[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-optimization-intent-router":3,"mdc-824mjc-key":43,"related-repo-nvidia-optimization-intent-router":1340,"related-org-nvidia-optimization-intent-router":1428},{"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":38,"sourceUrl":41,"mdContent":42},"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},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,20],{"name":13,"slug":14,"type":15},"Data Analysis","data-analysis","tag",{"name":17,"slug":18,"type":15},"Optimization","optimization",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Analytics","analytics",462,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcuopt-examples","2026-07-14T05:32:53.475243","Apache-2.0",81,[29,30,31,32,33,18,34,35,36,37],"combinatorial-optimization","gpu","linear-programming","mixed-integer-programming","operations-research","optimization-algorithms","route-optimization","traveling-salesman-problem","vehicle-routing-problem",{"repoUrl":24,"stars":23,"forks":27,"topics":39,"description":40},[29,30,31,32,33,18,34,35,36,37],"NVIDIA cuOpt examples for decision optimization","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcuopt-examples\u002Ftree\u002FHEAD\u002Fcuopt_on_nemoclaw\u002Fopenclaw-skills\u002Foptimization-intent-router","---\nname: optimization-intent-router\nversion: \"26.06.01\"\ndescription: Classify whether a data-backed request is LP, MILP, QP, routing, or non-optimization analytics.\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 Intent Router\n\nUse this skill when a user asks a question in natural language and it is not yet clear whether the request should be handled as:\n\n- a numerical optimization problem (LP \u002F MILP \u002F QP)\n- a routing problem (VRP \u002F TSP \u002F PDP)\n- or a non-optimization task such as analytics, filtering, aggregation, forecasting, or explanation\n\nThis skill exists to decide whether cuOpt should be involved at all, and if so, which downstream formulation path should be used.\n\nThe classification from this skill is provisional. A later data-ingestion step may refine or correct the problem family if the uploaded tables clearly support a different interpretation.\n\n## Purpose\n\nThe user often does not say:\n- \"this is an LP\"\n- \"this is a MILP\"\n- \"this is a routing problem\"\n\nInstead they ask questions like:\n- \"What’s the best production plan?\"\n- \"How should we allocate inventory?\"\n- \"What’s the optimal sales mix?\"\n- \"How many trucks should go to each depot?\"\n- \"Can you minimize delivery cost?\"\n\nYour job is to recognize when the request is really asking for an optimization model.\n\n## Core rule\n\nRoute to optimization when the user wants a **plan produced under\nconstraints** — assignment, schedule, allocation, route, mix, or mix of\nresources — whether they say **feasible\u002Fvalid**, **minimize\u002Fmaximize**, or\n**best\u002Foptimal**.\n\nUse **structure**, not exact keywords:\n\n| Signal | Route to cuOpt? |\n|---|---|\n| Constructive + constraint tables (uploaded or referenced) | Yes — family TBD (MILP\u002FLP\u002FQP\u002Frouting) |\n| Minimize \u002F maximize \u002F best \u002F optimal \u002F least cost | Yes |\n| Feasible \u002F valid \u002F make it work \u002F respect rules (constructive) | Yes — feasibility = hard constraints |\n| Summarize \u002F chart \u002F filter \u002F explain data only | No |\n\n**NemoClaw sandbox:** When constructive + constraint data →\n**`cuopt-sandbox` gates immediately** before any schedule, roster,\nassignment, or route output — regardless of whether the user said\nminimize or optimal. Do not use a greedy Python builder as a first\ndeliverable.\n\nParaphrase examples (non-exhaustive): \"plan the season\", \"set game times\",\n\"fill the calendar\", \"assign shifts\", \"put jobs on machines\", \"build a\nschedule\" — same intent. See `cuopt-sandbox\u002Freferences\u002Fintent-and-triggers.md`.\n\nIf the user is only asking for:\n- descriptive statistics\n- filtering\n- counting\n- sorting\n- charting\n- SQL-style aggregation\n- explanation of existing data\n\nthen do **not** force the request into optimization.\n\n## Signals that this is an optimization task\n\nStrong signals:\n- words like **optimize**, **optimal**, **best**, **maximize**, **minimize**\n- **build \u002F create \u002F assign** a schedule, roster, shift plan, or calendar\n  from structured data (sandbox: treat as MILP even without minimize)\n- resource tradeoffs: capacity, budget, time, labor, inventory, demand, hours, materials\n- decision language: how much, how many, which, assign, allocate, route, schedule, choose\n- explicit constraints: at most, at least, must, cannot exceed, within budget, limited by\n- competing objectives: profit vs capacity, cost vs service, coverage vs distance\n\nWeaker but meaningful signals:\n- \"What should we do?\"\n- \"What is the best plan?\"\n- \"How should we allocate this?\"\n- \"How can we reduce cost while meeting demand?\"\n\nWhen weaker signals appear, inspect whether there are real constraints and decisions. If yes, treat it as optimization — and in the NemoClaw sandbox, **`cuopt-sandbox` gates apply immediately** (probe before any schedule\u002Fheuristic output).\n\n## Route classification\n\n### Route to LP\nUse LP when:\n- the objective is linear\n- the constraints are linear\n- decision variables can be continuous or fractional\n\nCommon examples:\n- product mix\n- production planning\n- blending\n- budget allocation\n- transportation flow without integrality requirements\n\n### Route to MILP\nUse MILP when:\n- the model is otherwise linear\n- but some decisions must be integer or binary\n\nCommon signals:\n- units must be whole numbers\n- yes\u002Fno decisions\n- open\u002Fclose decisions\n- assign-or-not decisions\n- minimum lot sizes\n- discrete staffing or vehicle counts\n\nCommon examples:\n- facility opening\n- workforce scheduling with headcounts\n- assignment with binary decisions\n- product counts that must be whole\n- **slot\u002Fresource scheduling** (games, shifts, appointments → time slots\n  and resources) — including when the user only says \"build a schedule\"\n\n### Route to QP\nUse QP when:\n- the objective contains squared terms or interactions\n- the user wants quadratic minimization under linear constraints\n\nCommon signals:\n- variance minimization\n- least squares\n- portfolio optimization with covariance \u002F risk terms\n- quadratic penalty terms\n\nIf the user is maximizing a quadratic expression, note that the modeling path may require reformulation as a minimization of the negated objective.\n\n### Route to routing\nUse routing when the decisions are fundamentally about movement through locations.\n\nCommon signals:\n- deliveries\n- pickups and dropoffs\n- stops\n- depots\n- vehicles\n- routes\n- travel time \u002F distance \u002F cost matrices\n- time windows\n- capacities on vehicles\n- pickup-delivery pairing\n\nCommon routing types:\n- TSP: one route, visit locations\n- VRP: multiple vehicles, capacities and\u002For time limits\n- PDP: paired pickups and deliveries\n\n### Route to non-optimization handling\nDo **not** route to cuOpt when the user is asking for:\n- summaries\n- analytics\n- cleaning\n- forecasting without optimization\n- dashboards \u002F charts\n- explanation of data patterns\n- SQL \u002F filtering tasks\n\n## Ambiguity handling\n\nIf it is unclear whether the request is optimization or analytics, ask a concise clarifier such as:\n\n- **\"Are you asking for a summary of the data, or the best decision under constraints?\"**\n- **\"Do you want analytics from the data, or should I optimize a plan from it?\"**\n\nIf it is clearly optimization but unclear which family applies, ask the minimum clarifier needed.\n\nExamples:\n- **\"Do these quantities need to be whole numbers, or can fractional values be allowed?\"**\n- **\"Is this about assigning routes\u002Fvehicles to locations, or just allocating quantities across products\u002Fresources?\"**\n- **\"Is the objective linear, or does it include variance \u002F squared penalties \u002F interaction terms?\"**\n\n## Required output of this skill\n\nBefore handing off, produce an internal working conclusion with at least:\n- `is_optimization`: yes \u002F no\n- `problem_family`: lp | milp | qp | routing | unknown\n- `why`: short explanation grounded in the user’s wording and available data\n- `missing_information`: only the minimum unresolved items\n\n## Handoff guidance\n\n- If `problem_family = lp | milp`:\n  - hand off to `numerical-optimization-formulation`\n  - then to `cuopt-numerical-optimization-api-python` (or\n    `cuopt-numerical-optimization-api-cli` for MPS inputs)\n\n- If `problem_family = qp`:\n  - hand off to `numerical-optimization-formulation`\n  - then to `cuopt-numerical-optimization-api-python`\n\n- If `problem_family = routing`:\n  - hand off to `routing-formulation`\n  - then to `cuopt-routing-api-python`\n\n- If `is_optimization = yes` and there is a meaningful signal that replayability, audit, export, or reuse may matter:\n  - use `optimization-mode-router` before deep data interpretation or model construction\n\n- If `is_optimization = no`:\n  - do not invoke cuOpt just because the data could in theory be optimized\n\n## Examples\n\n### Example 1: product mix question\nUser says:\n> \"I uploaded a product table and a capacity table. What’s the best production plan?\"\n\nInterpretation:\n- this is optimization\n- likely LP or MILP\n- the user is asking for the best decision under resource constraints\n\n### Example 2: routing question\nUser says:\n> \"I uploaded depots, customers, and a travel-time matrix. What’s the cheapest delivery plan?\"\n\nInterpretation:\n- this is optimization\n- specifically routing\n- path construction and travel costs are central to the decision\n\n### Example 3: analytics question\nUser says:\n> \"I uploaded sales.csv. Which product had the highest revenue last month?\"\n\nInterpretation:\n- this is not optimization\n- it is descriptive analysis over existing data\n- do not route to cuOpt\n\n## Behavioral guardrails\n\n- Do not label a request as optimization only because it contains numbers.\n- Do not force a routing interpretation when the task is really allocation or assignment without path planning.\n- Do not force LP\u002FMILP when the question is clearly descriptive analytics.\n- Prefer one short clarifying question over building the wrong model.\n- When the user asks for the best decision under constraints, prefer optimization even if they do not use formal mathematical language.\n",{"data":44,"body":53},{"name":4,"version":45,"description":6,"license":26,"metadata":46,"origin":52},"26.06.01",{"author":47,"tags":48},"NVIDIA cuOpt Team",[49,50,51],"cuopt","nemoclaw","orchestration","skill-evolution",{"type":54,"children":55},"root",[56,64,70,90,95,100,107,112,130,135,163,168,174,208,220,300,324,336,341,379,391,397,402,469,474,497,514,520,527,532,550,555,583,589,594,607,612,645,649,682,688,693,706,710,733,738,744,749,753,806,811,829,835,846,883,889,894,913,918,923,950,956,961,1008,1014,1178,1184,1190,1195,1204,1209,1227,1233,1237,1245,1249,1266,1272,1276,1284,1288,1306,1312],{"type":57,"tag":58,"props":59,"children":60},"element","h1",{"id":4},[61],{"type":62,"value":63},"text","Optimization Intent Router",{"type":57,"tag":65,"props":66,"children":67},"p",{},[68],{"type":62,"value":69},"Use this skill when a user asks a question in natural language and it is not yet clear whether the request should be handled as:",{"type":57,"tag":71,"props":72,"children":73},"ul",{},[74,80,85],{"type":57,"tag":75,"props":76,"children":77},"li",{},[78],{"type":62,"value":79},"a numerical optimization problem (LP \u002F MILP \u002F QP)",{"type":57,"tag":75,"props":81,"children":82},{},[83],{"type":62,"value":84},"a routing problem (VRP \u002F TSP \u002F PDP)",{"type":57,"tag":75,"props":86,"children":87},{},[88],{"type":62,"value":89},"or a non-optimization task such as analytics, filtering, aggregation, forecasting, or explanation",{"type":57,"tag":65,"props":91,"children":92},{},[93],{"type":62,"value":94},"This skill exists to decide whether cuOpt should be involved at all, and if so, which downstream formulation path should be used.",{"type":57,"tag":65,"props":96,"children":97},{},[98],{"type":62,"value":99},"The classification from this skill is provisional. A later data-ingestion step may refine or correct the problem family if the uploaded tables clearly support a different interpretation.",{"type":57,"tag":101,"props":102,"children":104},"h2",{"id":103},"purpose",[105],{"type":62,"value":106},"Purpose",{"type":57,"tag":65,"props":108,"children":109},{},[110],{"type":62,"value":111},"The user often does not say:",{"type":57,"tag":71,"props":113,"children":114},{},[115,120,125],{"type":57,"tag":75,"props":116,"children":117},{},[118],{"type":62,"value":119},"\"this is an LP\"",{"type":57,"tag":75,"props":121,"children":122},{},[123],{"type":62,"value":124},"\"this is a MILP\"",{"type":57,"tag":75,"props":126,"children":127},{},[128],{"type":62,"value":129},"\"this is a routing problem\"",{"type":57,"tag":65,"props":131,"children":132},{},[133],{"type":62,"value":134},"Instead they ask questions like:",{"type":57,"tag":71,"props":136,"children":137},{},[138,143,148,153,158],{"type":57,"tag":75,"props":139,"children":140},{},[141],{"type":62,"value":142},"\"What’s the best production plan?\"",{"type":57,"tag":75,"props":144,"children":145},{},[146],{"type":62,"value":147},"\"How should we allocate inventory?\"",{"type":57,"tag":75,"props":149,"children":150},{},[151],{"type":62,"value":152},"\"What’s the optimal sales mix?\"",{"type":57,"tag":75,"props":154,"children":155},{},[156],{"type":62,"value":157},"\"How many trucks should go to each depot?\"",{"type":57,"tag":75,"props":159,"children":160},{},[161],{"type":62,"value":162},"\"Can you minimize delivery cost?\"",{"type":57,"tag":65,"props":164,"children":165},{},[166],{"type":62,"value":167},"Your job is to recognize when the request is really asking for an optimization model.",{"type":57,"tag":101,"props":169,"children":171},{"id":170},"core-rule",[172],{"type":62,"value":173},"Core rule",{"type":57,"tag":65,"props":175,"children":176},{},[177,179,185,187,192,194,199,201,206],{"type":62,"value":178},"Route to optimization when the user wants a ",{"type":57,"tag":180,"props":181,"children":182},"strong",{},[183],{"type":62,"value":184},"plan produced under\nconstraints",{"type":62,"value":186}," — assignment, schedule, allocation, route, mix, or mix of\nresources — whether they say ",{"type":57,"tag":180,"props":188,"children":189},{},[190],{"type":62,"value":191},"feasible\u002Fvalid",{"type":62,"value":193},", ",{"type":57,"tag":180,"props":195,"children":196},{},[197],{"type":62,"value":198},"minimize\u002Fmaximize",{"type":62,"value":200},", or\n",{"type":57,"tag":180,"props":202,"children":203},{},[204],{"type":62,"value":205},"best\u002Foptimal",{"type":62,"value":207},".",{"type":57,"tag":65,"props":209,"children":210},{},[211,213,218],{"type":62,"value":212},"Use ",{"type":57,"tag":180,"props":214,"children":215},{},[216],{"type":62,"value":217},"structure",{"type":62,"value":219},", not exact keywords:",{"type":57,"tag":221,"props":222,"children":223},"table",{},[224,243],{"type":57,"tag":225,"props":226,"children":227},"thead",{},[228],{"type":57,"tag":229,"props":230,"children":231},"tr",{},[232,238],{"type":57,"tag":233,"props":234,"children":235},"th",{},[236],{"type":62,"value":237},"Signal",{"type":57,"tag":233,"props":239,"children":240},{},[241],{"type":62,"value":242},"Route to cuOpt?",{"type":57,"tag":244,"props":245,"children":246},"tbody",{},[247,261,274,287],{"type":57,"tag":229,"props":248,"children":249},{},[250,256],{"type":57,"tag":251,"props":252,"children":253},"td",{},[254],{"type":62,"value":255},"Constructive + constraint tables (uploaded or referenced)",{"type":57,"tag":251,"props":257,"children":258},{},[259],{"type":62,"value":260},"Yes — family TBD (MILP\u002FLP\u002FQP\u002Frouting)",{"type":57,"tag":229,"props":262,"children":263},{},[264,269],{"type":57,"tag":251,"props":265,"children":266},{},[267],{"type":62,"value":268},"Minimize \u002F maximize \u002F best \u002F optimal \u002F least cost",{"type":57,"tag":251,"props":270,"children":271},{},[272],{"type":62,"value":273},"Yes",{"type":57,"tag":229,"props":275,"children":276},{},[277,282],{"type":57,"tag":251,"props":278,"children":279},{},[280],{"type":62,"value":281},"Feasible \u002F valid \u002F make it work \u002F respect rules (constructive)",{"type":57,"tag":251,"props":283,"children":284},{},[285],{"type":62,"value":286},"Yes — feasibility = hard constraints",{"type":57,"tag":229,"props":288,"children":289},{},[290,295],{"type":57,"tag":251,"props":291,"children":292},{},[293],{"type":62,"value":294},"Summarize \u002F chart \u002F filter \u002F explain data only",{"type":57,"tag":251,"props":296,"children":297},{},[298],{"type":62,"value":299},"No",{"type":57,"tag":65,"props":301,"children":302},{},[303,308,310,322],{"type":57,"tag":180,"props":304,"children":305},{},[306],{"type":62,"value":307},"NemoClaw sandbox:",{"type":62,"value":309}," When constructive + constraint data →\n",{"type":57,"tag":180,"props":311,"children":312},{},[313,320],{"type":57,"tag":314,"props":315,"children":317},"code",{"className":316},[],[318],{"type":62,"value":319},"cuopt-sandbox",{"type":62,"value":321}," gates immediately",{"type":62,"value":323}," before any schedule, roster,\nassignment, or route output — regardless of whether the user said\nminimize or optimal. Do not use a greedy Python builder as a first\ndeliverable.",{"type":57,"tag":65,"props":325,"children":326},{},[327,329,335],{"type":62,"value":328},"Paraphrase examples (non-exhaustive): \"plan the season\", \"set game times\",\n\"fill the calendar\", \"assign shifts\", \"put jobs on machines\", \"build a\nschedule\" — same intent. See ",{"type":57,"tag":314,"props":330,"children":332},{"className":331},[],[333],{"type":62,"value":334},"cuopt-sandbox\u002Freferences\u002Fintent-and-triggers.md",{"type":62,"value":207},{"type":57,"tag":65,"props":337,"children":338},{},[339],{"type":62,"value":340},"If the user is only asking for:",{"type":57,"tag":71,"props":342,"children":343},{},[344,349,354,359,364,369,374],{"type":57,"tag":75,"props":345,"children":346},{},[347],{"type":62,"value":348},"descriptive statistics",{"type":57,"tag":75,"props":350,"children":351},{},[352],{"type":62,"value":353},"filtering",{"type":57,"tag":75,"props":355,"children":356},{},[357],{"type":62,"value":358},"counting",{"type":57,"tag":75,"props":360,"children":361},{},[362],{"type":62,"value":363},"sorting",{"type":57,"tag":75,"props":365,"children":366},{},[367],{"type":62,"value":368},"charting",{"type":57,"tag":75,"props":370,"children":371},{},[372],{"type":62,"value":373},"SQL-style aggregation",{"type":57,"tag":75,"props":375,"children":376},{},[377],{"type":62,"value":378},"explanation of existing data",{"type":57,"tag":65,"props":380,"children":381},{},[382,384,389],{"type":62,"value":383},"then do ",{"type":57,"tag":180,"props":385,"children":386},{},[387],{"type":62,"value":388},"not",{"type":62,"value":390}," force the request into optimization.",{"type":57,"tag":101,"props":392,"children":394},{"id":393},"signals-that-this-is-an-optimization-task",[395],{"type":62,"value":396},"Signals that this is an optimization task",{"type":57,"tag":65,"props":398,"children":399},{},[400],{"type":62,"value":401},"Strong signals:",{"type":57,"tag":71,"props":403,"children":404},{},[405,439,449,454,459,464],{"type":57,"tag":75,"props":406,"children":407},{},[408,410,415,416,421,422,427,428,433,434],{"type":62,"value":409},"words like ",{"type":57,"tag":180,"props":411,"children":412},{},[413],{"type":62,"value":414},"optimize",{"type":62,"value":193},{"type":57,"tag":180,"props":417,"children":418},{},[419],{"type":62,"value":420},"optimal",{"type":62,"value":193},{"type":57,"tag":180,"props":423,"children":424},{},[425],{"type":62,"value":426},"best",{"type":62,"value":193},{"type":57,"tag":180,"props":429,"children":430},{},[431],{"type":62,"value":432},"maximize",{"type":62,"value":193},{"type":57,"tag":180,"props":435,"children":436},{},[437],{"type":62,"value":438},"minimize",{"type":57,"tag":75,"props":440,"children":441},{},[442,447],{"type":57,"tag":180,"props":443,"children":444},{},[445],{"type":62,"value":446},"build \u002F create \u002F assign",{"type":62,"value":448}," a schedule, roster, shift plan, or calendar\nfrom structured data (sandbox: treat as MILP even without minimize)",{"type":57,"tag":75,"props":450,"children":451},{},[452],{"type":62,"value":453},"resource tradeoffs: capacity, budget, time, labor, inventory, demand, hours, materials",{"type":57,"tag":75,"props":455,"children":456},{},[457],{"type":62,"value":458},"decision language: how much, how many, which, assign, allocate, route, schedule, choose",{"type":57,"tag":75,"props":460,"children":461},{},[462],{"type":62,"value":463},"explicit constraints: at most, at least, must, cannot exceed, within budget, limited by",{"type":57,"tag":75,"props":465,"children":466},{},[467],{"type":62,"value":468},"competing objectives: profit vs capacity, cost vs service, coverage vs distance",{"type":57,"tag":65,"props":470,"children":471},{},[472],{"type":62,"value":473},"Weaker but meaningful signals:",{"type":57,"tag":71,"props":475,"children":476},{},[477,482,487,492],{"type":57,"tag":75,"props":478,"children":479},{},[480],{"type":62,"value":481},"\"What should we do?\"",{"type":57,"tag":75,"props":483,"children":484},{},[485],{"type":62,"value":486},"\"What is the best plan?\"",{"type":57,"tag":75,"props":488,"children":489},{},[490],{"type":62,"value":491},"\"How should we allocate this?\"",{"type":57,"tag":75,"props":493,"children":494},{},[495],{"type":62,"value":496},"\"How can we reduce cost while meeting demand?\"",{"type":57,"tag":65,"props":498,"children":499},{},[500,502,512],{"type":62,"value":501},"When weaker signals appear, inspect whether there are real constraints and decisions. If yes, treat it as optimization — and in the NemoClaw sandbox, ",{"type":57,"tag":180,"props":503,"children":504},{},[505,510],{"type":57,"tag":314,"props":506,"children":508},{"className":507},[],[509],{"type":62,"value":319},{"type":62,"value":511}," gates apply immediately",{"type":62,"value":513}," (probe before any schedule\u002Fheuristic output).",{"type":57,"tag":101,"props":515,"children":517},{"id":516},"route-classification",[518],{"type":62,"value":519},"Route classification",{"type":57,"tag":521,"props":522,"children":524},"h3",{"id":523},"route-to-lp",[525],{"type":62,"value":526},"Route to LP",{"type":57,"tag":65,"props":528,"children":529},{},[530],{"type":62,"value":531},"Use LP when:",{"type":57,"tag":71,"props":533,"children":534},{},[535,540,545],{"type":57,"tag":75,"props":536,"children":537},{},[538],{"type":62,"value":539},"the objective is linear",{"type":57,"tag":75,"props":541,"children":542},{},[543],{"type":62,"value":544},"the constraints are linear",{"type":57,"tag":75,"props":546,"children":547},{},[548],{"type":62,"value":549},"decision variables can be continuous or fractional",{"type":57,"tag":65,"props":551,"children":552},{},[553],{"type":62,"value":554},"Common examples:",{"type":57,"tag":71,"props":556,"children":557},{},[558,563,568,573,578],{"type":57,"tag":75,"props":559,"children":560},{},[561],{"type":62,"value":562},"product mix",{"type":57,"tag":75,"props":564,"children":565},{},[566],{"type":62,"value":567},"production planning",{"type":57,"tag":75,"props":569,"children":570},{},[571],{"type":62,"value":572},"blending",{"type":57,"tag":75,"props":574,"children":575},{},[576],{"type":62,"value":577},"budget allocation",{"type":57,"tag":75,"props":579,"children":580},{},[581],{"type":62,"value":582},"transportation flow without integrality requirements",{"type":57,"tag":521,"props":584,"children":586},{"id":585},"route-to-milp",[587],{"type":62,"value":588},"Route to MILP",{"type":57,"tag":65,"props":590,"children":591},{},[592],{"type":62,"value":593},"Use MILP when:",{"type":57,"tag":71,"props":595,"children":596},{},[597,602],{"type":57,"tag":75,"props":598,"children":599},{},[600],{"type":62,"value":601},"the model is otherwise linear",{"type":57,"tag":75,"props":603,"children":604},{},[605],{"type":62,"value":606},"but some decisions must be integer or binary",{"type":57,"tag":65,"props":608,"children":609},{},[610],{"type":62,"value":611},"Common signals:",{"type":57,"tag":71,"props":613,"children":614},{},[615,620,625,630,635,640],{"type":57,"tag":75,"props":616,"children":617},{},[618],{"type":62,"value":619},"units must be whole numbers",{"type":57,"tag":75,"props":621,"children":622},{},[623],{"type":62,"value":624},"yes\u002Fno decisions",{"type":57,"tag":75,"props":626,"children":627},{},[628],{"type":62,"value":629},"open\u002Fclose decisions",{"type":57,"tag":75,"props":631,"children":632},{},[633],{"type":62,"value":634},"assign-or-not decisions",{"type":57,"tag":75,"props":636,"children":637},{},[638],{"type":62,"value":639},"minimum lot sizes",{"type":57,"tag":75,"props":641,"children":642},{},[643],{"type":62,"value":644},"discrete staffing or vehicle counts",{"type":57,"tag":65,"props":646,"children":647},{},[648],{"type":62,"value":554},{"type":57,"tag":71,"props":650,"children":651},{},[652,657,662,667,672],{"type":57,"tag":75,"props":653,"children":654},{},[655],{"type":62,"value":656},"facility opening",{"type":57,"tag":75,"props":658,"children":659},{},[660],{"type":62,"value":661},"workforce scheduling with headcounts",{"type":57,"tag":75,"props":663,"children":664},{},[665],{"type":62,"value":666},"assignment with binary decisions",{"type":57,"tag":75,"props":668,"children":669},{},[670],{"type":62,"value":671},"product counts that must be whole",{"type":57,"tag":75,"props":673,"children":674},{},[675,680],{"type":57,"tag":180,"props":676,"children":677},{},[678],{"type":62,"value":679},"slot\u002Fresource scheduling",{"type":62,"value":681}," (games, shifts, appointments → time slots\nand resources) — including when the user only says \"build a schedule\"",{"type":57,"tag":521,"props":683,"children":685},{"id":684},"route-to-qp",[686],{"type":62,"value":687},"Route to QP",{"type":57,"tag":65,"props":689,"children":690},{},[691],{"type":62,"value":692},"Use QP when:",{"type":57,"tag":71,"props":694,"children":695},{},[696,701],{"type":57,"tag":75,"props":697,"children":698},{},[699],{"type":62,"value":700},"the objective contains squared terms or interactions",{"type":57,"tag":75,"props":702,"children":703},{},[704],{"type":62,"value":705},"the user wants quadratic minimization under linear constraints",{"type":57,"tag":65,"props":707,"children":708},{},[709],{"type":62,"value":611},{"type":57,"tag":71,"props":711,"children":712},{},[713,718,723,728],{"type":57,"tag":75,"props":714,"children":715},{},[716],{"type":62,"value":717},"variance minimization",{"type":57,"tag":75,"props":719,"children":720},{},[721],{"type":62,"value":722},"least squares",{"type":57,"tag":75,"props":724,"children":725},{},[726],{"type":62,"value":727},"portfolio optimization with covariance \u002F risk terms",{"type":57,"tag":75,"props":729,"children":730},{},[731],{"type":62,"value":732},"quadratic penalty terms",{"type":57,"tag":65,"props":734,"children":735},{},[736],{"type":62,"value":737},"If the user is maximizing a quadratic expression, note that the modeling path may require reformulation as a minimization of the negated objective.",{"type":57,"tag":521,"props":739,"children":741},{"id":740},"route-to-routing",[742],{"type":62,"value":743},"Route to routing",{"type":57,"tag":65,"props":745,"children":746},{},[747],{"type":62,"value":748},"Use routing when the decisions are fundamentally about movement through locations.",{"type":57,"tag":65,"props":750,"children":751},{},[752],{"type":62,"value":611},{"type":57,"tag":71,"props":754,"children":755},{},[756,761,766,771,776,781,786,791,796,801],{"type":57,"tag":75,"props":757,"children":758},{},[759],{"type":62,"value":760},"deliveries",{"type":57,"tag":75,"props":762,"children":763},{},[764],{"type":62,"value":765},"pickups and dropoffs",{"type":57,"tag":75,"props":767,"children":768},{},[769],{"type":62,"value":770},"stops",{"type":57,"tag":75,"props":772,"children":773},{},[774],{"type":62,"value":775},"depots",{"type":57,"tag":75,"props":777,"children":778},{},[779],{"type":62,"value":780},"vehicles",{"type":57,"tag":75,"props":782,"children":783},{},[784],{"type":62,"value":785},"routes",{"type":57,"tag":75,"props":787,"children":788},{},[789],{"type":62,"value":790},"travel time \u002F distance \u002F cost matrices",{"type":57,"tag":75,"props":792,"children":793},{},[794],{"type":62,"value":795},"time windows",{"type":57,"tag":75,"props":797,"children":798},{},[799],{"type":62,"value":800},"capacities on vehicles",{"type":57,"tag":75,"props":802,"children":803},{},[804],{"type":62,"value":805},"pickup-delivery pairing",{"type":57,"tag":65,"props":807,"children":808},{},[809],{"type":62,"value":810},"Common routing types:",{"type":57,"tag":71,"props":812,"children":813},{},[814,819,824],{"type":57,"tag":75,"props":815,"children":816},{},[817],{"type":62,"value":818},"TSP: one route, visit locations",{"type":57,"tag":75,"props":820,"children":821},{},[822],{"type":62,"value":823},"VRP: multiple vehicles, capacities and\u002For time limits",{"type":57,"tag":75,"props":825,"children":826},{},[827],{"type":62,"value":828},"PDP: paired pickups and deliveries",{"type":57,"tag":521,"props":830,"children":832},{"id":831},"route-to-non-optimization-handling",[833],{"type":62,"value":834},"Route to non-optimization handling",{"type":57,"tag":65,"props":836,"children":837},{},[838,840,844],{"type":62,"value":839},"Do ",{"type":57,"tag":180,"props":841,"children":842},{},[843],{"type":62,"value":388},{"type":62,"value":845}," route to cuOpt when the user is asking for:",{"type":57,"tag":71,"props":847,"children":848},{},[849,854,858,863,868,873,878],{"type":57,"tag":75,"props":850,"children":851},{},[852],{"type":62,"value":853},"summaries",{"type":57,"tag":75,"props":855,"children":856},{},[857],{"type":62,"value":22},{"type":57,"tag":75,"props":859,"children":860},{},[861],{"type":62,"value":862},"cleaning",{"type":57,"tag":75,"props":864,"children":865},{},[866],{"type":62,"value":867},"forecasting without optimization",{"type":57,"tag":75,"props":869,"children":870},{},[871],{"type":62,"value":872},"dashboards \u002F charts",{"type":57,"tag":75,"props":874,"children":875},{},[876],{"type":62,"value":877},"explanation of data patterns",{"type":57,"tag":75,"props":879,"children":880},{},[881],{"type":62,"value":882},"SQL \u002F filtering tasks",{"type":57,"tag":101,"props":884,"children":886},{"id":885},"ambiguity-handling",[887],{"type":62,"value":888},"Ambiguity handling",{"type":57,"tag":65,"props":890,"children":891},{},[892],{"type":62,"value":893},"If it is unclear whether the request is optimization or analytics, ask a concise clarifier such as:",{"type":57,"tag":71,"props":895,"children":896},{},[897,905],{"type":57,"tag":75,"props":898,"children":899},{},[900],{"type":57,"tag":180,"props":901,"children":902},{},[903],{"type":62,"value":904},"\"Are you asking for a summary of the data, or the best decision under constraints?\"",{"type":57,"tag":75,"props":906,"children":907},{},[908],{"type":57,"tag":180,"props":909,"children":910},{},[911],{"type":62,"value":912},"\"Do you want analytics from the data, or should I optimize a plan from it?\"",{"type":57,"tag":65,"props":914,"children":915},{},[916],{"type":62,"value":917},"If it is clearly optimization but unclear which family applies, ask the minimum clarifier needed.",{"type":57,"tag":65,"props":919,"children":920},{},[921],{"type":62,"value":922},"Examples:",{"type":57,"tag":71,"props":924,"children":925},{},[926,934,942],{"type":57,"tag":75,"props":927,"children":928},{},[929],{"type":57,"tag":180,"props":930,"children":931},{},[932],{"type":62,"value":933},"\"Do these quantities need to be whole numbers, or can fractional values be allowed?\"",{"type":57,"tag":75,"props":935,"children":936},{},[937],{"type":57,"tag":180,"props":938,"children":939},{},[940],{"type":62,"value":941},"\"Is this about assigning routes\u002Fvehicles to locations, or just allocating quantities across products\u002Fresources?\"",{"type":57,"tag":75,"props":943,"children":944},{},[945],{"type":57,"tag":180,"props":946,"children":947},{},[948],{"type":62,"value":949},"\"Is the objective linear, or does it include variance \u002F squared penalties \u002F interaction terms?\"",{"type":57,"tag":101,"props":951,"children":953},{"id":952},"required-output-of-this-skill",[954],{"type":62,"value":955},"Required output of this skill",{"type":57,"tag":65,"props":957,"children":958},{},[959],{"type":62,"value":960},"Before handing off, produce an internal working conclusion with at least:",{"type":57,"tag":71,"props":962,"children":963},{},[964,975,986,997],{"type":57,"tag":75,"props":965,"children":966},{},[967,973],{"type":57,"tag":314,"props":968,"children":970},{"className":969},[],[971],{"type":62,"value":972},"is_optimization",{"type":62,"value":974},": yes \u002F no",{"type":57,"tag":75,"props":976,"children":977},{},[978,984],{"type":57,"tag":314,"props":979,"children":981},{"className":980},[],[982],{"type":62,"value":983},"problem_family",{"type":62,"value":985},": lp | milp | qp | routing | unknown",{"type":57,"tag":75,"props":987,"children":988},{},[989,995],{"type":57,"tag":314,"props":990,"children":992},{"className":991},[],[993],{"type":62,"value":994},"why",{"type":62,"value":996},": short explanation grounded in the user’s wording and available data",{"type":57,"tag":75,"props":998,"children":999},{},[1000,1006],{"type":57,"tag":314,"props":1001,"children":1003},{"className":1002},[],[1004],{"type":62,"value":1005},"missing_information",{"type":62,"value":1007},": only the minimum unresolved items",{"type":57,"tag":101,"props":1009,"children":1011},{"id":1010},"handoff-guidance",[1012],{"type":62,"value":1013},"Handoff guidance",{"type":57,"tag":71,"props":1015,"children":1016},{},[1017,1065,1097,1131,1159],{"type":57,"tag":75,"props":1018,"children":1019},{},[1020,1022,1028,1030],{"type":62,"value":1021},"If ",{"type":57,"tag":314,"props":1023,"children":1025},{"className":1024},[],[1026],{"type":62,"value":1027},"problem_family = lp | milp",{"type":62,"value":1029},":",{"type":57,"tag":71,"props":1031,"children":1032},{},[1033,1044],{"type":57,"tag":75,"props":1034,"children":1035},{},[1036,1038],{"type":62,"value":1037},"hand off to ",{"type":57,"tag":314,"props":1039,"children":1041},{"className":1040},[],[1042],{"type":62,"value":1043},"numerical-optimization-formulation",{"type":57,"tag":75,"props":1045,"children":1046},{},[1047,1049,1055,1057,1063],{"type":62,"value":1048},"then to ",{"type":57,"tag":314,"props":1050,"children":1052},{"className":1051},[],[1053],{"type":62,"value":1054},"cuopt-numerical-optimization-api-python",{"type":62,"value":1056}," (or\n",{"type":57,"tag":314,"props":1058,"children":1060},{"className":1059},[],[1061],{"type":62,"value":1062},"cuopt-numerical-optimization-api-cli",{"type":62,"value":1064}," for MPS inputs)",{"type":57,"tag":75,"props":1066,"children":1067},{},[1068,1069,1075,1076],{"type":62,"value":1021},{"type":57,"tag":314,"props":1070,"children":1072},{"className":1071},[],[1073],{"type":62,"value":1074},"problem_family = qp",{"type":62,"value":1029},{"type":57,"tag":71,"props":1077,"children":1078},{},[1079,1088],{"type":57,"tag":75,"props":1080,"children":1081},{},[1082,1083],{"type":62,"value":1037},{"type":57,"tag":314,"props":1084,"children":1086},{"className":1085},[],[1087],{"type":62,"value":1043},{"type":57,"tag":75,"props":1089,"children":1090},{},[1091,1092],{"type":62,"value":1048},{"type":57,"tag":314,"props":1093,"children":1095},{"className":1094},[],[1096],{"type":62,"value":1054},{"type":57,"tag":75,"props":1098,"children":1099},{},[1100,1101,1107,1108],{"type":62,"value":1021},{"type":57,"tag":314,"props":1102,"children":1104},{"className":1103},[],[1105],{"type":62,"value":1106},"problem_family = routing",{"type":62,"value":1029},{"type":57,"tag":71,"props":1109,"children":1110},{},[1111,1121],{"type":57,"tag":75,"props":1112,"children":1113},{},[1114,1115],{"type":62,"value":1037},{"type":57,"tag":314,"props":1116,"children":1118},{"className":1117},[],[1119],{"type":62,"value":1120},"routing-formulation",{"type":57,"tag":75,"props":1122,"children":1123},{},[1124,1125],{"type":62,"value":1048},{"type":57,"tag":314,"props":1126,"children":1128},{"className":1127},[],[1129],{"type":62,"value":1130},"cuopt-routing-api-python",{"type":57,"tag":75,"props":1132,"children":1133},{},[1134,1135,1141,1143],{"type":62,"value":1021},{"type":57,"tag":314,"props":1136,"children":1138},{"className":1137},[],[1139],{"type":62,"value":1140},"is_optimization = yes",{"type":62,"value":1142}," and there is a meaningful signal that replayability, audit, export, or reuse may matter:",{"type":57,"tag":71,"props":1144,"children":1145},{},[1146],{"type":57,"tag":75,"props":1147,"children":1148},{},[1149,1151,1157],{"type":62,"value":1150},"use ",{"type":57,"tag":314,"props":1152,"children":1154},{"className":1153},[],[1155],{"type":62,"value":1156},"optimization-mode-router",{"type":62,"value":1158}," before deep data interpretation or model construction",{"type":57,"tag":75,"props":1160,"children":1161},{},[1162,1163,1169,1170],{"type":62,"value":1021},{"type":57,"tag":314,"props":1164,"children":1166},{"className":1165},[],[1167],{"type":62,"value":1168},"is_optimization = no",{"type":62,"value":1029},{"type":57,"tag":71,"props":1171,"children":1172},{},[1173],{"type":57,"tag":75,"props":1174,"children":1175},{},[1176],{"type":62,"value":1177},"do not invoke cuOpt just because the data could in theory be optimized",{"type":57,"tag":101,"props":1179,"children":1181},{"id":1180},"examples",[1182],{"type":62,"value":1183},"Examples",{"type":57,"tag":521,"props":1185,"children":1187},{"id":1186},"example-1-product-mix-question",[1188],{"type":62,"value":1189},"Example 1: product mix question",{"type":57,"tag":65,"props":1191,"children":1192},{},[1193],{"type":62,"value":1194},"User says:",{"type":57,"tag":1196,"props":1197,"children":1198},"blockquote",{},[1199],{"type":57,"tag":65,"props":1200,"children":1201},{},[1202],{"type":62,"value":1203},"\"I uploaded a product table and a capacity table. What’s the best production plan?\"",{"type":57,"tag":65,"props":1205,"children":1206},{},[1207],{"type":62,"value":1208},"Interpretation:",{"type":57,"tag":71,"props":1210,"children":1211},{},[1212,1217,1222],{"type":57,"tag":75,"props":1213,"children":1214},{},[1215],{"type":62,"value":1216},"this is optimization",{"type":57,"tag":75,"props":1218,"children":1219},{},[1220],{"type":62,"value":1221},"likely LP or MILP",{"type":57,"tag":75,"props":1223,"children":1224},{},[1225],{"type":62,"value":1226},"the user is asking for the best decision under resource constraints",{"type":57,"tag":521,"props":1228,"children":1230},{"id":1229},"example-2-routing-question",[1231],{"type":62,"value":1232},"Example 2: routing question",{"type":57,"tag":65,"props":1234,"children":1235},{},[1236],{"type":62,"value":1194},{"type":57,"tag":1196,"props":1238,"children":1239},{},[1240],{"type":57,"tag":65,"props":1241,"children":1242},{},[1243],{"type":62,"value":1244},"\"I uploaded depots, customers, and a travel-time matrix. What’s the cheapest delivery plan?\"",{"type":57,"tag":65,"props":1246,"children":1247},{},[1248],{"type":62,"value":1208},{"type":57,"tag":71,"props":1250,"children":1251},{},[1252,1256,1261],{"type":57,"tag":75,"props":1253,"children":1254},{},[1255],{"type":62,"value":1216},{"type":57,"tag":75,"props":1257,"children":1258},{},[1259],{"type":62,"value":1260},"specifically routing",{"type":57,"tag":75,"props":1262,"children":1263},{},[1264],{"type":62,"value":1265},"path construction and travel costs are central to the decision",{"type":57,"tag":521,"props":1267,"children":1269},{"id":1268},"example-3-analytics-question",[1270],{"type":62,"value":1271},"Example 3: analytics question",{"type":57,"tag":65,"props":1273,"children":1274},{},[1275],{"type":62,"value":1194},{"type":57,"tag":1196,"props":1277,"children":1278},{},[1279],{"type":57,"tag":65,"props":1280,"children":1281},{},[1282],{"type":62,"value":1283},"\"I uploaded sales.csv. Which product had the highest revenue last month?\"",{"type":57,"tag":65,"props":1285,"children":1286},{},[1287],{"type":62,"value":1208},{"type":57,"tag":71,"props":1289,"children":1290},{},[1291,1296,1301],{"type":57,"tag":75,"props":1292,"children":1293},{},[1294],{"type":62,"value":1295},"this is not optimization",{"type":57,"tag":75,"props":1297,"children":1298},{},[1299],{"type":62,"value":1300},"it is descriptive analysis over existing data",{"type":57,"tag":75,"props":1302,"children":1303},{},[1304],{"type":62,"value":1305},"do not route to cuOpt",{"type":57,"tag":101,"props":1307,"children":1309},{"id":1308},"behavioral-guardrails",[1310],{"type":62,"value":1311},"Behavioral guardrails",{"type":57,"tag":71,"props":1313,"children":1314},{},[1315,1320,1325,1330,1335],{"type":57,"tag":75,"props":1316,"children":1317},{},[1318],{"type":62,"value":1319},"Do not label a request as optimization only because it contains numbers.",{"type":57,"tag":75,"props":1321,"children":1322},{},[1323],{"type":62,"value":1324},"Do not force a routing interpretation when the task is really allocation or assignment without path planning.",{"type":57,"tag":75,"props":1326,"children":1327},{},[1328],{"type":62,"value":1329},"Do not force LP\u002FMILP when the question is clearly descriptive analytics.",{"type":57,"tag":75,"props":1331,"children":1332},{},[1333],{"type":62,"value":1334},"Prefer one short clarifying question over building the wrong model.",{"type":57,"tag":75,"props":1336,"children":1337},{},[1338],{"type":62,"value":1339},"When the user asks for the best decision under constraints, prefer optimization even if they do not use formal mathematical language.",{"items":1341,"total":1427},[1342,1357,1369,1383,1396,1411,1418],{"slug":1343,"name":1343,"fn":1344,"description":1345,"org":1346,"tags":1347,"stars":23,"repoUrl":24,"updatedAt":1356},"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},[1348,1351,1354,1355],{"name":1349,"slug":1350,"type":15},"Debugging","debugging",{"name":1352,"slug":1353,"type":15},"Mathematics","mathematics",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-23T05:43:32.51428",{"slug":1358,"name":1358,"fn":1359,"description":1360,"org":1361,"tags":1362,"stars":23,"repoUrl":24,"updatedAt":1368},"cuopt-model-mapper","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},[1363,1366,1367],{"name":1364,"slug":1365,"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":319,"name":319,"fn":1370,"description":1371,"org":1372,"tags":1373,"stars":23,"repoUrl":24,"updatedAt":1382},"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},[1374,1377,1378,1379],{"name":1375,"slug":1376,"type":15},"CLI","cli",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"name":1380,"slug":1381,"type":15},"Simulation","simulation","2026-07-30T05:29:17.131364",{"slug":1384,"name":1384,"fn":1385,"description":1386,"org":1387,"tags":1388,"stars":23,"repoUrl":24,"updatedAt":1395},"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},[1389,1390,1391,1392],{"name":1364,"slug":1365,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"name":1393,"slug":1394,"type":15},"Supply Chain","supply-chain","2026-07-14T05:32:29.39577",{"slug":1397,"name":1397,"fn":1398,"description":1399,"org":1400,"tags":1401,"stars":23,"repoUrl":24,"updatedAt":1410},"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},[1402,1405,1408,1409],{"name":1403,"slug":1404,"type":15},"Automation","automation",{"name":1406,"slug":1407,"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":4,"name":4,"fn":5,"description":6,"org":1412,"tags":1413,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1414,1415,1416,1417],{"name":21,"slug":22,"type":15},{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"slug":1156,"name":1156,"fn":1419,"description":1420,"org":1421,"tags":1422,"stars":23,"repoUrl":24,"updatedAt":1426},"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},[1423,1424,1425],{"name":1403,"slug":1404,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-30T05:29:18.162351",8,{"items":1429,"total":1584},[1430,1448,1466,1477,1489,1501,1514,1528,1541,1552,1566,1575],{"slug":1431,"name":1431,"fn":1432,"description":1433,"org":1434,"tags":1435,"stars":1445,"repoUrl":1446,"updatedAt":1447},"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},[1436,1439,1442],{"name":1437,"slug":1438,"type":15},"Documentation","documentation",{"name":1440,"slug":1441,"type":15},"MCP","mcp",{"name":1443,"slug":1444,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1449,"name":1449,"fn":1450,"description":1451,"org":1452,"tags":1453,"stars":1463,"repoUrl":1464,"updatedAt":1465},"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},[1454,1457,1460],{"name":1455,"slug":1456,"type":15},"Containers","containers",{"name":1458,"slug":1459,"type":15},"Deployment","deployment",{"name":1461,"slug":1462,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1467,"name":1467,"fn":1468,"description":1469,"org":1470,"tags":1471,"stars":1463,"repoUrl":1464,"updatedAt":1476},"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},[1472,1475],{"name":1473,"slug":1474,"type":15},"CI\u002FCD","ci-cd",{"name":1458,"slug":1459,"type":15},"2026-07-14T05:25:59.97109",{"slug":1478,"name":1478,"fn":1479,"description":1480,"org":1481,"tags":1482,"stars":1463,"repoUrl":1464,"updatedAt":1488},"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},[1483,1484,1485],{"name":1473,"slug":1474,"type":15},{"name":1458,"slug":1459,"type":15},{"name":1486,"slug":1487,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1490,"name":1490,"fn":1491,"description":1492,"org":1493,"tags":1494,"stars":1463,"repoUrl":1464,"updatedAt":1500},"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},[1495,1496,1497],{"name":1349,"slug":1350,"type":15},{"name":1486,"slug":1487,"type":15},{"name":1498,"slug":1499,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1502,"name":1502,"fn":1503,"description":1504,"org":1505,"tags":1506,"stars":1463,"repoUrl":1464,"updatedAt":1513},"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},[1507,1510],{"name":1508,"slug":1509,"type":15},"Best Practices","best-practices",{"name":1511,"slug":1512,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1515,"name":1515,"fn":1516,"description":1517,"org":1518,"tags":1519,"stars":1463,"repoUrl":1464,"updatedAt":1527},"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},[1520,1523,1526],{"name":1521,"slug":1522,"type":15},"Machine Learning","machine-learning",{"name":1524,"slug":1525,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1529,"name":1529,"fn":1530,"description":1531,"org":1532,"tags":1533,"stars":1463,"repoUrl":1464,"updatedAt":1540},"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},[1534,1537],{"name":1535,"slug":1536,"type":15},"QA","qa",{"name":1538,"slug":1539,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":1542,"name":1542,"fn":1543,"description":1544,"org":1545,"tags":1546,"stars":1463,"repoUrl":1464,"updatedAt":1551},"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},[1547,1548],{"name":1458,"slug":1459,"type":15},{"name":1549,"slug":1550,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":1553,"name":1553,"fn":1554,"description":1555,"org":1556,"tags":1557,"stars":1463,"repoUrl":1464,"updatedAt":1565},"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},[1558,1561,1562],{"name":1559,"slug":1560,"type":15},"Code Review","code-review",{"name":1486,"slug":1487,"type":15},{"name":1563,"slug":1564,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1567,"name":1567,"fn":1568,"description":1569,"org":1570,"tags":1571,"stars":1463,"repoUrl":1464,"updatedAt":1574},"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},[1572,1573],{"name":1535,"slug":1536,"type":15},{"name":1538,"slug":1539,"type":15},"2026-07-14T05:25:54.928983",{"slug":1576,"name":1576,"fn":1577,"description":1578,"org":1579,"tags":1580,"stars":1463,"repoUrl":1464,"updatedAt":1583},"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},[1581,1582],{"name":1403,"slug":1404,"type":15},{"name":1473,"slug":1474,"type":15},"2026-07-30T05:29:03.275638",496]