[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6DJF68YHCTQA473PSH24HSK":3,"skill-same:01M6DJF68YHCTQA473PSH24HSK":97},{"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":31,"publishedAt":52,"linkOnly":53,"markdownContent":54,"markdownBody":55,"frontMatter":56,"refName":57,"commitSha":58,"skillDirectory":59,"forksCount":14,"repositoryHtmlUrl":60,"repositoryHomepageUrl":6,"ownerHtmlUrl":61,"ownerBlogUrl":62,"ownerDescription":63,"updatedAt":64,"install":65,"resources":78},"01M6DJF68YHCTQA473PSH24HSK","scientific-platform-architect",null,"Design scientific computing platforms combining HPC kernels with modern infrastructure. Enforces compute-layer separation (Rust\u002FPyTorch\u002FPython), emulator-first thinking, and right-sized infrastructure. Use for MCMC pipelines, neural emulators, platform design, or scientific code review.","dasobral\u002Fskills","dasobral","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F108470360?v=4","01M6DJBRR588S4TSNFT0ZYY6BV","Daniel SB","01M6DJBRWBDEBPCQ874SWDZWZR",0,"MIT","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fscientific-platform-architect\u002FSKILL.md",[18,23,26],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000015","视觉设计",{"categoryNo":27,"categoryName":28,"taxonomyType":29,"taxonomyTypeDesc":30},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51],"scientific platform architect","platform","architect","platform design","scientific computing","compute","thinking","Neural","pipelines","emulators","use","code review","first","code","infrastructure","pytorch","rust","python","review","design","2026-11-10 10:12:30",false,"---\nname: scientific-platform-architect\ndescription: >\n  Design scientific computing platforms combining HPC kernels with modern\n  infrastructure. Enforces compute-layer separation (Rust\u002FPyTorch\u002FPython),\n  emulator-first thinking, and right-sized infrastructure. Use for MCMC\n  pipelines, neural emulators, platform design, or scientific code review.\n---\n\n# Scientific Platform Architect\n\nSenior scientific computing architect. Apply compute-layer principles and\ninfrastructure constraints. Push back on over-engineered infra for small teams.\n\n## Step 1 — Classify Request\n\n| Category | Examples |\n|----------|----------|\n| Platform design | Bayesian inference platform structure |\n| Compute layer | Rust vs Python vs JAX\u002FPyTorch |\n| Infrastructure sizing | K8s vs Docker Compose vs bare metal |\n| Emulator strategy | Slow likelihood → scale MCMC |\n| Agent\u002FMCP integration | Wire agents to kernels correctly |\n| Code review | Rust kernels, PyO3, PyTorch loops |\n\n## Step 2 — Four Diagnostic Questions (design\u002Fsizing)\n\n1. How many concurrent users\u002Finstitutions?\n2. Bottleneck: likelihood eval or sampling?\n3. Data sharing requirements?\n4. Team size and ops capacity?\n\n## Step 3 — Layer Separation\n\n| Layer | Technology | Responsibility |\n|-------|-----------|----------------|\n| Compute kernels | Rust (CPU) \u002F CUDA (GPU) | Numerics only — no I\u002FO, no agents |\n| ML\u002Femulators | PyTorch\u002FJAX | Training, inference, uncertainty |\n| Orchestration | Python | Pipelines, config, scheduling |\n| Agents | MCP\u002FAPI boundary | Orchestrate — never do numerics |\n\n**Rule:** Agents orchestrate; kernels compute. Never let an LLM agent call\na likelihood function directly.\n\n## Step 4 — Infrastructure Sizing\n\n| Scale | Recommendation |\n|-------|---------------|\n| Solo \u002F 1–3 researchers | Docker Compose or bare metal |\n| Small team (4–10) | Managed services, minimal K8s |\n| Multi-institution | Kubernetes justified |\n\nSee `references\u002Finfrastructure-decision-guide.md`.\n\n## Step 5 — Emulator-First\n\nIf likelihood eval >100ms per call, emulator strategy before kernel optimization.\nSee `references\u002Femulator-benchmarks.md` and `templates\u002Fplatform-design.md`.\n\n## Rules\n\n- Operational complexity is a real cost\n- Right-size infrastructure to team scale\n- Examples in `examples\u002Fbayesnetes-design.md`\n","# Scientific Platform Architect\n\nSenior scientific computing architect. Apply compute-layer principles and\ninfrastructure constraints. Push back on over-engineered infra for small teams.\n\n## Step 1 — Classify Request\n\n| Category | Examples |\n|----------|----------|\n| Platform design | Bayesian inference platform structure |\n| Compute layer | Rust vs Python vs JAX\u002FPyTorch |\n| Infrastructure sizing | K8s vs Docker Compose vs bare metal |\n| Emulator strategy | Slow likelihood → scale MCMC |\n| Agent\u002FMCP integration | Wire agents to kernels correctly |\n| Code review | Rust kernels, PyO3, PyTorch loops |\n\n## Step 2 — Four Diagnostic Questions (design\u002Fsizing)\n\n1. How many concurrent users\u002Finstitutions?\n2. Bottleneck: likelihood eval or sampling?\n3. Data sharing requirements?\n4. Team size and ops capacity?\n\n## Step 3 — Layer Separation\n\n| Layer | Technology | Responsibility |\n|-------|-----------|----------------|\n| Compute kernels | Rust (CPU) \u002F CUDA (GPU) | Numerics only — no I\u002FO, no agents |\n| ML\u002Femulators | PyTorch\u002FJAX | Training, inference, uncertainty |\n| Orchestration | Python | Pipelines, config, scheduling |\n| Agents | MCP\u002FAPI boundary | Orchestrate — never do numerics |\n\n**Rule:** Agents orchestrate; kernels compute. Never let an LLM agent call\na likelihood function directly.\n\n## Step 4 — Infrastructure Sizing\n\n| Scale | Recommendation |\n|-------|---------------|\n| Solo \u002F 1–3 researchers | Docker Compose or bare metal |\n| Small team (4–10) | Managed services, minimal K8s |\n| Multi-institution | Kubernetes justified |\n\nSee `references\u002Finfrastructure-decision-guide.md`.\n\n## Step 5 — Emulator-First\n\nIf likelihood eval >100ms per call, emulator strategy before kernel optimization.\nSee `references\u002Femulator-benchmarks.md` and `templates\u002Fplatform-design.md`.\n\n## Rules\n\n- Operational complexity is a real cost\n- Right-size infrastructure to team scale\n- Examples in `examples\u002Fbayesnetes-design.md`\n",{"name":5,"description":7},"main","1e047a4b00be0c07efa464cd97f36fa39d4b1718","skills\u002Fscientific-platform-architect","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills","https:\u002F\u002Fgithub.com\u002Fdasobral","https:\u002F\u002Forcid.org\u002Fmy-orcid?orcid=0000-0001-9559-3651","Software Engineer with a Ph. D. in Physics","2026-08-15 21:49:02",{"source":8,"skillName":5,"commands":66,"prompt":76,"zipAvailable":77},[67,70,73],{"tool":68,"command":69},"npx","npx skills add dasobral\u002Fskills --skill scientific-platform-architect",{"tool":71,"command":72},"bunx","bunx skills add dasobral\u002Fskills --skill scientific-platform-architect",{"tool":74,"command":75},"pnpm","pnpm dlx skills add dasobral\u002Fskills --skill scientific-platform-architect","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fscientific-platform-architect\u002FSKILL.md",true,[79,85,89,93],{"path":80,"resourceType":81,"resourceTypeDesc":82,"sizeBytes":83,"sourceHtmlUrl":84},"examples\u002Fbayesnetes-design.md","REFERENCE","参考文档",6286,"https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fscientific-platform-architect\u002Fexamples\u002Fbayesnetes-design.md",{"path":86,"resourceType":81,"resourceTypeDesc":82,"sizeBytes":87,"sourceHtmlUrl":88},"references\u002Femulator-benchmarks.md",3962,"https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fscientific-platform-architect\u002Freferences\u002Femulator-benchmarks.md",{"path":90,"resourceType":81,"resourceTypeDesc":82,"sizeBytes":91,"sourceHtmlUrl":92},"references\u002Finfrastructure-decision-guide.md",5080,"https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fscientific-platform-architect\u002Freferences\u002Finfrastructure-decision-guide.md",{"path":94,"resourceType":81,"resourceTypeDesc":82,"sizeBytes":95,"sourceHtmlUrl":96},"templates\u002Fplatform-design.md",4203,"https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fscientific-platform-architect\u002Ftemplates\u002Fplatform-design.md",[98,113,143,163,187,202],{"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":52},"01M6DJFKJASZX8BMFGRQBEKFY9","fixture-skill","Fixture skill used to prove trust inventory coverage.","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Ftools\u002Fskills-export\u002Ftests\u002Ffixtures\u002Fworkflows\u002Fagent-security\u002Frepository\u002F.cursor\u002Fskills\u002Ffixture-skill\u002FSKILL.md",[104],{"categoryNo":105,"categoryName":106,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",[108,109,110,111,112],"fixture skill","fixture","inventory","coverage","trust",{"skillNo":114,"name":115,"nameZh":6,"description":116,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":117,"categories":118,"keywords":128,"publishedAt":52},"01M6DJFG24D9X2WAP411JJP71W","write-conformant-code","Write code in any language that matches project conventions exactly. Reads CODING_REQUIREMENTS.md, prior review output, or built-in defaults before writing. Part of Codecraft pipeline: analyze → write → review → audit. Use when implementing features, functions, or any code that should fit seamlessly into an existing codebase.","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwrite-conformant-code\u002FSKILL.md",[119,122,125],{"categoryNo":120,"categoryName":121,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":123,"categoryName":124,"taxonomyType":29,"taxonomyTypeDesc":30},"01JBSEEDCAT000000000000030","开发工程师",{"categoryNo":126,"categoryName":127,"taxonomyType":29,"taxonomyTypeDesc":30},"01JBSEEDCAT000000000000036","内容创作者",[45,129,130,131,132,50,133,134,135,136,42,137,138,139,140,141,142],"write","write conformant code","analyze","writing","project conventions","audit","pipeline","coding","functions","language","conventions","project","requirements","fit",{"skillNo":144,"name":145,"nameZh":6,"description":146,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":147,"categories":148,"keywords":152,"publishedAt":52},"01M6DJFDDCWB5VMAYE79YTZB40","test-crypto-interoperability","Record reproducible cryptographic interoperability cases, including negotiation, fallback, sizes, latency, fragmentation, and negative tests.","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Ftest-crypto-interoperability\u002FSKILL.md",[149],{"categoryNo":150,"categoryName":151,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",[153,154,155,156,157,158,159,160,161,162],"interoperability","test","crypto","test crypto interoperability","Reproducible","fallback","record","latency","tests","negotiation",{"skillNo":164,"name":165,"nameZh":6,"description":166,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":167,"categories":168,"keywords":170,"publishedAt":52},"01M6DJF2D8ZNG4439WQ21CQ75D","rust-systems","Rust systems engineering for daemons, HTTP APIs, and AOS components. Enforces Tokio\u002FAxum async-first, LiteLLM-compatible OpenAI APIs, workspace layouts, tracing, and systemd-compatible daemon design. Part of AOS Stack plugin. Use for .rs files, Cargo.toml, daemons, or Rust orchestrator services.","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Frust-systems\u002FSKILL.md",[169],{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},[48,171,172,173,51,174,175,176,177,178,44,179,180,181,182,42,183,184,185,186],"systems","rust systems","openai","components","http","plugin","tracing","services","stack","workspace","litellm","orchestrator","async","files","toml","cargo",{"skillNo":188,"name":189,"nameZh":6,"description":190,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":191,"categories":192,"keywords":194,"publishedAt":52},"01M6DJEYP0JJX0XJPJYR36VRTW","run-agent-attack-replay","Aggregate isolated agent attack trials with deterministic side-effect assertions, Wilson intervals, and preserved trace evidence.","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Frun-agent-attack-replay\u002FSKILL.md",[193],{"categoryNo":105,"categoryName":106,"taxonomyType":21,"taxonomyTypeDesc":22},[195,196,197,198,199,200,201],"run","replay","run agent attack replay","trace","effect","evidence","deterministic",{"skillNo":203,"name":204,"nameZh":6,"description":205,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":206,"categories":207,"keywords":209,"publishedAt":52},"01M6DJERMEJ072CKPA19PDDH0P","review-quality","Language-agnostic code review across four pillars: Clarity, Simplicity, Good Practices, and Security. Produces severity-tagged findings (CRITICAL\u002FWARNING\u002FSUGGESTION). Complements audit-cognitive-debt for dead code and duplication. Part of Codecraft plugin. Use when asked to review code quality, audit security, check best practices, or find code smells.","https:\u002F\u002Fgithub.com\u002Fdasobral\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Freview-quality\u002FSKILL.md",[208],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[210,50,211,212,213,134,214,45,176,215,43,216,217,218,42,219,220,221,222,223],"review quality","quality","review code","severity","best practices","code quality","critical","good","simplicity","check","find","across","dead code","clarity"]