":18,"type":15},{"name":9,"slug":8,"type":15},{"slug":765,"name":765,"fn":766,"description":767,"org":768,"tags":769,"stars":23,"repoUrl":24,"updatedAt":776},"msresearch-dayhoff","generate and score protein sequences with Dayhoff","Microsoft Research's Dayhoff — generates novel protein sequences and scores variants using hybrid Mamba+Transformer models trained on the 3.34B-sequence Dayhoff Atlas. Use when users mention Dayhoff by name, want to generate protein sequences from scratch or from a prompt, score mutations \u002F variants \u002F fitness without experimental data, design sequences in a protein family from homologs, scaffold sequences around a motif, or are setting up \u002F running \u002F troubleshooting the Dayhoff prototype locally.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[770,771,772,773],{"name":20,"slug":21,"type":15},{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},{"name":774,"slug":775,"type":15},"Proteomics","proteomics","2026-07-07T06:53:47.296995",3,{"items":779,"total":972},[780,802,821,842,857,874,885,898,913,928,947,960],{"slug":781,"name":781,"fn":782,"description":783,"org":784,"tags":785,"stars":799,"repoUrl":800,"updatedAt":801},"rushstack-best-practices","manage Rush monorepos with best practices","Provides best practices and guidance for working with Rush monorepos. Use when the user is working in a Rush-based repository, asks about Rush commands (install, update, build, rebuild), needs help with project selection, dependency management, build caching, subspace configuration, or troubleshooting Rush-specific issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[786,789,792,793,796],{"name":787,"slug":788,"type":15},"Engineering","engineering",{"name":790,"slug":791,"type":15},"Local Development","local-development",{"name":9,"slug":8,"type":15},{"name":794,"slug":795,"type":15},"Project Management","project-management",{"name":797,"slug":798,"type":15},"Rush","rush",6484,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Frushstack","2026-04-06T18:34:44.965032",{"slug":803,"name":803,"fn":804,"description":805,"org":806,"tags":807,"stars":818,"repoUrl":819,"updatedAt":820},"azure-ai-agents-persistent-dotnet","build AI agents with Azure .NET SDK","Azure AI Agents Persistent SDK for .NET. Low-level SDK for creating and managing AI agents with threads, messages, runs, and tools. Use for agent CRUD, conversation threads, streaming responses, function calling, file search, and code interpreter. Triggers: \"PersistentAgentsClient\", \"persistent agents\", \"agent threads\", \"agent runs\", \"streaming agents\", \"function calling agents .NET\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[808,811,814,817],{"name":809,"slug":810,"type":15},".NET","net",{"name":812,"slug":813,"type":15},"Agents","agents",{"name":815,"slug":816,"type":15},"Azure","azure",{"name":747,"slug":748,"type":15},2804,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fskills","2026-07-03T16:32:10.297433",{"slug":822,"name":822,"fn":823,"description":824,"org":825,"tags":826,"stars":818,"repoUrl":819,"updatedAt":841},"azure-ai-anomalydetector-java","build anomaly detection applications with Java","Build anomaly detection applications with Azure AI Anomaly Detector SDK for Java. Use when implementing univariate\u002Fmultivariate anomaly detection, time-series analysis, or AI-powered monitoring.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[827,830,831,834,837,838],{"name":828,"slug":829,"type":15},"Analytics","analytics",{"name":815,"slug":816,"type":15},{"name":832,"slug":833,"type":15},"Data Analysis","data-analysis",{"name":835,"slug":836,"type":15},"Java","java",{"name":9,"slug":8,"type":15},{"name":839,"slug":840,"type":15},"Monitoring","monitoring","2026-05-13T06:14:16.261754",{"slug":843,"name":843,"fn":844,"description":845,"org":846,"tags":847,"stars":818,"repoUrl":819,"updatedAt":856},"azure-ai-contentsafety-java","build content moderation applications with Azure AI","Build content moderation applications with Azure AI Content Safety SDK for Java. Use when implementing text\u002Fimage analysis, blocklist management, or harm detection for hate, violence, sexual content, and self-harm.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[848,851,852,853],{"name":849,"slug":850,"type":15},"AI Infrastructure","ai-infrastructure",{"name":815,"slug":816,"type":15},{"name":835,"slug":836,"type":15},{"name":854,"slug":855,"type":15},"Security","security","2026-07-07T06:53:31.293235",{"slug":858,"name":858,"fn":859,"description":860,"org":861,"tags":862,"stars":818,"repoUrl":819,"updatedAt":873},"azure-ai-contentsafety-py","detect harmful content with Azure AI Content Safety","Azure AI Content Safety SDK for Python. Use for detecting harmful content in text and images with multi-severity classification.\nTriggers: \"azure-ai-contentsafety\", \"ContentSafetyClient\", \"content moderation\", \"harmful content\", \"text analysis\", \"image analysis\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[863,864,867,868,869,872],{"name":815,"slug":816,"type":15},{"name":865,"slug":866,"type":15},"Compliance","compliance",{"name":747,"slug":748,"type":15},{"name":9,"slug":8,"type":15},{"name":870,"slug":871,"type":15},"Python","python",{"name":854,"slug":855,"type":15},"2026-07-18T05:14:23.017504",{"slug":875,"name":875,"fn":876,"description":877,"org":878,"tags":879,"stars":818,"repoUrl":819,"updatedAt":884},"azure-ai-language-conversations-py","implement conversational language understanding with Python","Implement Conversational Language Understanding (CLU) using the azure-ai-language-conversations Python SDK. Use when working with ConversationAnalysisClient to analyze conversation intent and entities, building NLP features, or integrating language understanding into applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[880,881,882,883],{"name":828,"slug":829,"type":15},{"name":815,"slug":816,"type":15},{"name":747,"slug":748,"type":15},{"name":870,"slug":871,"type":15},"2026-07-31T05:54:29.068751",{"slug":886,"name":886,"fn":887,"description":888,"org":889,"tags":890,"stars":818,"repoUrl":819,"updatedAt":897},"azure-ai-translation-text-py","translate text using Azure AI services","Azure AI Text Translation SDK for real-time text translation, transliteration, language detection, and dictionary lookup. Use for translating text content in applications.\nTriggers: \"text translation\", \"translator\", \"translate text\", \"transliterate\", \"TextTranslationClient\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[891,894,895,896],{"name":892,"slug":893,"type":15},"API Development","api-development",{"name":815,"slug":816,"type":15},{"name":9,"slug":8,"type":15},{"name":870,"slug":871,"type":15},"2026-07-18T05:14:16.988376",{"slug":899,"name":899,"fn":900,"description":901,"org":902,"tags":903,"stars":818,"repoUrl":819,"updatedAt":912},"azure-ai-vision-imageanalysis-py","analyze images with Azure AI Vision","Azure AI Vision Image Analysis SDK for captions, tags, objects, OCR, people detection, and smart cropping. Use for computer vision and image understanding tasks.\nTriggers: \"image analysis\", \"computer vision\", \"OCR\", \"object detection\", \"ImageAnalysisClient\", \"image caption\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[904,905,908,911],{"name":815,"slug":816,"type":15},{"name":906,"slug":907,"type":15},"Computer Vision","computer-vision",{"name":909,"slug":910,"type":15},"Images","images",{"name":870,"slug":871,"type":15},"2026-07-18T05:14:18.007737",{"slug":914,"name":914,"fn":915,"description":916,"org":917,"tags":918,"stars":818,"repoUrl":819,"updatedAt":927},"azure-appconfiguration-java","manage configuration with Azure App Configuration","Azure App Configuration SDK for Java. Centralized application configuration management with key-value settings, feature flags, and snapshots.\nTriggers: \"ConfigurationClient java\", \"app configuration java\", \"feature flag java\", \"configuration setting java\", \"azure config java\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[919,920,923,926],{"name":815,"slug":816,"type":15},{"name":921,"slug":922,"type":15},"Configuration","configuration",{"name":924,"slug":925,"type":15},"Feature Flags","feature-flags",{"name":835,"slug":836,"type":15},"2026-07-03T16:32:01.278468",{"slug":929,"name":929,"fn":930,"description":931,"org":932,"tags":933,"stars":818,"repoUrl":819,"updatedAt":946},"azure-cosmos-rust","build applications with Azure Cosmos DB","Azure Cosmos DB library for Rust (NoSQL API). Document CRUD, containers, and globally distributed data.\nTriggers: \"cosmos db rust\", \"CosmosClient rust\", \"document crud rust\", \"NoSQL rust\", \"partition key rust\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[934,937,940,943],{"name":935,"slug":936,"type":15},"Cosmos DB","cosmos-db",{"name":938,"slug":939,"type":15},"Database","database",{"name":941,"slug":942,"type":15},"NoSQL","nosql",{"name":944,"slug":945,"type":15},"Rust","rust","2026-07-31T05:54:27.021432",{"slug":948,"name":948,"fn":930,"description":949,"org":950,"tags":951,"stars":818,"repoUrl":819,"updatedAt":959},"azure-cosmos-ts","Azure Cosmos DB JavaScript\u002FTypeScript SDK (@azure\u002Fcosmos) for data plane operations. Use for CRUD operations on documents, queries, bulk operations, and container management. Triggers: \"Cosmos DB\", \"@azure\u002Fcosmos\", \"CosmosClient\", \"document CRUD\", \"NoSQL queries\", \"bulk operations\", \"partition key\", \"container.items\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[952,953,954,955,956],{"name":935,"slug":936,"type":15},{"name":938,"slug":939,"type":15},{"name":9,"slug":8,"type":15},{"name":941,"slug":942,"type":15},{"name":957,"slug":958,"type":15},"TypeScript","typescript","2026-07-03T16:31:19.368382",{"slug":961,"name":961,"fn":962,"description":963,"org":964,"tags":965,"stars":818,"repoUrl":819,"updatedAt":971},"azure-data-tables-java","build table storage applications with Java","Build table storage applications with Azure Tables SDK for Java. Use when working with Azure Table Storage or Cosmos DB Table API for NoSQL key-value data, schemaless storage, or structured data at scale.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[966,967,968,969,970],{"name":815,"slug":816,"type":15},{"name":935,"slug":936,"type":15},{"name":938,"slug":939,"type":15},{"name":835,"slug":836,"type":15},{"name":941,"slug":942,"type":15},"2026-05-13T06:14:17.582229",267]