[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-78455":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":8,"htmlUrl":8,"language":9,"languages":8,"totalLinesOfCode":8,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":14,"forks30d":14,"starsTrendScore":18,"compositeScore":19,"rankGlobal":8,"rankLanguage":8,"license":8,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":20,"topics":23,"createdAt":8,"pushedAt":8,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":14,"starSnapshotCount":14,"syncStatus":27,"lastSyncTime":28,"discoverSource":29},78455,"kernel-design-agents","mit-han-lab\u002Fkernel-design-agents","mit-han-lab",null,"Python",549,46,3,1,0,6,97,430,45,9.02,false,"main",true,[],"2026-06-12 02:03:47","# Kernel Design Agents\n\nKernel Design Agents (KDA) is a agent-centric workflow for using coding agents to research, implement, verify, and iterate on performance-sensitive CUDA kernel tasks.\n\nThis repository documents the early research prototype and is still under active development (we are looking for community feedbacks!). If you are interested in HAN Lab Mafia's  solution ranking #1~3 on tracks at MLSys Kernel Contest, please refer to [mit-han-lab\u002Fmlsys2026-flashinfer-contest](https:\u002F\u002Fgithub.com\u002Fmit-han-lab\u002Fmlsys2026-flashinfer-contest) for perform evaluation and reproducement.\n\n## Contents\n\n| Path | Purpose |\n|---|---|\n| `docs\u002Fagent-flow.md` | Minimal end-to-end KDA workflow. |\n| `prompts\u002FREADME.md` | How to use prompt templates. |\n| `prompts\u002Fbasic-flow.md` | Generic starter prompt for a new task. |\n| `CLAUDE.md` | Repository-facing agent instructions. |\n\n## Getting Started\nInstall the agent workflow dependencies before starting the agent session:\n\n```bash\ngit clone --recurse-submodules https:\u002F\u002Fgithub.com\u002Fmit-han-lab\u002Fkernel-design-agents.git\ncd kernel-design-agents\n\n# link skills\nmkdir -p ~\u002F.claude\u002Fskills\nln -s \"$(pwd)\u002Fskills\u002Fncu-report-skill\" ~\u002F.claude\u002Fskills\u002Fncu-report-skill\nln -s \"$(pwd)\u002Fskills\u002FKernelWiki\" ~\u002F.claude\u002Fskills\u002FKernelWiki\n\n# or clone skills directly\nmkdir -p ~\u002F.claude\u002Fskills && cd ~\u002F.claude\u002Fskills\ngit clone https:\u002F\u002Fgithub.com\u002Fmit-han-lab\u002Fncu-report-skill.git\ngit clone https:\u002F\u002Fgithub.com\u002Fmit-han-lab\u002FKernelWiki.git\n```\n\nInstall the `humanize` Claude Code plugin from the Claude Code plugin UI:\n\n```text\n\u002Fplugin marketplace add PolyArch\u002Fhumanize\n\u002Fplugin install humanize@PolyArch\n```\n\n## Minimal Flow\n\n1. Create a separate implementation workspace for the target task.\n2. Define the task contract: objective, constraints, validation command, and promotion criteria.\n3. Start an agent session in the implementation workspace.\n4. Give the agent `prompts\u002Fbasic-flow.md`, filled in with the task-specific details.\n5. Ask the agent to write a short plan draft to `docs\u002Fdraft.md` in the implementation workspace.\n6. Convert the draft into an executable plan, either manually or with a planning tool such as Humanize.\n7. Implement in small iterations, verifying after each meaningful change.\n8. Record candidates, benchmark or evaluation results, profiling evidence, and final promotion decisions.\n\nThe workflow is intentionally independent of any single benchmark harness or hardware target. A downstream task can add its own evaluator, datasets, profiling tools, and domain-specific references.\n\n## Recommended Workspace Layout\n\nUse this repository as reference material, then do implementation work elsewhere:\n\n```text\ntask-workspace\u002F\n  docs\u002F\n    draft.md\n    plan.md\n  runs\u002F\n  outputs\u002F\n  profile\u002F\n  benchmark.csv\n  candidates.jsonl\n```\n\nThe exact files can change by domain. The important rule is that the agent records enough context for another engineer to understand what was tried, what passed validation, and why the final candidate was selected.\n\n","Kernel Design Agents (KDA) 是一个以代理为中心的工作流程，用于通过编码代理来研究、实现、验证和迭代性能敏感的CUDA内核任务。其核心功能包括使用预定义的提示模板启动新任务，以及利用特定技能如ncu-report-skill和KernelWiki增强代理能力，支持在不同硬件目标上进行独立于具体基准测试工具的任务执行。此外，该项目推荐了一种工作区布局方式，便于记录每次迭代的过程与结果，确保其他工程师能够理解并复现整个开发过程。KDA适用于需要高效优化GPU计算性能的研究场景或工业应用中。",2,"2026-06-11 03:56:51","CREATED_QUERY"]