rauchg 分享 Agent 全流程发代码工作流：功能开关+推送验证+指标监控
rauchg 演示了一套不用 localhost 的发版流程：Agent 写 feature-gated 代码，push 时自动过安全和证明检查，上线后盯指标。Marc Lou 补充了他一个月只在生产环境自测的做法，连查日志都让 Agent 定时提醒。写 AGENTS.md 的开发者可以直接抄这套规则。
Vercel CEO Guillermo Rauch 提出一种 Agent 交付代码的三段式流程：Agent 先写带 feature flag 的功能门控代码，推送阶段由验证门把关（包括 proof checking 和 security audits），上线后由 Agent 持续监控指标。他给出可用命令：vercel flags、vercel metrics、vercel deploy，全部在 Vercel CLI 中即可跑通整条链路。
开发者 Marc Lou 在引用中补充了自己的实践：已一个月没用 localhost，而是让 AI 把新功能只对生产环境中的自己可见，确认可用后再放给用户，并让 Agent 设置每小时/每日提醒去查日志和数据库。他的 AGENTS.md 中还写入了强制测试规则，使代码从 prompt 直接进入生产环境。
这条工作流的核心是把安全发布机制（feature flags、验证门、监控）交给 Agent 编排，开发者只做评估决策。对使用 Vercel 部署的独立开发者和团队来说，这相当于把 localhost 开发环节整个移除了。
This is how agents will ship all code. → Agent writes feature-gated code → Verification gates on push (proof checking, security audits) → Agent monitors metrics You can live the future today with Vercel CLI: 𝚟𝚎𝚛𝚌𝚎𝚕 𝚏𝚕𝚊𝚐𝚜, 𝚟𝚎𝚛𝚌𝚎𝚕 𝚖𝚎𝚝𝚛𝚒𝚌𝚜, 𝚟𝚎𝚛𝚌𝚎𝚕 𝚍𝚎𝚙𝚕𝚘𝚢. Marc Lou @marclou I haven't touched localhost in month. I ask AI to make the feature only visible to me in production, so I can assess and iterate. If it's good, I release it to users, and ask the agent to set hourly/daily reminders to check logs and database to fix any issues. My AGENTS·md also has rules to force the agent to test everything. Now AI is so good, everything goes straight from prompt to prod. 🔗 View Quoted Tweet📊 39 ⚡
This report is categorized under AITOP's "How-to" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Vercel, AGENTS.md, feature flags, agents.
Source attribution: curated from Guillermo Rauch (@rauchg).
