Nacos 3.3.0 adds task-based AI resource discovery
Nacos 3.3.0 adds Agentic Resource Discovery (ARD), allowing agents to search an existing Nacos registry for MCP servers, Skills, Prompts, and other agents relevant to a task. Juejin · 人工智能 provides configuration steps and a demonstration in which Codex uses ARD and AgentFinder to locate sales data and reporting resources; the figures shown are a demo, not a verified customer result. The report also notes that ARD remains a proposal-stage specification and requires teams to provide and maintain the underlying services, rules, permissions, and optional vector-search components.
ARD gives agents a shared discovery layer for reusing governed MCP services, Skills, Prompts, and agent resources, while Nacos can retain responsibility for publishing and versioning those resources.
作者:席翁
如今,工作中用 Agent 帮忙已经越来越常见。研发同学用 Coding Agent 写代码、补测试、排查问题,业务同学用 Work Agent 整理资料、分析表格、制作演示文稿。不少过去需要自己一步步操作的事情,现在可以把需求交给 Agent,让它协助完成。
月底的经营汇报就是一个例子:看看华东区这个月卖得怎么样,和上个月比有哪些变化,再整理成几页 PPT。这类任务很适合让 Agent 帮忙。
不过,在 Agent 开始干活之前,员工往往还得自己做不少准备。
一份经营汇报背后的准备工作
假设 Agent 还没有接入公司的销售系统。你得先找数据团队,说明要查华东区哪两个月、需要哪些指标,请他们导出 Excel 或 CSV。拿到文件后,再确认字段含义、检查数据有没有缺漏,才能交给 Agent。
#Nacos 3.3.0#Agentic Resource Discovery#Nacos AI Registry#AgentFinder Skill#Codex