[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-optimization-mode-router":3,"mdc--259g83-key":40,"related-repo-nvidia-optimization-mode-router":829,"related-org-nvidia-optimization-mode-router":919},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":20,"repoUrl":21,"updatedAt":22,"license":23,"forks":24,"topics":25,"repo":35,"sourceUrl":38,"mdContent":39},"optimization-mode-router","route optimization requests to cuOpt solvers","Choose fast direct-to-cuOpt solve versus replayable or auditable model artifact mode.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19],{"name":13,"slug":14,"type":15},"Automation","automation","tag",{"name":17,"slug":18,"type":15},"Optimization","optimization",{"name":9,"slug":8,"type":15},462,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcuopt-examples","2026-07-30T05:29:18.162351","Apache-2.0",81,[26,27,28,29,30,18,31,32,33,34],"combinatorial-optimization","gpu","linear-programming","mixed-integer-programming","operations-research","optimization-algorithms","route-optimization","traveling-salesman-problem","vehicle-routing-problem",{"repoUrl":21,"stars":20,"forks":24,"topics":36,"description":37},[26,27,28,29,30,18,31,32,33,34],"NVIDIA cuOpt examples for decision optimization","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcuopt-examples\u002Ftree\u002FHEAD\u002Fcuopt_on_nemoclaw\u002Fopenclaw-skills\u002Foptimization-mode-router","---\nname: optimization-mode-router\nversion: \"26.06.01\"\ndescription: Choose fast direct-to-cuOpt solve versus replayable or auditable model artifact mode.\nlicense: Apache-2.0\nmetadata:\n  author: NVIDIA cuOpt Team\n  tags:\n    - cuopt\n    - nemoclaw\n    - orchestration\norigin: skill-evolution\n---\n\n# Optimization Mode Router\n\nUse this skill when a user asks a question that may be answered by solving an optimization problem from uploaded or provided data, and you need to decide whether to:\n\n- proceed with a **fast direct-to-cuOpt solve**, or\n- offer a **replayable\u002Fauditable path** that preserves a structured model artifact for later reruns, review, export, or audit.\n\nThis skill is about **mode selection**, not full formulation. Its purpose is to keep the common path fast while surfacing stronger reproducibility only when it is actually useful.\n\n## Read this when\n\nRead this skill when all of the following are true:\n\n1. The user appears to be asking a question that could become an LP, MILP, QP, or routing problem.\n2. The user has provided data, or is expected to provide data.\n3. You need to decide whether to:\n   - go straight to a direct cuOpt solve, or\n   - preserve a replayable\u002Fauditable artifact as part of the workflow.\n\n## Do not use this skill for\n\n- Pure formulation work after the execution mode has already been chosen.\n- Pure cuOpt API usage when the user has already clearly chosen fast vs replayable mode.\n- Non-optimization analytics questions.\n\n## Default behavior\n\n- Default to **Fast mode**.\n- Default to **direct cuOpt solve** for one-off requests from uploaded CSVs\n  (schedule, assignment, allocation, routing) — proceed to\n  `cuopt-model-mapper` without asking fast vs replayable unless the user\n  signals audit\u002Fexport\u002Frerun.\n- Do **not** ask about replayability\u002Fauditability unless there is a real signal that it matters.\n- Avoid turning a straightforward optimization request into a heavy upfront questionnaire.\n- **NemoClaw sandbox:** Fast mode means cuOpt after `cuopt-sandbox` gates —\n  never a custom greedy\u002Fheuristic builder as the solve path.\n\n## Two modes\n\n### Fast mode\nUse Fast mode when the user appears to want the quickest route to an answer.\n\nBehavior:\n- inspect the provided data\n- identify whether the request is LP, MILP, QP, or routing\n- ask only the minimum necessary clarifying questions\n- build the model directly in cuOpt\n- solve and explain the result\n- do not preserve a replayable model artifact unless requested\n\n### Replayable \u002F Auditable mode\nUse this mode when the user wants reuse, traceability, or formal review.\n\nBehavior:\n- capture explicit assumptions\n- record data-to-model mappings\n- preserve a structured model specification or equivalent reusable artifact\n- solve with cuOpt\n- return both the answer and reusable\u002Freviewable model metadata\n\n## When to ask the mode-selection question\n\nAsk the user to choose between Fast mode and Replayable\u002FAuditable mode if any of the following are true:\n\n1. The user explicitly asks for any of these:\n   - save the model\n   - rerun on new data\n   - replay later\n   - recurring or scheduled runs\n   - audit trail\n   - export the model\n   - document assumptions\n   - review the formulation\n   - show how the data maps into the model\n\n2. The request appears operational or recurring rather than one-off:\n   - daily \u002F weekly \u002F monthly planning\n   - production workflow\n   - repeated scenario analysis\n   - future datasets are expected\n\n3. The user indicates a need for traceability or justification:\n   - compliance\n   - internal or external review\n   - reproducibility\n   - explicit explanation of assumptions or mappings\n\n## When not to ask\n\nDo **not** ask the mode-selection question when the request is clearly:\n- one-off\n- exploratory\n- ad hoc\n- speed-oriented\n\nIn those cases, proceed in **Fast mode** unless the user later asks for replayability, audit, export, or model persistence.\n\n## Recommended user-facing phrasing\n\nPreferred short form:\n\n**Should I treat this as a one-off solve, or make it replayable\u002Fauditable too?**\n\nAlternative longer form:\n\n**I can do this in two ways: Fast mode for the quickest answer, or Replayable mode that also keeps a structured spec for reruns, audit, and reuse. Which do you want?**\n\n## Decision rule\n\n- If there is **no meaningful signal** for replayability\u002Fauditability → use **Fast mode** silently.\n- If there **is** a meaningful signal → ask the mode-selection clarifier early.\n- If the user chooses replayable\u002Fauditable mode → preserve the structured artifact and proceed.\n- If the user chooses fast mode → proceed directly to cuOpt.\n\n## Handoff guidance\n\nAfter selecting a mode, hand off based on problem type:\n\n- If the request is LP \u002F MILP:\n  - use `numerical-optimization-formulation`\n  - then use `cuopt-numerical-optimization-api-python` (or\n    `cuopt-numerical-optimization-api-cli` for MPS inputs)\n  - in sandbox contexts, follow `cuopt-sandbox` (gates + selected gRPC path)\n    before any LP\u002FMILP solve\n\n- If the request is QP:\n  - use `numerical-optimization-formulation`\n  - then use `cuopt-numerical-optimization-api-python`\n  - in sandbox contexts, follow `cuopt-sandbox` (gates + selected gRPC path)\n    before any QP solve\n\n- If the request is routing (VRP \u002F TSP \u002F PDP):\n  - use `routing-formulation`\n  - then use `cuopt-routing-api-python`\n  - in sandbox contexts, follow `cuopt-sandbox` (gates + REST)\n    before any routing solve\n\n- If the user is asking about server usage or deployment rather than solving a model directly:\n  - use `cuopt-server-common` or `cuopt-server-api-python` as appropriate\n\n- In all cuOpt user tasks:\n  - follow `cuopt-user-rules`\n\n## Operational notes\n\n- This skill is about **execution mode selection**, not full mathematical modeling.\n- Keep the user experience lightweight by default.\n- Prefer direct-to-cuOpt for one-off work.\n- Use replayable\u002Fauditable mode only when the user’s needs justify the extra structure.\n",{"data":41,"body":50},{"name":4,"version":42,"description":6,"license":23,"metadata":43,"origin":49},"26.06.01",{"author":44,"tags":45},"NVIDIA cuOpt Team",[46,47,48],"cuopt","nemoclaw","orchestration","skill-evolution",{"type":51,"children":52},"root",[53,61,67,97,109,116,121,153,159,177,183,253,259,265,270,275,308,314,319,323,351,357,362,474,480,490,513,524,530,535,543,548,556,562,605,611,616,794,800],{"type":54,"tag":55,"props":56,"children":57},"element","h1",{"id":4},[58],{"type":59,"value":60},"text","Optimization Mode Router",{"type":54,"tag":62,"props":63,"children":64},"p",{},[65],{"type":59,"value":66},"Use this skill when a user asks a question that may be answered by solving an optimization problem from uploaded or provided data, and you need to decide whether to:",{"type":54,"tag":68,"props":69,"children":70},"ul",{},[71,85],{"type":54,"tag":72,"props":73,"children":74},"li",{},[75,77,83],{"type":59,"value":76},"proceed with a ",{"type":54,"tag":78,"props":79,"children":80},"strong",{},[81],{"type":59,"value":82},"fast direct-to-cuOpt solve",{"type":59,"value":84},", or",{"type":54,"tag":72,"props":86,"children":87},{},[88,90,95],{"type":59,"value":89},"offer a ",{"type":54,"tag":78,"props":91,"children":92},{},[93],{"type":59,"value":94},"replayable\u002Fauditable path",{"type":59,"value":96}," that preserves a structured model artifact for later reruns, review, export, or audit.",{"type":54,"tag":62,"props":98,"children":99},{},[100,102,107],{"type":59,"value":101},"This skill is about ",{"type":54,"tag":78,"props":103,"children":104},{},[105],{"type":59,"value":106},"mode selection",{"type":59,"value":108},", not full formulation. Its purpose is to keep the common path fast while surfacing stronger reproducibility only when it is actually useful.",{"type":54,"tag":110,"props":111,"children":113},"h2",{"id":112},"read-this-when",[114],{"type":59,"value":115},"Read this when",{"type":54,"tag":62,"props":117,"children":118},{},[119],{"type":59,"value":120},"Read this skill when all of the following are true:",{"type":54,"tag":122,"props":123,"children":124},"ol",{},[125,130,135],{"type":54,"tag":72,"props":126,"children":127},{},[128],{"type":59,"value":129},"The user appears to be asking a question that could become an LP, MILP, QP, or routing problem.",{"type":54,"tag":72,"props":131,"children":132},{},[133],{"type":59,"value":134},"The user has provided data, or is expected to provide data.",{"type":54,"tag":72,"props":136,"children":137},{},[138,140],{"type":59,"value":139},"You need to decide whether to:\n",{"type":54,"tag":68,"props":141,"children":142},{},[143,148],{"type":54,"tag":72,"props":144,"children":145},{},[146],{"type":59,"value":147},"go straight to a direct cuOpt solve, or",{"type":54,"tag":72,"props":149,"children":150},{},[151],{"type":59,"value":152},"preserve a replayable\u002Fauditable artifact as part of the workflow.",{"type":54,"tag":110,"props":154,"children":156},{"id":155},"do-not-use-this-skill-for",[157],{"type":59,"value":158},"Do not use this skill for",{"type":54,"tag":68,"props":160,"children":161},{},[162,167,172],{"type":54,"tag":72,"props":163,"children":164},{},[165],{"type":59,"value":166},"Pure formulation work after the execution mode has already been chosen.",{"type":54,"tag":72,"props":168,"children":169},{},[170],{"type":59,"value":171},"Pure cuOpt API usage when the user has already clearly chosen fast vs replayable mode.",{"type":54,"tag":72,"props":173,"children":174},{},[175],{"type":59,"value":176},"Non-optimization analytics questions.",{"type":54,"tag":110,"props":178,"children":180},{"id":179},"default-behavior",[181],{"type":59,"value":182},"Default behavior",{"type":54,"tag":68,"props":184,"children":185},{},[186,198,218,230,235],{"type":54,"tag":72,"props":187,"children":188},{},[189,191,196],{"type":59,"value":190},"Default to ",{"type":54,"tag":78,"props":192,"children":193},{},[194],{"type":59,"value":195},"Fast mode",{"type":59,"value":197},".",{"type":54,"tag":72,"props":199,"children":200},{},[201,202,207,209,216],{"type":59,"value":190},{"type":54,"tag":78,"props":203,"children":204},{},[205],{"type":59,"value":206},"direct cuOpt solve",{"type":59,"value":208}," for one-off requests from uploaded CSVs\n(schedule, assignment, allocation, routing) — proceed to\n",{"type":54,"tag":210,"props":211,"children":213},"code",{"className":212},[],[214],{"type":59,"value":215},"cuopt-model-mapper",{"type":59,"value":217}," without asking fast vs replayable unless the user\nsignals audit\u002Fexport\u002Frerun.",{"type":54,"tag":72,"props":219,"children":220},{},[221,223,228],{"type":59,"value":222},"Do ",{"type":54,"tag":78,"props":224,"children":225},{},[226],{"type":59,"value":227},"not",{"type":59,"value":229}," ask about replayability\u002Fauditability unless there is a real signal that it matters.",{"type":54,"tag":72,"props":231,"children":232},{},[233],{"type":59,"value":234},"Avoid turning a straightforward optimization request into a heavy upfront questionnaire.",{"type":54,"tag":72,"props":236,"children":237},{},[238,243,245,251],{"type":54,"tag":78,"props":239,"children":240},{},[241],{"type":59,"value":242},"NemoClaw sandbox:",{"type":59,"value":244}," Fast mode means cuOpt after ",{"type":54,"tag":210,"props":246,"children":248},{"className":247},[],[249],{"type":59,"value":250},"cuopt-sandbox",{"type":59,"value":252}," gates —\nnever a custom greedy\u002Fheuristic builder as the solve path.",{"type":54,"tag":110,"props":254,"children":256},{"id":255},"two-modes",[257],{"type":59,"value":258},"Two modes",{"type":54,"tag":260,"props":261,"children":263},"h3",{"id":262},"fast-mode",[264],{"type":59,"value":195},{"type":54,"tag":62,"props":266,"children":267},{},[268],{"type":59,"value":269},"Use Fast mode when the user appears to want the quickest route to an answer.",{"type":54,"tag":62,"props":271,"children":272},{},[273],{"type":59,"value":274},"Behavior:",{"type":54,"tag":68,"props":276,"children":277},{},[278,283,288,293,298,303],{"type":54,"tag":72,"props":279,"children":280},{},[281],{"type":59,"value":282},"inspect the provided data",{"type":54,"tag":72,"props":284,"children":285},{},[286],{"type":59,"value":287},"identify whether the request is LP, MILP, QP, or routing",{"type":54,"tag":72,"props":289,"children":290},{},[291],{"type":59,"value":292},"ask only the minimum necessary clarifying questions",{"type":54,"tag":72,"props":294,"children":295},{},[296],{"type":59,"value":297},"build the model directly in cuOpt",{"type":54,"tag":72,"props":299,"children":300},{},[301],{"type":59,"value":302},"solve and explain the result",{"type":54,"tag":72,"props":304,"children":305},{},[306],{"type":59,"value":307},"do not preserve a replayable model artifact unless requested",{"type":54,"tag":260,"props":309,"children":311},{"id":310},"replayable-auditable-mode",[312],{"type":59,"value":313},"Replayable \u002F Auditable mode",{"type":54,"tag":62,"props":315,"children":316},{},[317],{"type":59,"value":318},"Use this mode when the user wants reuse, traceability, or formal review.",{"type":54,"tag":62,"props":320,"children":321},{},[322],{"type":59,"value":274},{"type":54,"tag":68,"props":324,"children":325},{},[326,331,336,341,346],{"type":54,"tag":72,"props":327,"children":328},{},[329],{"type":59,"value":330},"capture explicit assumptions",{"type":54,"tag":72,"props":332,"children":333},{},[334],{"type":59,"value":335},"record data-to-model mappings",{"type":54,"tag":72,"props":337,"children":338},{},[339],{"type":59,"value":340},"preserve a structured model specification or equivalent reusable artifact",{"type":54,"tag":72,"props":342,"children":343},{},[344],{"type":59,"value":345},"solve with cuOpt",{"type":54,"tag":72,"props":347,"children":348},{},[349],{"type":59,"value":350},"return both the answer and reusable\u002Freviewable model metadata",{"type":54,"tag":110,"props":352,"children":354},{"id":353},"when-to-ask-the-mode-selection-question",[355],{"type":59,"value":356},"When to ask the mode-selection question",{"type":54,"tag":62,"props":358,"children":359},{},[360],{"type":59,"value":361},"Ask the user to choose between Fast mode and Replayable\u002FAuditable mode if any of the following are true:",{"type":54,"tag":122,"props":363,"children":364},{},[365,418,446],{"type":54,"tag":72,"props":366,"children":367},{},[368,370],{"type":59,"value":369},"The user explicitly asks for any of these:",{"type":54,"tag":68,"props":371,"children":372},{},[373,378,383,388,393,398,403,408,413],{"type":54,"tag":72,"props":374,"children":375},{},[376],{"type":59,"value":377},"save the model",{"type":54,"tag":72,"props":379,"children":380},{},[381],{"type":59,"value":382},"rerun on new data",{"type":54,"tag":72,"props":384,"children":385},{},[386],{"type":59,"value":387},"replay later",{"type":54,"tag":72,"props":389,"children":390},{},[391],{"type":59,"value":392},"recurring or scheduled runs",{"type":54,"tag":72,"props":394,"children":395},{},[396],{"type":59,"value":397},"audit trail",{"type":54,"tag":72,"props":399,"children":400},{},[401],{"type":59,"value":402},"export the model",{"type":54,"tag":72,"props":404,"children":405},{},[406],{"type":59,"value":407},"document assumptions",{"type":54,"tag":72,"props":409,"children":410},{},[411],{"type":59,"value":412},"review the formulation",{"type":54,"tag":72,"props":414,"children":415},{},[416],{"type":59,"value":417},"show how the data maps into the model",{"type":54,"tag":72,"props":419,"children":420},{},[421,423],{"type":59,"value":422},"The request appears operational or recurring rather than one-off:",{"type":54,"tag":68,"props":424,"children":425},{},[426,431,436,441],{"type":54,"tag":72,"props":427,"children":428},{},[429],{"type":59,"value":430},"daily \u002F weekly \u002F monthly planning",{"type":54,"tag":72,"props":432,"children":433},{},[434],{"type":59,"value":435},"production workflow",{"type":54,"tag":72,"props":437,"children":438},{},[439],{"type":59,"value":440},"repeated scenario analysis",{"type":54,"tag":72,"props":442,"children":443},{},[444],{"type":59,"value":445},"future datasets are expected",{"type":54,"tag":72,"props":447,"children":448},{},[449,451],{"type":59,"value":450},"The user indicates a need for traceability or justification:",{"type":54,"tag":68,"props":452,"children":453},{},[454,459,464,469],{"type":54,"tag":72,"props":455,"children":456},{},[457],{"type":59,"value":458},"compliance",{"type":54,"tag":72,"props":460,"children":461},{},[462],{"type":59,"value":463},"internal or external review",{"type":54,"tag":72,"props":465,"children":466},{},[467],{"type":59,"value":468},"reproducibility",{"type":54,"tag":72,"props":470,"children":471},{},[472],{"type":59,"value":473},"explicit explanation of assumptions or mappings",{"type":54,"tag":110,"props":475,"children":477},{"id":476},"when-not-to-ask",[478],{"type":59,"value":479},"When not to ask",{"type":54,"tag":62,"props":481,"children":482},{},[483,484,488],{"type":59,"value":222},{"type":54,"tag":78,"props":485,"children":486},{},[487],{"type":59,"value":227},{"type":59,"value":489}," ask the mode-selection question when the request is clearly:",{"type":54,"tag":68,"props":491,"children":492},{},[493,498,503,508],{"type":54,"tag":72,"props":494,"children":495},{},[496],{"type":59,"value":497},"one-off",{"type":54,"tag":72,"props":499,"children":500},{},[501],{"type":59,"value":502},"exploratory",{"type":54,"tag":72,"props":504,"children":505},{},[506],{"type":59,"value":507},"ad hoc",{"type":54,"tag":72,"props":509,"children":510},{},[511],{"type":59,"value":512},"speed-oriented",{"type":54,"tag":62,"props":514,"children":515},{},[516,518,522],{"type":59,"value":517},"In those cases, proceed in ",{"type":54,"tag":78,"props":519,"children":520},{},[521],{"type":59,"value":195},{"type":59,"value":523}," unless the user later asks for replayability, audit, export, or model persistence.",{"type":54,"tag":110,"props":525,"children":527},{"id":526},"recommended-user-facing-phrasing",[528],{"type":59,"value":529},"Recommended user-facing phrasing",{"type":54,"tag":62,"props":531,"children":532},{},[533],{"type":59,"value":534},"Preferred short form:",{"type":54,"tag":62,"props":536,"children":537},{},[538],{"type":54,"tag":78,"props":539,"children":540},{},[541],{"type":59,"value":542},"Should I treat this as a one-off solve, or make it replayable\u002Fauditable too?",{"type":54,"tag":62,"props":544,"children":545},{},[546],{"type":59,"value":547},"Alternative longer form:",{"type":54,"tag":62,"props":549,"children":550},{},[551],{"type":54,"tag":78,"props":552,"children":553},{},[554],{"type":59,"value":555},"I can do this in two ways: Fast mode for the quickest answer, or Replayable mode that also keeps a structured spec for reruns, audit, and reuse. Which do you want?",{"type":54,"tag":110,"props":557,"children":559},{"id":558},"decision-rule",[560],{"type":59,"value":561},"Decision rule",{"type":54,"tag":68,"props":563,"children":564},{},[565,583,595,600],{"type":54,"tag":72,"props":566,"children":567},{},[568,570,575,577,581],{"type":59,"value":569},"If there is ",{"type":54,"tag":78,"props":571,"children":572},{},[573],{"type":59,"value":574},"no meaningful signal",{"type":59,"value":576}," for replayability\u002Fauditability → use ",{"type":54,"tag":78,"props":578,"children":579},{},[580],{"type":59,"value":195},{"type":59,"value":582}," silently.",{"type":54,"tag":72,"props":584,"children":585},{},[586,588,593],{"type":59,"value":587},"If there ",{"type":54,"tag":78,"props":589,"children":590},{},[591],{"type":59,"value":592},"is",{"type":59,"value":594}," a meaningful signal → ask the mode-selection clarifier early.",{"type":54,"tag":72,"props":596,"children":597},{},[598],{"type":59,"value":599},"If the user chooses replayable\u002Fauditable mode → preserve the structured artifact and proceed.",{"type":54,"tag":72,"props":601,"children":602},{},[603],{"type":59,"value":604},"If the user chooses fast mode → proceed directly to cuOpt.",{"type":54,"tag":110,"props":606,"children":608},{"id":607},"handoff-guidance",[609],{"type":59,"value":610},"Handoff guidance",{"type":54,"tag":62,"props":612,"children":613},{},[614],{"type":59,"value":615},"After selecting a mode, hand off based on problem type:",{"type":54,"tag":68,"props":617,"children":618},{},[619,671,708,747,775],{"type":54,"tag":72,"props":620,"children":621},{},[622,624],{"type":59,"value":623},"If the request is LP \u002F MILP:",{"type":54,"tag":68,"props":625,"children":626},{},[627,638,659],{"type":54,"tag":72,"props":628,"children":629},{},[630,632],{"type":59,"value":631},"use ",{"type":54,"tag":210,"props":633,"children":635},{"className":634},[],[636],{"type":59,"value":637},"numerical-optimization-formulation",{"type":54,"tag":72,"props":639,"children":640},{},[641,643,649,651,657],{"type":59,"value":642},"then use ",{"type":54,"tag":210,"props":644,"children":646},{"className":645},[],[647],{"type":59,"value":648},"cuopt-numerical-optimization-api-python",{"type":59,"value":650}," (or\n",{"type":54,"tag":210,"props":652,"children":654},{"className":653},[],[655],{"type":59,"value":656},"cuopt-numerical-optimization-api-cli",{"type":59,"value":658}," for MPS inputs)",{"type":54,"tag":72,"props":660,"children":661},{},[662,664,669],{"type":59,"value":663},"in sandbox contexts, follow ",{"type":54,"tag":210,"props":665,"children":667},{"className":666},[],[668],{"type":59,"value":250},{"type":59,"value":670}," (gates + selected gRPC path)\nbefore any LP\u002FMILP solve",{"type":54,"tag":72,"props":672,"children":673},{},[674,676],{"type":59,"value":675},"If the request is QP:",{"type":54,"tag":68,"props":677,"children":678},{},[679,688,697],{"type":54,"tag":72,"props":680,"children":681},{},[682,683],{"type":59,"value":631},{"type":54,"tag":210,"props":684,"children":686},{"className":685},[],[687],{"type":59,"value":637},{"type":54,"tag":72,"props":689,"children":690},{},[691,692],{"type":59,"value":642},{"type":54,"tag":210,"props":693,"children":695},{"className":694},[],[696],{"type":59,"value":648},{"type":54,"tag":72,"props":698,"children":699},{},[700,701,706],{"type":59,"value":663},{"type":54,"tag":210,"props":702,"children":704},{"className":703},[],[705],{"type":59,"value":250},{"type":59,"value":707}," (gates + selected gRPC path)\nbefore any QP solve",{"type":54,"tag":72,"props":709,"children":710},{},[711,713],{"type":59,"value":712},"If the request is routing (VRP \u002F TSP \u002F PDP):",{"type":54,"tag":68,"props":714,"children":715},{},[716,726,736],{"type":54,"tag":72,"props":717,"children":718},{},[719,720],{"type":59,"value":631},{"type":54,"tag":210,"props":721,"children":723},{"className":722},[],[724],{"type":59,"value":725},"routing-formulation",{"type":54,"tag":72,"props":727,"children":728},{},[729,730],{"type":59,"value":642},{"type":54,"tag":210,"props":731,"children":733},{"className":732},[],[734],{"type":59,"value":735},"cuopt-routing-api-python",{"type":54,"tag":72,"props":737,"children":738},{},[739,740,745],{"type":59,"value":663},{"type":54,"tag":210,"props":741,"children":743},{"className":742},[],[744],{"type":59,"value":250},{"type":59,"value":746}," (gates + REST)\nbefore any routing solve",{"type":54,"tag":72,"props":748,"children":749},{},[750,752],{"type":59,"value":751},"If the user is asking about server usage or deployment rather than solving a model directly:",{"type":54,"tag":68,"props":753,"children":754},{},[755],{"type":54,"tag":72,"props":756,"children":757},{},[758,759,765,767,773],{"type":59,"value":631},{"type":54,"tag":210,"props":760,"children":762},{"className":761},[],[763],{"type":59,"value":764},"cuopt-server-common",{"type":59,"value":766}," or ",{"type":54,"tag":210,"props":768,"children":770},{"className":769},[],[771],{"type":59,"value":772},"cuopt-server-api-python",{"type":59,"value":774}," as appropriate",{"type":54,"tag":72,"props":776,"children":777},{},[778,780],{"type":59,"value":779},"In all cuOpt user tasks:",{"type":54,"tag":68,"props":781,"children":782},{},[783],{"type":54,"tag":72,"props":784,"children":785},{},[786,788],{"type":59,"value":787},"follow ",{"type":54,"tag":210,"props":789,"children":791},{"className":790},[],[792],{"type":59,"value":793},"cuopt-user-rules",{"type":54,"tag":110,"props":795,"children":797},{"id":796},"operational-notes",[798],{"type":59,"value":799},"Operational notes",{"type":54,"tag":68,"props":801,"children":802},{},[803,814,819,824],{"type":54,"tag":72,"props":804,"children":805},{},[806,807,812],{"type":59,"value":101},{"type":54,"tag":78,"props":808,"children":809},{},[810],{"type":59,"value":811},"execution mode selection",{"type":59,"value":813},", not full mathematical modeling.",{"type":54,"tag":72,"props":815,"children":816},{},[817],{"type":59,"value":818},"Keep the user experience lightweight by default.",{"type":54,"tag":72,"props":820,"children":821},{},[822],{"type":59,"value":823},"Prefer direct-to-cuOpt for one-off work.",{"type":54,"tag":72,"props":825,"children":826},{},[827],{"type":59,"value":828},"Use replayable\u002Fauditable mode only when the user’s needs justify the extra structure.",{"items":830,"total":918},[831,846,857,871,884,897,912],{"slug":832,"name":832,"fn":833,"description":834,"org":835,"tags":836,"stars":20,"repoUrl":21,"updatedAt":845},"cuopt-debugging","debug NVIDIA cuOpt optimization problems","Troubleshoot cuOpt LP\u002FMILP problems including errors, wrong results, infeasible solutions, performance issues, and status codes. Use when the user says something isn't working, gets unexpected results, or needs help diagnosing issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[837,840,843,844],{"name":838,"slug":839,"type":15},"Debugging","debugging",{"name":841,"slug":842,"type":15},"Mathematics","mathematics",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-23T05:43:32.51428",{"slug":215,"name":215,"fn":847,"description":848,"org":849,"tags":850,"stars":20,"repoUrl":21,"updatedAt":856},"map optimization problems to cuOpt models","Map interpreted optimization problems into cuOpt-native models for the fast path with minimal clarifying questions.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[851,854,855],{"name":852,"slug":853,"type":15},"Data Modeling","data-modeling",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-30T05:29:19.401656",{"slug":250,"name":250,"fn":858,"description":859,"org":860,"tags":861,"stars":20,"repoUrl":21,"updatedAt":870},"run cuOpt optimization in NemoClaw sandbox","Run cuOpt in the NemoClaw sandbox — probe\u002Fsmoke gates, prefer cancelable Python gRPC jobs, use legacy remote execution only when that API is unavailable, then vendored cuOpt skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[862,865,866,867],{"name":863,"slug":864,"type":15},"CLI","cli",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"name":868,"slug":869,"type":15},"Simulation","simulation","2026-07-30T05:29:17.131364",{"slug":872,"name":872,"fn":873,"description":874,"org":875,"tags":876,"stars":20,"repoUrl":21,"updatedAt":883},"generic-max-supply","plan supply chain models with cuOpt","Multi-period supply chain planning model: data files, BOM structure, variable\u002Fconstraint reference for the max-supply base model.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[877,878,879,880],{"name":852,"slug":853,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"name":881,"slug":882,"type":15},"Supply Chain","supply-chain","2026-07-14T05:32:29.39577",{"slug":885,"name":885,"fn":886,"description":887,"org":888,"tags":889,"stars":20,"repoUrl":21,"updatedAt":896},"optimization-from-data-orchestrator","orchestrate optimization data and solving","Coordinate uploaded data plus a natural-language question into interpretation, clarification, cuOpt solve, and a user-facing answer.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[890,891,894,895],{"name":13,"slug":14,"type":15},{"name":892,"slug":893,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-14T05:32:59.680482",{"slug":898,"name":898,"fn":899,"description":900,"org":901,"tags":902,"stars":20,"repoUrl":21,"updatedAt":911},"optimization-intent-router","classify optimization and analytics requests","Classify whether a data-backed request is LP, MILP, QP, routing, or non-optimization analytics.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[903,906,909,910],{"name":904,"slug":905,"type":15},"Analytics","analytics",{"name":907,"slug":908,"type":15},"Data Analysis","data-analysis",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-14T05:32:53.475243",{"slug":4,"name":4,"fn":5,"description":6,"org":913,"tags":914,"stars":20,"repoUrl":21,"updatedAt":22},{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[915,916,917],{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},8,{"items":920,"total":1075},[921,939,957,968,980,992,1005,1019,1032,1043,1057,1066],{"slug":922,"name":922,"fn":923,"description":924,"org":925,"tags":926,"stars":936,"repoUrl":937,"updatedAt":938},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[927,930,933],{"name":928,"slug":929,"type":15},"Documentation","documentation",{"name":931,"slug":932,"type":15},"MCP","mcp",{"name":934,"slug":935,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":940,"name":940,"fn":941,"description":942,"org":943,"tags":944,"stars":954,"repoUrl":955,"updatedAt":956},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[945,948,951],{"name":946,"slug":947,"type":15},"Containers","containers",{"name":949,"slug":950,"type":15},"Deployment","deployment",{"name":952,"slug":953,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":958,"name":958,"fn":959,"description":960,"org":961,"tags":962,"stars":954,"repoUrl":955,"updatedAt":967},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[963,966],{"name":964,"slug":965,"type":15},"CI\u002FCD","ci-cd",{"name":949,"slug":950,"type":15},"2026-07-14T05:25:59.97109",{"slug":969,"name":969,"fn":970,"description":971,"org":972,"tags":973,"stars":954,"repoUrl":955,"updatedAt":979},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[974,975,976],{"name":964,"slug":965,"type":15},{"name":949,"slug":950,"type":15},{"name":977,"slug":978,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":981,"name":981,"fn":982,"description":983,"org":984,"tags":985,"stars":954,"repoUrl":955,"updatedAt":991},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[986,987,988],{"name":838,"slug":839,"type":15},{"name":977,"slug":978,"type":15},{"name":989,"slug":990,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":993,"name":993,"fn":994,"description":995,"org":996,"tags":997,"stars":954,"repoUrl":955,"updatedAt":1004},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[998,1001],{"name":999,"slug":1000,"type":15},"Best Practices","best-practices",{"name":1002,"slug":1003,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1006,"name":1006,"fn":1007,"description":1008,"org":1009,"tags":1010,"stars":954,"repoUrl":955,"updatedAt":1018},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1011,1014,1017],{"name":1012,"slug":1013,"type":15},"Machine Learning","machine-learning",{"name":1015,"slug":1016,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1020,"name":1020,"fn":1021,"description":1022,"org":1023,"tags":1024,"stars":954,"repoUrl":955,"updatedAt":1031},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1025,1028],{"name":1026,"slug":1027,"type":15},"QA","qa",{"name":1029,"slug":1030,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":1033,"name":1033,"fn":1034,"description":1035,"org":1036,"tags":1037,"stars":954,"repoUrl":955,"updatedAt":1042},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1038,1039],{"name":949,"slug":950,"type":15},{"name":1040,"slug":1041,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":1044,"name":1044,"fn":1045,"description":1046,"org":1047,"tags":1048,"stars":954,"repoUrl":955,"updatedAt":1056},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1049,1052,1053],{"name":1050,"slug":1051,"type":15},"Code Review","code-review",{"name":977,"slug":978,"type":15},{"name":1054,"slug":1055,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1058,"name":1058,"fn":1059,"description":1060,"org":1061,"tags":1062,"stars":954,"repoUrl":955,"updatedAt":1065},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1063,1064],{"name":1026,"slug":1027,"type":15},{"name":1029,"slug":1030,"type":15},"2026-07-14T05:25:54.928983",{"slug":1067,"name":1067,"fn":1068,"description":1069,"org":1070,"tags":1071,"stars":954,"repoUrl":955,"updatedAt":1074},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1072,1073],{"name":13,"slug":14,"type":15},{"name":964,"slug":965,"type":15},"2026-07-30T05:29:03.275638",496]