[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-kiro-aws-healthomics-bioinformatics-workflow-zh":3,"article-related-kiro-aws-healthomics-bioinformatics-workflow-zh":28,"series-tools-4cbaa31a-f971-4bb3-8607-96d12e6548aa":85},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":11,"views":25,"created_at":26,"published_at":27,"topic_cluster_id":11},"4cbaa31a-f971-4bb3-8607-96d12e6548aa","kiro-aws-healthomics-bioinformatics-workflow-zh","Kiro 搭 HealthOmics，生資流程少卡關","\u003Cp>生資 workflow 很磨人。你要懂 biology，也要會 WDL、Nextflow，還得顧雲端設定。AWS 這次把 \u003Ca href=\"https:\u002F\u002Fkiro.dev\" target=\"_blank\" rel=\"noopener\">Kiro\u003C\u002Fa> 接上 \u003Ca href=\"https:\u002F\u002Faws.amazon.com\u002Fhealthomics\u002F\" target=\"_blank\" rel=\"noopener\">AWS HealthOmics\u003C\u002Fa>，直接把痛點搬進 IDE。AWS 說，workflow 建置和 migration 的速度，能拉到 2 倍以上。還有一個 RNA-seq migration，從數天壓到半天內。\u003C\u002Fp>\u003Cp>講白了，這不是在比誰 p\u003Ca href=\"\u002Fnews\u002Fprompt-engineering-explained-without-the-hype-zh\">romp\u003C\u002Fa>t 寫得帥。這是在比誰少踩坑。生資工程最貴的，常常不是寫 code，而是等跑完才發現 container 壞了、directive 不支援，或參數根本不合規。\u003C\u002Fp>\u003Cp>這次 AWS 的做法很直白。它想把平台知識塞進工具裡。讓你不用每次都重新教 AI 一遍 HealthOmics 規則。這點我覺得蠻實際的。\u003C\u002Fp>\u003Ch2>AWS 這次到底做了什麼\u003C\u002Fh2>\u003Cp>核心是 \u003Ca href=\"https:\u002F\u002Faws.amazon.com\u002Fblogs\u002Findustries\u002Ffrom-prompt-to-pipeline-ai-powered-bioinformatics-workflow-development-with-kiro-and-aws-healthomics\u002F\" target=\"_blank\" rel=\"noopener\">AWS HealthOmics extension for Kiro\u003C\u002Fa>，再加上 AWS HealthOmics Kiro Power。兩個東西合起來，讓 Kiro 多了 HealthOmics 的語境。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775167427064-u1q3.png\" alt=\"Kiro 搭 HealthOmics，生資流程少卡關\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>這代表 IDE 不再只會補字。它還能幫你看部署、驗證、除錯，甚至更新 workflow。對一般 app 開發者來說，這很像把 lint、deploy、runbook 全塞回編輯器。\u003C\u002Fp>\u003Cp>對生資團隊來說，這種整合很重要。因為一個錯誤常常不是小 bug，而是幾小時 compute 白燒。你不想等到跑完才知道 container image 不對。你想在寫的時候就被攔下來。\u003C\u002Fp>\u003Cul>\u003Cli>支援 \u003Ca href=\"https:\u002F\u002Fwww.nextflow.io\u002F\" target=\"_blank\" rel=\"noopener\">Nextflow\u003C\u002Fa> 和 \u003Ca href=\"https:\u002F\u002Fopenwdl.org\u002F\" target=\"_blank\" rel=\"noopener\">WDL\u003C\u002Fa>\u003C\u002Fli>\u003Cli>提供 IntelliSense 和即時診斷\u003C\u002Fli>\u003Cli>內建 HealthOmics Explorer，可看 workflow 和 run\u003C\u002Fli>\u003Cli>可先檢查不支援的 directive 與 container 格式\u003C\u002Fli>\u003Cli>透過 \u003Ca href=\"https:\u002F\u002Fmodelcontextprotocol.io\u002F\" target=\"_blank\" rel=\"noopener\">Model Context Protocol\u003C\u002Fa> 做自然語言操作\u003C\u002Fli>\u003C\u002Ful>\u003Cp>這裡的重點不是 AI 很會聊。重點是它知道你在 HealthOmics 裡做事。這種 context 才值錢。沒有 context 的 AI，常常只是在猜。\u003C\u002Fp>\u003Cp>而且 AWS 沒叫你改用自家新語言。它還是吃 Nextflow 和 WDL。這很務實。因為生資團隊早就有既有 pi\u003Ca href=\"\u002Fnews\u002Fopenai-vs-deepmind-models-apps-2026-zh\">pe\u003C\u002Fa>line，不可能為了 AI 再重寫一輪。\u003C\u002Fp>\u003Ch2>為什麼 MCP 這層很關鍵\u003C\u002Fh2>\u003Cp>MCP 是這次最有意思的地方。因為它決定 AI 回答的品質。沒有 domain context，模型可以寫 code，但常常漏掉平台限制。對 HealthOmics 來說，漏一條規則就可能整個部署失敗。\u003C\u002Fp>\u003Cp>AWS 說，Kiro Power 會自動設定 HealthOmics MCP server，還會加上 steering guides。這些 guide 會管第一次設定、spec-driven 開發、跨平台 migration。也就是說，Kiro 不只會寫，還知道怎麼包、怎麼 deploy、怎麼查錯。\u003C\u002Fp>\u003Cp>這種設計很像把資深同事的腦袋，做成工作區設定。你不用每次問「這個 container 可以嗎」。工具自己先幫你擋一次。這對團隊協作也有幫助，因為規則會比較一致。\u003C\u002Fp>\u003Cblockquote>“You can ask Kiro to create a totally new workflow definition from only a natural language description,” AWS wrote in the blog post announcing the extension and power.\u003C\u002Fblockquote>\u003Cp>這句話很直球。AWS 想把起點放在自然語言，再一路走到 spec、部署、執行。流程不是聊天而已。它是把聊天接到可跑的 pipeline。\u003C\u002Fp>\u003Cp>我覺得這種模式比空泛的 AI demo 實在。因為它把問題縮小了。不是問 AI 會不會萬用。是問它懂不懂這個系統。\u003C\u002Fp>\u003Ch2>跟舊流程比，差在哪\u003C\u002Fh2>\u003Cp>傳統生資流程很像打地鼠。你改一版，跑一次，等很久，然後爆掉。再改，再跑，再等。這種迴圈很耗人。也很耗雲端費用。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775167434634-kugb.png\" alt=\"Kiro 搭 HealthOmics，生資流程少卡關\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>AWS 在文章裡給了幾個數字。這些數字比形容詞有用多了。因為你可以直接拿來估工時，也能拿來算成本。\u003C\u002Fp>\u003Cp>它的說法很簡單：把錯誤提早抓出來，把重複工作交給工具。這樣一來，很多本來要靠手動檢查的步驟，就能在 editor 裡先過濾掉。\u003C\u002Fp>\u003Cul>\u003Cli>一般 workflow 建置與 migration：速度超過 2 倍\u003C\u002Fli>\u003Cli>一個 RNA-seq migration：從數天縮到半天內\u003C\u002Fli>\u003Cli>失敗診斷：從 run 後才知道，改成寫的時候先檢查\u003C\u002Fli>\u003Cli>部署方式：直接從 IDE 送出，不用一直切 console\u003C\u002Fli>\u003C\u002Ful>\u003Cp>這種差異很像一般軟體開發從命令列，走到完整 IDE 的過程。差別是，生資的錯誤代價更高。你不是只浪費時間。你還可能浪費樣本、算力，甚至研究排程。\u003C\u002Fp>\u003Cp>如果拿競品來看，\u003Ca href=\"https:\u002F\u002Fwww.nextflow.io\u002F\" target=\"_blank\" rel=\"noopener\">Nextflow\u003C\u002Fa> 本身已經很成熟。\u003Ca href=\"https:\u002F\u002Fopenwdl.org\u002F\" target=\"_blank\" rel=\"noopener\">WDL\u003C\u002Fa> 也有穩定社群。AWS 的差別，不在語言本身，而在整合層。\u003C\u002Fp>\u003Cp>它想做的是把 workflow 語言、雲端資源、AI 助手綁在一起。這跟單純賣一個 chatbot 不一樣。它比較像工作台。\u003C\u002Fp>\u003Ch2>這件事對 AI 工具的意思\u003C\u002Fh2>\u003Cp>這次更新其實在講一件事。AI coding 工具要有領域知識，才真的有用。通用模型很強，但不代表懂每個雲端服務的規則。\u003C\u002Fp>\u003Cp>在科學軟體裡，這件事更明顯。你寫得對，不代表能跑。你跑得動，不代表合規。你合規，也不代表省錢。這些條件常常同時存在。\u003C\u002Fp>\u003Cp>所以 HealthOmics extension 的價值，不是它會聊天。是它知道 HealthOmics 的邊界。知道什麼可以做，什麼不行。這種知識，才是生資團隊真正需要的。\u003C\u002Fp>\u003Cp>我也覺得這會影響其他領域。像 lab automation、影像 pipeline、金融風控流程，都很吃 domain context。未來如果 AI 工具只會泛用回答，競爭力會很有限。\u003C\u002Fp>\u003Cp>真正有用的工具，會把模型包進工作流程。不是叫你換腦袋，而是叫工具先學會你現在的規矩。\u003C\u002Fp>\u003Ch2>這波更新背後的產業脈絡\u003C\u002Fh2>\u003Cp>生資軟體本來就很碎。研究人員、資料工程師、雲端工程師，常常要一起配合。每個人看問題的角度都不同。這也是 workflow 開發一直難搞的原因。\u003C\u002Fp>\u003Cp>另外，雲端算力越來越貴。尤其是長時間跑的分析。像 RNA-seq、variant calling、meta\u003Ca href=\"\u002Fnews\u002Fprompt-engineering-agents-structured-outputs-zh\">gen\u003C\u002Fa>omics，任何一個節點出錯，都會拖掉整條鏈。這時候，早點發現問題，比事後修正重要很多。\u003C\u002Fp>\u003Cp>從產業角度看，AWS 這次是把 AI 助手往垂直場景推。不是再做一個萬用聊天介面，而是把它塞進既有平台。這種做法比較容易被企業買單。因為它能直接對到工時、錯誤率、和 cloud bill。\u003C\u002Fp>\u003Cp>如果你問我，這類工具會不會取代生資工程師？我覺得不會。它比較像把重複勞動壓縮掉。人還是要負責判斷。尤其是資料品質、實驗設計、和結果解讀，AI 還差很遠。\u003C\u002Fp>\u003Cp>但如果你問它會不會改變團隊寫 pipeline 的方式？我覺得會。至少從現在開始，大家會更期待 IDE 直接懂平台，而不是只會幫你補幾行 code。\u003C\u002Fp>\u003Ch2>接下來該看什麼\u003C\u002Fh2>\u003Cp>如果你已經在用 HealthOmics，我會先試 Kiro 和 extension。不要先想大改架構。先拿一條現有 pipeline 試水溫。看它能不能真的幫你抓錯、補 spec、縮短 migration。\u003C\u002Fp>\u003Cp>如果你是做一般軟體開發，也可以把這次當成一個訊號。AI 工具正在往「懂場景」走。下一個有用的助手，不一定是最會講話的那個。可能是最懂你專案規則的那個。\u003C\u002Fp>\u003Cp>我自己的判斷很直接。這類整合如果真的穩，接下來 12 個月，會有更多雲端服務學這一套。問題不是會不會跟上。問題是你的團隊準備好把流程標準化了沒。\u003C\u002Fp>\u003Cp>你如果是生資或資料平台團隊，現在就該問一句：我們的 pipeline，有沒有哪一段，其實可以先交給工具處理？\u003C\u002Fp>","AWS 把 Kiro 接上 HealthOmics，主打把生資 workflow 建置速度拉高 2 倍以上；一個 RNA-seq migration 甚至從數天縮到半天內。","aws.amazon.com","https:\u002F\u002Faws.amazon.com\u002Fblogs\u002Findustries\u002Ffrom-prompt-to-pipeline-ai-powered-bioinformatics-workflow-development-with-kiro-and-aws-healthomics\u002F",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775167427064-u1q3.png","tools","zh","9bf2d3e1-bb4a-4feb-9ad4-6cfcfdbfa5c5",[17,18,19,20,21,22,23,24],"Kiro","AWS HealthOmics","生物資訊","Nextflow","WDL","MCP","AI 工具","workflow",5,"2026-04-02T22:03:31.71144+00:00","2026-04-02T22:03:31.359+00:00",{"tags":29,"relatedLang":44,"relatedPosts":48},[30,32,34,36,38,39,41,43],{"name":23,"slug":31},"ai-工具",{"name":18,"slug":33},"aws-healthomics",{"name":22,"slug":35},"mcp",{"name":20,"slug":37},"nextflow",{"name":24,"slug":24},{"name":17,"slug":40},"kiro",{"name":21,"slug":42},"wdl",{"name":19,"slug":19},{"id":15,"slug":45,"title":46,"language":47},"kiro-aws-healthomics-bioinformatics-workflow-en","Kiro and AWS HealthOmics Cut Workflow Friction","en",[49,55,61,67,73,79],{"id":50,"slug":51,"title":52,"cover_image":53,"image_url":53,"created_at":54,"category":13},"d3ec03a8-a805-4a21-9826-72a74a72b625","databricks-model-serving-llm-deploy-guide-zh","Databricks Model Serving 讓 LLM 部署變簡單","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780525998117-7ur8.png","2026-06-03T22:32:51.005996+00:00",{"id":56,"slug":57,"title":58,"cover_image":59,"image_url":59,"created_at":60,"category":13},"4dd225a8-bf6c-4768-a486-a27956c7033d","opencode-digitalocean-model-freedom-zh","OpenCode+DigitalOcean 讓你切換模型","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780525116428-1q7g.png","2026-06-03T22:18:06.969758+00:00",{"id":62,"slug":63,"title":64,"cover_image":65,"image_url":65,"created_at":66,"category":13},"4bdcf208-fb80-484e-b4b6-06af035a6df1","modulate-aws-voice-chats-into-signals-zh","Modulate 用 AWS 把語音聊天做成訊號","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780519733892-rxue.png","2026-06-03T20:48:22.697917+00:00",{"id":68,"slug":69,"title":70,"cover_image":71,"image_url":71,"created_at":72,"category":13},"f44a28d3-2305-43de-b5fa-21217d561054","amazon-rekognition-content-moderation-filter-zh","Amazon Rekognition把審核變成過濾器","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780517005409-bxfc.png","2026-06-03T20:02:57.634353+00:00",{"id":74,"slug":75,"title":76,"cover_image":77,"image_url":77,"created_at":78,"category":13},"80f6f40b-3217-45e4-acff-7b2f6d261779","codex-workspace-limits-tell-you-why-zh","Codex 讓工作區限額錯誤說人話","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780514293711-ltqa.png","2026-06-03T19:17:41.340056+00:00",{"id":80,"slug":81,"title":82,"cover_image":83,"image_url":83,"created_at":84,"category":13},"daa3d568-4bc5-4f29-aa64-225928ace9b4","book-2-turns-sneaker-drop-into-merch-zh","Book 2 把球鞋發售變成周邊系統","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780513400116-8jeh.png","2026-06-03T19:02:49.03795+00:00",[86,91,96,101,106,111,116,121,126,131],{"id":87,"slug":88,"title":89,"created_at":90},"855cd52f-6fab-46cc-a7c1-42195e8a0de4","surepath-real-time-mcp-policy-controls-zh","SurePath 推出即時 MCP 政策控管","2026-03-26T07:57:40.77233+00:00",{"id":92,"slug":93,"title":94,"created_at":95},"9b19ab54-edef-4dbd-9ce4-a51e4bae4ebb","mcp-in-2026-the-ai-tool-layer-teams-use-zh","2026 年 MCP：團隊真的在用的 AI 工具層","2026-03-26T08:01:46.589694+00:00",{"id":97,"slug":98,"title":99,"created_at":100},"af9c46c3-7a28-410b-9f04-32b3de30a68c","prompting-in-2026-what-actually-works-zh","2026 提示工程，真正有用的是什麼","2026-03-26T08:08:12.453028+00:00",{"id":102,"slug":103,"title":104,"created_at":105},"05553086-6ed0-4758-81fd-6cab24b575e0","garry-tan-open-sources-claude-code-toolkit-zh","Garry Tan 開源 Claude Code 工具包","2026-03-26T08:26:20.068737+00:00",{"id":107,"slug":108,"title":109,"created_at":110},"042a73a2-18a2-433d-9e8f-9802b9559aac","github-ai-projects-to-watch-in-2026-zh","2026 必看 20 個 GitHub AI 專案","2026-03-26T08:28:09.619964+00:00",{"id":112,"slug":113,"title":114,"created_at":115},"a5f94120-ac0d-4483-9a8b-63590071ac6a","claude-code-vs-cursor-2026-zh","Claude Code 與 Cursor 深度對比：202…","2026-03-26T13:27:14.279193+00:00",{"id":117,"slug":118,"title":119,"created_at":120},"0975afa1-e0c7-4130-a20d-d890eaed995e","practical-github-guide-learning-ml-2026-zh","2026 機器學習入門 GitHub 實用指南","2026-03-27T01:16:49.712576+00:00",{"id":122,"slug":123,"title":124,"created_at":125},"bfdb467a-290f-4a80-b3a9-6f081afb6dff","aiml-2026-student-ai-ml-lab-repo-review-zh","AIML-2026：像課綱的學生實驗 Repo","2026-03-27T01:21:51.467798+00:00",{"id":127,"slug":128,"title":129,"created_at":130},"80cabc3e-09fc-4ff5-8f07-b8d68f5ae545","ai-trending-github-repos-and-research-feeds-zh","AI Trending：把 AI 資源收成一張表","2026-03-27T01:31:35.262183+00:00",{"id":132,"slug":133,"title":134,"created_at":135},"3ce6e6e2-bac5-463e-9f8d-45caabcc61f7","awesome-ai-for-science-research-tools-map-zh","AI 科研工具清單，開始像地圖了","2026-03-27T01:46:50.521945+00:00"]