[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-tilegym-monkey-patch-kernels-to-transformers":3,"mdc-xysc37-key":34,"related-repo-nvidia-tilegym-monkey-patch-kernels-to-transformers":225,"related-org-nvidia-tilegym-monkey-patch-kernels-to-transformers":329},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"tilegym-monkey-patch-kernels-to-transformers","integrate TileGym kernels into Hugging Face transformers","Integrate TileGym kernels into Hugging Face `transformers` models by replacing the library's submodule(s) and certain class(es)' implementations, and patching certain class(es)' init\u002Fforward\u002Fload weight methods prior to instantiating models. Used when the user requires integrating TileGym kernels into `transformers` models.",{"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},"Performance","performance","tag",{"name":17,"slug":18,"type":15},"Hugging Face","hugging-face",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Engineering","engineering",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:27:13.964423","CC-BY-4.0 AND Apache-2.0",281,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Ftilegym-monkey-patch-kernels-to-transformers","---\nname: tilegym-monkey-patch-kernels-to-transformers\ndescription: Integrate TileGym kernels into Hugging Face `transformers` models by replacing the library's submodule(s) and certain class(es)' implementations, and patching certain class(es)' init\u002Fforward\u002Fload weight methods prior to instantiating models. Used when the user requires integrating TileGym kernels into `transformers` models.\nlicense: CC-BY-4.0 AND Apache-2.0\ncompatibility: Verified on Claude Code with Opus-4.6 and onward, CodeX with GPT-5.5 and onward, and Cursor (Agent mode) with GPT-5.3-CodeX and stronger models.\nmetadata:\n  author: \"TileGym Team \u003CTileGym@nvidia.com>\"\n  version: \"2026.06.03\"\n  tags:\n    - tilegym\n    - transformers\n    - integration\n    - kernel\n    - monkey-patch\n---\n# Integrate and create cuTile kernels into 🤗 Transformers\nThe main purpose of TileGym project is to provide performant kernels for LLM training and inference. We will integrate proper kernels available in TileGym project to LLM models provided by Hugging Face `transformers` library to validate end-to-end functional correctness and performance improvements. Instead of modifying `transformers` source code, we will take a non-intrusive monkey-patch approach: We will replace certain modules\u002Fclasses\u002Fmethods in `transformers` library that implement the Transformer model we would like to integrate, such that at model instantiation, that model's core components will be replaced by TileGym implementations. At runtime the model will actually invoke TileGym kernels under the hood. In addition, we will follow an auto-research-style agent harness loop to create and integrate new cuTile kernels to the target model to improve kernel coverage and end-to-end throughput.\n\n## Instructions\nThis is for human readers: Simply prompt your favorite AI Agent with skill name and target model ID. E.g.,:\n```Claude\u002FCodeX\nHi, please \u002Fmonkey-patch-kernels-to-transformers Qwen\u002FQwen3.5-0.8B.\n```\nThe Agent might ask you several questions. Make clarifications and give a go confirmation.\n\n## Workflow\n1. Prepare experiment environment. Follow [environment-setup.md](.\u002Freferences\u002Fenvironment-setup.md)\n2. Integrate existing TileGym kernels to the target model. Follow [kernel-integration.md](.\u002Freferences\u002Fkernel-integration.md)\n3. Autonomously create new cuTile kernels for uncovered PyTorch code. Follow [auto-kernelize.md](.\u002Freferences\u002Fauto-kernelize.md)\n   * Feel free to add new cuTile kernels with constraints in mind\n   * Do not stop until meet auto-kernelize loop stop conditions\n4. Summarize and report\n\n## Disciplines\nThis is for AI Agents executing this workflow.\n\n### Kernel inventory\nReusable transformer-local kernels must be represented with FlashInfer-Bench-style Definition and Solution metadata. Follow [kernel-inventory-schema.md](.\u002Freferences\u002Fkernel-inventory-schema.md) when researching compute requirements, inventorying existing kernels, proposing candidates, or creating new generated kernels.\n",{"data":35,"body":46},{"name":4,"description":6,"license":26,"compatibility":36,"metadata":37},"Verified on Claude Code with Opus-4.6 and onward, CodeX with GPT-5.5 and onward, and Cursor (Agent mode) with GPT-5.3-CodeX and stronger models.",{"author":38,"version":39,"tags":40},"TileGym Team \u003CTileGym@nvidia.com>","2026.06.03",[41,42,43,44,45],"tilegym","transformers","integration","kernel","monkey-patch",{"type":47,"children":48},"root",[49,58,86,93,98,119,124,130,188,194,199,206,219],{"type":50,"tag":51,"props":52,"children":54},"element","h1",{"id":53},"integrate-and-create-cutile-kernels-into-transformers",[55],{"type":56,"value":57},"text","Integrate and create cuTile kernels into 🤗 Transformers",{"type":50,"tag":59,"props":60,"children":61},"p",{},[62,64,70,72,77,79,84],{"type":56,"value":63},"The main purpose of TileGym project is to provide performant kernels for LLM training and inference. We will integrate proper kernels available in TileGym project to LLM models provided by Hugging Face ",{"type":50,"tag":65,"props":66,"children":68},"code",{"className":67},[],[69],{"type":56,"value":42},{"type":56,"value":71}," library to validate end-to-end functional correctness and performance improvements. Instead of modifying ",{"type":50,"tag":65,"props":73,"children":75},{"className":74},[],[76],{"type":56,"value":42},{"type":56,"value":78}," source code, we will take a non-intrusive monkey-patch approach: We will replace certain modules\u002Fclasses\u002Fmethods in ",{"type":50,"tag":65,"props":80,"children":82},{"className":81},[],[83],{"type":56,"value":42},{"type":56,"value":85}," library that implement the Transformer model we would like to integrate, such that at model instantiation, that model's core components will be replaced by TileGym implementations. At runtime the model will actually invoke TileGym kernels under the hood. In addition, we will follow an auto-research-style agent harness loop to create and integrate new cuTile kernels to the target model to improve kernel coverage and end-to-end throughput.",{"type":50,"tag":87,"props":88,"children":90},"h2",{"id":89},"instructions",[91],{"type":56,"value":92},"Instructions",{"type":50,"tag":59,"props":94,"children":95},{},[96],{"type":56,"value":97},"This is for human readers: Simply prompt your favorite AI Agent with skill name and target model ID. E.g.,:",{"type":50,"tag":99,"props":100,"children":105},"pre",{"className":101,"code":102,"language":103,"meta":104,"style":104},"language-Claude\u002FCodeX shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","Hi, please \u002Fmonkey-patch-kernels-to-transformers Qwen\u002FQwen3.5-0.8B.\n","Claude\u002FCodeX","",[106],{"type":50,"tag":65,"props":107,"children":108},{"__ignoreMap":104},[109],{"type":50,"tag":110,"props":111,"children":114},"span",{"class":112,"line":113},"line",1,[115],{"type":50,"tag":110,"props":116,"children":117},{},[118],{"type":56,"value":102},{"type":50,"tag":59,"props":120,"children":121},{},[122],{"type":56,"value":123},"The Agent might ask you several questions. Make clarifications and give a go confirmation.",{"type":50,"tag":87,"props":125,"children":127},{"id":126},"workflow",[128],{"type":56,"value":129},"Workflow",{"type":50,"tag":131,"props":132,"children":133},"ol",{},[134,147,158,183],{"type":50,"tag":135,"props":136,"children":137},"li",{},[138,140],{"type":56,"value":139},"Prepare experiment environment. Follow ",{"type":50,"tag":141,"props":142,"children":144},"a",{"href":143},".\u002Freferences\u002Fenvironment-setup.md",[145],{"type":56,"value":146},"environment-setup.md",{"type":50,"tag":135,"props":148,"children":149},{},[150,152],{"type":56,"value":151},"Integrate existing TileGym kernels to the target model. Follow ",{"type":50,"tag":141,"props":153,"children":155},{"href":154},".\u002Freferences\u002Fkernel-integration.md",[156],{"type":56,"value":157},"kernel-integration.md",{"type":50,"tag":135,"props":159,"children":160},{},[161,163,169],{"type":56,"value":162},"Autonomously create new cuTile kernels for uncovered PyTorch code. Follow ",{"type":50,"tag":141,"props":164,"children":166},{"href":165},".\u002Freferences\u002Fauto-kernelize.md",[167],{"type":56,"value":168},"auto-kernelize.md",{"type":50,"tag":170,"props":171,"children":172},"ul",{},[173,178],{"type":50,"tag":135,"props":174,"children":175},{},[176],{"type":56,"value":177},"Feel free to add new cuTile kernels with constraints in mind",{"type":50,"tag":135,"props":179,"children":180},{},[181],{"type":56,"value":182},"Do not stop until meet auto-kernelize loop stop conditions",{"type":50,"tag":135,"props":184,"children":185},{},[186],{"type":56,"value":187},"Summarize and report",{"type":50,"tag":87,"props":189,"children":191},{"id":190},"disciplines",[192],{"type":56,"value":193},"Disciplines",{"type":50,"tag":59,"props":195,"children":196},{},[197],{"type":56,"value":198},"This is for AI Agents executing this workflow.",{"type":50,"tag":200,"props":201,"children":203},"h3",{"id":202},"kernel-inventory",[204],{"type":56,"value":205},"Kernel inventory",{"type":50,"tag":59,"props":207,"children":208},{},[209,211,217],{"type":56,"value":210},"Reusable transformer-local kernels must be represented with FlashInfer-Bench-style Definition and Solution metadata. Follow ",{"type":50,"tag":141,"props":212,"children":214},{"href":213},".\u002Freferences\u002Fkernel-inventory-schema.md",[215],{"type":56,"value":216},"kernel-inventory-schema.md",{"type":56,"value":218}," when researching compute requirements, inventorying existing kernels, proposing candidates, or creating new generated kernels.",{"type":50,"tag":220,"props":221,"children":222},"style",{},[223],{"type":56,"value":224},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":226,"total":328},[227,242,256,270,282,299,314],{"slug":228,"name":228,"fn":229,"description":230,"org":231,"tags":232,"stars":23,"repoUrl":24,"updatedAt":241},"accelerated-computing-cudf","accelerate data processing with cuDF","Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV\u002FParquet I\u002FO, nullable semantics, and multi-GPU DataFrame workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[233,236,239,240],{"name":234,"slug":235,"type":15},"Data Analysis","data-analysis",{"name":237,"slug":238,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},"2026-07-14T05:28:43.176466",{"slug":243,"name":243,"fn":244,"description":245,"org":246,"tags":247,"stars":23,"repoUrl":24,"updatedAt":255},"aiq-deploy","deploy and manage NVIDIA AI-Q infrastructure","Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[248,251,254],{"name":249,"slug":250,"type":15},"Deployment","deployment",{"name":252,"slug":253,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":257,"name":257,"fn":258,"description":259,"org":260,"tags":261,"stars":23,"repoUrl":24,"updatedAt":269},"aiq-research","conduct deep research with AI-Q","Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[262,265,266],{"name":263,"slug":264,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":267,"slug":268,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":271,"name":271,"fn":272,"description":273,"org":274,"tags":275,"stars":23,"repoUrl":24,"updatedAt":281},"amc-run-sample-calibration","run AMC sample dataset calibration","Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[276,277,278],{"name":234,"slug":235,"type":15},{"name":9,"slug":8,"type":15},{"name":279,"slug":280,"type":15},"Testing","testing","2026-07-17T05:29:03.913266",{"slug":283,"name":283,"fn":284,"description":285,"org":286,"tags":287,"stars":23,"repoUrl":24,"updatedAt":298},"amc-run-video-calibration","calibrate video datasets with AutoMagicCalib","Calibrate a new dataset from pre-recorded video files via the AutoMagicCalib REST API. Use when user has local MP4s and says 'calibrate my videos', 'run AMC on these videos', or similar. For RTSP\u002Flive streams, use amc-run-rtsp-calibration instead.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[288,291,294,295],{"name":289,"slug":290,"type":15},"Automation","automation",{"name":292,"slug":293,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":296,"slug":297,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":300,"name":300,"fn":301,"description":302,"org":303,"tags":304,"stars":23,"repoUrl":24,"updatedAt":313},"amc-setup-calibration-stack","deploy AutoMagicCalib microservice with Docker","Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[305,306,309,310],{"name":249,"slug":250,"type":15},{"name":307,"slug":308,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":311,"slug":312,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":315,"name":315,"fn":316,"description":317,"org":318,"tags":319,"stars":23,"repoUrl":24,"updatedAt":327},"cudaq-guide","develop quantum applications with CUDA-Q","CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[320,321,324],{"name":9,"slug":8,"type":15},{"name":322,"slug":323,"type":15},"Quantum Computing","quantum-computing",{"name":325,"slug":326,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":330,"total":481},[331,349,365,376,388,402,415,429,440,449,463,472],{"slug":332,"name":332,"fn":333,"description":334,"org":335,"tags":336,"stars":346,"repoUrl":347,"updatedAt":348},"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},[337,340,343],{"name":338,"slug":339,"type":15},"Documentation","documentation",{"name":341,"slug":342,"type":15},"MCP","mcp",{"name":344,"slug":345,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":350,"name":350,"fn":351,"description":352,"org":353,"tags":354,"stars":362,"repoUrl":363,"updatedAt":364},"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},[355,358,359],{"name":356,"slug":357,"type":15},"Containers","containers",{"name":249,"slug":250,"type":15},{"name":360,"slug":361,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":366,"name":366,"fn":367,"description":368,"org":369,"tags":370,"stars":362,"repoUrl":363,"updatedAt":375},"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},[371,374],{"name":372,"slug":373,"type":15},"CI\u002FCD","ci-cd",{"name":249,"slug":250,"type":15},"2026-07-14T05:25:59.97109",{"slug":377,"name":377,"fn":378,"description":379,"org":380,"tags":381,"stars":362,"repoUrl":363,"updatedAt":387},"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},[382,383,384],{"name":372,"slug":373,"type":15},{"name":249,"slug":250,"type":15},{"name":385,"slug":386,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":389,"name":389,"fn":390,"description":391,"org":392,"tags":393,"stars":362,"repoUrl":363,"updatedAt":401},"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},[394,397,398],{"name":395,"slug":396,"type":15},"Debugging","debugging",{"name":385,"slug":386,"type":15},{"name":399,"slug":400,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":403,"name":403,"fn":404,"description":405,"org":406,"tags":407,"stars":362,"repoUrl":363,"updatedAt":414},"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},[408,411],{"name":409,"slug":410,"type":15},"Best Practices","best-practices",{"name":412,"slug":413,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":416,"name":416,"fn":417,"description":418,"org":419,"tags":420,"stars":362,"repoUrl":363,"updatedAt":428},"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},[421,424,427],{"name":422,"slug":423,"type":15},"Machine Learning","machine-learning",{"name":425,"slug":426,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":430,"name":430,"fn":431,"description":432,"org":433,"tags":434,"stars":362,"repoUrl":363,"updatedAt":439},"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},[435,438],{"name":436,"slug":437,"type":15},"QA","qa",{"name":279,"slug":280,"type":15},"2026-07-14T05:25:53.673039",{"slug":441,"name":441,"fn":442,"description":443,"org":444,"tags":445,"stars":362,"repoUrl":363,"updatedAt":448},"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},[446,447],{"name":249,"slug":250,"type":15},{"name":252,"slug":253,"type":15},"2026-07-14T05:25:49.362534",{"slug":450,"name":450,"fn":451,"description":452,"org":453,"tags":454,"stars":362,"repoUrl":363,"updatedAt":462},"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},[455,458,459],{"name":456,"slug":457,"type":15},"Code Review","code-review",{"name":385,"slug":386,"type":15},{"name":460,"slug":461,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":464,"name":464,"fn":465,"description":466,"org":467,"tags":468,"stars":362,"repoUrl":363,"updatedAt":471},"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},[469,470],{"name":436,"slug":437,"type":15},{"name":279,"slug":280,"type":15},"2026-07-14T05:25:54.928983",{"slug":473,"name":473,"fn":474,"description":475,"org":476,"tags":477,"stars":362,"repoUrl":363,"updatedAt":480},"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},[478,479],{"name":289,"slug":290,"type":15},{"name":372,"slug":373,"type":15},"2026-07-30T05:29:03.275638",496]