[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-prompt-engineering":3},{"tag":4,"articles":11},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":10},"226e7da3-c3ac-432a-9d56-30f18876fce1","prompt engineering","prompt-engineering",16,"Prompt engineering 已從「把話說好」變成 AI 工作流程的一部分，涵蓋標準化提示、結構化輸出、代理迴圈、長上下文與治理需求。對開發者來說，它直接影響錯誤率、token 成本與可審核性。","Prompt engineering is now part of the AI stack, not just wording. It covers shared prompt standards, structured outputs, agent loops, long-context handling, and governance concerns that affect error rates, token cost, and auditability in production.",[12,21,29,37,44,51,58,65,72,80,87],{"id":13,"slug":14,"title":15,"summary":16,"category":17,"image_url":18,"cover_image":18,"language":19,"created_at":20},"cdf45047-63fd-4393-b5a9-f2dbefe7738c","prompt-engineering-jobs-2026-worth-it-zh","2026 還值得學 Prompt Engineering 嗎","2026 年的 prompt engineering 不是消失，而是融進 AI 產品、工程和營運職位。","industry","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778537464048-diid.png","zh","2026-05-11T22:10:36.504242+00:00",{"id":22,"slug":23,"title":24,"summary":25,"category":26,"image_url":27,"cover_image":27,"language":19,"created_at":28},"8a126231-b5e1-4c89-a338-73d1d73148ff","why-prompt-engineering-is-dead-for-ai-agents-zh","為什麼 AI Agent 時代，Prompt Engineering 已經死了","AI Agent 的關鍵不在於把提示詞寫得更漂亮，而在於把上下文選對、排好、壓縮好；context engineering 才是可靠性的核心。","ai-agent","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778184658982-xgnl.png","2026-05-07T20:10:24.866536+00:00",{"id":30,"slug":31,"title":32,"summary":33,"category":34,"image_url":35,"cover_image":35,"language":19,"created_at":36},"eeeff79e-4789-40ce-a55d-dba97d54ada2","why-rag-needs-self-healing-layer-zh","為什麼 RAG 需要自癒層，而不只是更好的提示詞","RAG 應被視為會失敗的系統，真正該補的是即時自癒層，而不是繼續迷信提示詞調校。","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778098242230-wbbc.png","2026-05-06T20:10:22.158933+00:00",{"id":38,"slug":39,"title":40,"summary":41,"category":34,"image_url":42,"cover_image":42,"language":19,"created_at":43},"6b0e0e51-acc2-46dc-a376-6b7fc78f7918","prompt-engineering-becoming-infrastructure-zh","Prompt 工程正在變成基礎設施","Springer 新章節指出，Prompt engineering 已不只是寫得巧，而是牽涉倫理、治理與領域知識的系統工作。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1776742218129-fisd.png","2026-04-21T00:15:42.239955+00:00",{"id":45,"slug":46,"title":47,"summary":48,"category":34,"image_url":49,"cover_image":49,"language":19,"created_at":50},"b82d0062-8cef-4869-8a1e-e7a314d24478","why-prompt-standards-matter-for-ai-work-zh","AI 工作為何需要提示標準","Springer 新章節主張提示工程需要共通標準，才能減少 Token 浪費、降低錯誤，並讓 AI 工作更好追蹤與審核。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1776738622916-xipq.png","2026-04-21T00:12:38.635108+00:00",{"id":52,"slug":53,"title":54,"summary":55,"category":17,"image_url":56,"cover_image":56,"language":19,"created_at":57},"fcc8d167-dc0f-4514-8d6b-4f4230547616","prompt-to-harness-ai-engineering-shift-zh","從 Prompt 到 Harness 工程","OpenAI 透露，一個 3 人團隊用 Codex、5 個月，合併約 1,500 個 PR，做出超過 100 萬行程式碼的產品。重點不在寫 prompt，而是怎麼設計讓 AI 能穩定工作的 harness。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775630042932-7zhw.png","2026-04-08T06:33:43.724227+00:00",{"id":59,"slug":60,"title":61,"summary":62,"category":17,"image_url":63,"cover_image":63,"language":19,"created_at":64},"a0660205-5b41-49a6-8119-ee9105a7e1f5","chatgpt-ads-format-standardization-data-zh","ChatGPT 廣告越來越一致","40,000 筆廣告版位分析顯示，ChatGPT 廣告正變得更短、更直白、更標準化。這反映 OpenAI 在優化轉換，也透露 LLM 使用習慣正在往任務導向收斂。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775218190861-p9x8.png","2026-04-03T12:09:37.164139+00:00",{"id":66,"slug":67,"title":68,"summary":69,"category":26,"image_url":70,"cover_image":70,"language":19,"created_at":71},"f8c44ca5-e1b5-4b51-a7e5-61cdf8fa5ab9","prompt-engineering-agents-structured-outputs-zh","Agent 與結構化輸出提示詞實戰","LLM 進到生產環境後，提示詞不再是寫得漂亮就好。這篇拆解推理、長上下文、JSON 合約與 agent 迴圈，講清楚怎麼把 GPT、Claude 和本地模型用得更穩。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775164928194-j63i.png","2026-04-02T21:21:45.59991+00:00",{"id":73,"slug":74,"title":75,"summary":76,"category":77,"image_url":78,"cover_image":78,"language":19,"created_at":79},"13819f2d-e9a1-4af2-88f3-7dbe4cb4ce61","prompt-engineering-explained-without-the-hype-zh","別把 Prompt Engineering 想太神","Prompt engineering 不是玄學。AWS 直接把方法、用途和取捨講清楚，重點是把模糊需求變成可用輸出，讓 LLM 更穩、更好控。","tools","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775164736525-0uq1.png","2026-04-02T21:18:35.227524+00:00",{"id":81,"slug":82,"title":83,"summary":84,"category":34,"image_url":85,"cover_image":85,"language":19,"created_at":86},"87335969-ee48-4021-bd27-6731750537ff","duplicate-prompts-can-lift-accuracy-fast-zh","重複提示詞，準確率真的會上升","Google Research 研究發現，提示詞複製一次可讓 70 組模型與基準測試中的 47 組提升準確率，NameIndex 甚至從 21.33% 衝到 97.33%。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775122500397-vvmh.png","2026-04-02T08:39:34.363421+00:00",{"id":88,"slug":89,"title":90,"summary":91,"category":17,"image_url":92,"cover_image":92,"language":19,"created_at":93},"418192da-88ae-4126-b586-bf079402c91e","prompt-engineering-2026-skill-tool-or-job-zh","2026 提示工程：技能、工具，還是工作？","提示工程還有用，但已經不是單靠幾句 Prompt 就能拿高薪的神話。從工具鏈、評估流程到職缺變化，2026 年更像是把 Prompt 納進產品、軟體與營運流程的一項實用技能。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1774517431347-1wth.png","2026-03-26T08:10:10.548201+00:00"]