慣性聚合 高效追讀感興趣之博客、新聞、科技資訊
閱原文 以慣性聚合開啟

推薦訂閱源

让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
WordPress大学
WordPress大学
量子位
M
Microsoft Research Blog - Microsoft Research
Microsoft Azure Blog
Microsoft Azure Blog
Jina AI
Jina AI
罗磊的独立博客
V
Visual Studio Blog
Last Week in AI
Last Week in AI
阮一峰的网络日志
阮一峰的网络日志
IT之家
IT之家
aimingoo的专栏
aimingoo的专栏
雷峰网
雷峰网
酷 壳 – CoolShell
酷 壳 – CoolShell
美团技术团队
博客园 - 三生石上(FineUI控件)
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
MongoDB | Blog
MongoDB | Blog
小众软件
小众软件
IntelliJ IDEA : IntelliJ IDEA – the Leading IDE for Professional Development in Java and Kotlin | The JetBrains Blog
IntelliJ IDEA : IntelliJ IDEA – the Leading IDE for Professional Development in Java and Kotlin | The JetBrains Blog

DEV Community

Autonomous Agents Need Receipts, Not Just Reasoning What 3.9M powerlifting records tell us about competition strategy — an EDA with Python Dev.to Article Draft #13 Beyond the Context Window: How to Build a Self-Improving AI Agent with Persistent Memory Supply Chain Attacks + Stale Credentials: Why This Combination Is So Dangerous in 2026 Daily Briefing Platform Banning Agent PRs Won't Save Open Source Hitting Merge: Mentally Preparing for Your First Push to Production Learning Progress Pt.17 Monitoring Containers on AWS ECS with CloudWatch Tier 4 — Entity and Authority: Wikidata, KG, sameAs threading LocalFind Gemma — AI-Powered Semantic Search and Chat for Your Local Files AI-dy: On-Device Emergency First Aid with Gemma 4 Datrix: Chat With Your Data Using Gemma 4 — Charts, ML Models, No Code Understanding Reinforcement Learning with Human Feedback Part 4: Teaching Models Human Preferences The Architect’s Pivot: Mastering Parallel Agent Orchestration with Antigravity 2.0 Quidditch - Powered By PostgreSQL and ASP.NET Build a Database Connection Framework In 133 Lines Of Code How I mapped 600+ GPS audio-guides as a solo dev (and why I finally did it after 8 years) Installing Terminal & WSL (Windows Subsystem for Linux) A Floating Productivity Panel I Built for Android The Microsecond Lie: Why your Go timers are lying about the GPU Google used 6,000 open-source contributors then locked the door. Classic. Terceira semana tentando voltar ao mercado de trabalho How I turned a Python function into a web app in one decorator I Got Tired of Heavy Design Tools… So I Built My Own 😩 The Google I/O 2026 Moment That Quietly Changed How I See AI Getting Started: Run Your First Local LLM in 5 Minutes Building a 1% Fee Web3 Marketplace for Study Notes: Is a 5% Shift Sustainable? Full Agentic Stack - 5 Ideias da Arquitetura 'AI-First' que Vão Mudar a Forma Como Você Desenvolve Software Build Club Week Four: the part of Themis Lex I never explained I Tried Google Antigravity 2.0 Here's What It Actually Feels Like to Code With AI Agents By Isaac Yakubu | Google I/O 2026 Challenge Submission The growth quest picks what you avoid, not what you're already good at Firebase AI Logic's Template-Only Mode Is the Security Feature We Actually Needed Hardware Guide: What Do You Actually Need to Run Local LLMs? Constitutional Exception Committees: A Pattern for AI Agent Constraint Governance Veltrix's Treasure Hunt Engine: Optimized for Long-Term Survival, Not Just Scalability Open WebUI: Your Local ChatGPT Build a streaming UI without overcomplicating it The Cost of Kernel CVE Patching Frequency in SLA Commitments Gemma 4 Runs on a Raspberry Pi. Let That Sink In. The Git Filesystem - Recreating the Content-Addressable Database Why I Still Believe Our Event-Driven Architecture Was The Right Call For Veltrix Local RAG: Chat With Your Documents (Open Source, Private) GGUF & Modelfile: The Power User's Guide to Local LLMs What Excited Me Most at Google I/O 2026 OSS assemble! Kilo Code is launching on Product Hunt. Join the launch! https://www.producthunt.com/products/kilocode Your Organizational AI Adoption Metrics Are Lying (Plus How to Measure Real Adoption) Building a Production-Grade MLOps Home Lab on Windows — K8s, LLM, RAG & GitLab CI The Moment I Realized AI Agents are Changing Software Forever
全权能栈——五项“AI优先”架构理念,将革新您软件开发之形
suissAI · 2026-05-24 · via DEV Community

数十年间,"全栈开发者"之称号,象征技艺之巅峰。此辈能于前端、后端、数据库及用户体验间,游刃有余,驾驭自如。然今,根本之变,已然发生。

A 语言模型之升,为技术之堆添一"认知之层"。融智于器,已非可有可无之选,乃结构之必也,犹若数据库之不可或缺。其喻直而力:五至七年,无本智之系统,其废也,犹无响应之网页。

是时,新范式出焉:曰"全能代理栈"。此乃架构之模,生而具灵,初构即谋思虑,行止有度。今当探其五义之尤彰者,此新模正易软件之制也。


  1. 智械非插件,乃新核之枢也。

至深之变,在于文脉与构架。旧式之制,AI常为后加之器,以应专项之能。新法以AI为先,则AI为系统运作之基——遂成枢要。

此新构架,名曰"全智能栈"(Full Agentic Stack),其义可释如下:

全能栈乃一完整软件生态,由认知层、自主层与反应层构成,诸层协舞,以解、决、行实时之变。

为匠者,此定义化为一组熟稔之工程范式,相得益彰:事驱动架构(EDA)、如CQRS及事件溯源之范式,并辅以由认知代理协奏而成之基础,此代理与反应式、多模态之数据层相交互。

忽此知性之层,犹若无库之建新用,非其本也。无库之系,不可为用;无知性之系,不可争胜。盖其无学之能,无适之巧,无复繁务之协,此乃旧制之理,固不能及也。


  1. 尔次之API乃自然语言

依传统模式,与系统交互乃直通HTTP请求,须用户或他系统识API之结构。全能代理栈则逆此理。交互始于自然语言,继而译为"意向"之结构.

模型 请求
依古之HTTP径直至一端点(GET /api/orders?client=...)
全能栈 自然言 → 结构意

今举商贾之实例。一客书曰:

“吾欲周知约翰客近周所为之货。”

吾有项目名曰"CogGate",此项目通过大语言模型,将此句解读,并转化为系统所能理解的JSON结构化指令。

{
"意旨": "依客列单"
"参数": {
"客户": "João",
"期间": "本周"
}
}

其主效在于交互之直觉大增。系统应用户之需,用户无须谙悉SQL,亦不必识API之名,更不需知数据库之内部结构。其繁复,乃为智能之层所化。

O CogGate 乃一项目,能化任何旧有系统为认知MCP服务器。其能将自然语言提示转为执行既有功能,唯需添一提示以辨功能之用,并定其所需之架构,于v0.1.0版本中,吾以zod为之。是故,但运行openintent.forger.ts,即可读取汝之coggate.json,此即其事也。

{
  "$schema": "https://neurohive.dev/schemas/cognitivegateway-config.schema.json",
  "version": "1.0.0",
  "mode": "cognitive",
  "description": "Configuração base do Cognitive Gateway para execução de intents via OIP.",
  "oip": {
    "file": "openintent.json",
    "autoRegenerate": true
  },
  "llm": {
    "provider": "openai",
    "model": "gpt-4o-mini",
    "apiKey": "sk-proj-..."
  },
  "controllers": [
    "src/modules/*/controllers/**/*.ts"
  ],
  "gateway": {
    "promptRoute": "/prompt",
    "port": 8080,
    "logLevel": "info",
    "language": "pt-BR"
  },
  "execution": {
    "sandbox": false,
    "maxTokens": 512,
    "timeout": 10000
  }
}

入全景模式 出全景模式

Nessa v0.1.0 eu fiz em cima da arquitetura modularizada onde cada action de um Controler 乃一模块化、原子化之文件,须知其控驭含诸函数之对象,此函数乃由路由所执行。于v0.2.0版本,吾将施此机制于类文件及控制器文件,其于同一文件内含有多函数者,吾将采其常见之例,俾开发者无需更易一码,惟增提示与架构耳。

openintent.forger.ts 将生成openintent.json 乃自然语言意图之 OpenAPI,其联接各功能之提示,及其模式,俾 LLM 得辨何功能应用,据输入提示而定。

LLM 亦不执行一事,惟言何功能应行,及何数据应供是提示。

吾惟以此示 LLM:

const list = Object.entries(intents)
      .map(([id, v]) => `${id}: ${v.prompt}`)
      .join("\n");

//...

{ role: "system", content: "Classifique o prompt abaixo em uma das intenções listadas." },
{ role: "user", content: `${list}\n\nPrompt: ${prompt}` }

Enter fullscreen mode 退出全屏模式

list所含之文,乃"[函数名]:身份提示"之语。

欲使大语言模型依此提示,唯应回其Intent(函数名)。此乃openintent.json中例之貌:

 "intents": {
    "update": {
      "prompt": "Use esta intenção quando o usuário quiser **atualizar dados existentes**.\n\nExtraia o identificador (nome ou id) e os campos a atualizar.\nRetorne **somente JSON** no formato:\n{\n  \"identifier\": string | number,\n  \"fields\": {\n    \"name\"?: string,\n    \"email\"?: string,\n    \"phone\"?: string,\n    \"role\"?: string\n  }\n}",
      "examples": [],
      "behavior": {
        "controller": "modules.user.controllers.update",
        "operationId": "update"
      },
      "context": {
        "domain": "users",
        "language": "pt-BR"
      }
    },

进入全屏模式 退出全屏模式

察之,秘在于提示,唯其上,开发者须明示所求之架构,非必送JSON、架构、物之属。

此乃User/Controller/update.ts之文。

import { z } from "zod";
import { UserService } from "../services/index.js";

/**
 * @intent update
 * @example Mude o email da Maria para teste@teste.com
 * @context domain:users, language:pt-BR
 */
export const prompt = `Use esta intenção quando o usuário quiser **atualizar dados existentes**.

Extraia o identificador (nome ou id) e os campos a atualizar.
Retorne **somente JSON** no formato:
{
  "identifier": string | number,
  "fields": {
    "name"?: string,
    "email"?: string,
    "phone"?: string,
    "role"?: string
  }
}
`;

export const schema = z.object({
  identifier: z.union([z.string(), z.number()]),
  fields: z.object({
    name: z.string().optional(),
    email: z.string().email().optional(),
    phone: z.string().optional(),
    role: z.enum(["admin", "user", "guest"]).optional()
  })
});

export async function update(req, reply) {
  try {
    const { identifier, fields } = req.body;
    const service = new UserService();
    const updated = await service.update(identifier, fields);
    reply.send({ success: true, data: updated });
  } catch (error) {
    reply.status(400).send({ success: false, error: error.message });
  }
}

入全景模式 出全景模式

此架构当为典范,盖吾欲弘传此原子模块之制,使尔等制函数,若尔悟之,吾可为之:

//actions/controller.update.ts

入全景模式 出全景模式

盖因执行于路由之功能,当无业务规则,唯应验契约/架构,俾领域层可用之。吾所知至简之结构乃:

路由(输入/输出之关口)->控制器(数据之验)-> 服务(业务规则执行)—> 仓库(数据持久化)

君以为,无复增责于单层,事岂更简乎?

故,诸路之务,当皆通约,尔惟用一工坊制路,但纳实体之名与行名于是途。

app.post("/users", RoutesActionsFactory.forge("User", "create")

入全屏模式 出全屏模式


ts

无他,盖汝已有此架构于数据平面/src/entities/user/schema.json察之,文件之名,当不含实体之名,惟其本也,盖其已存于实体之文件夹中,如是则助生成动态之码,例若JSON架构:

{
  "$schema": "https://json-schema.org",
  "title": "User",
  "type": "object",
  "properties": {
    "name": {
      "type": "string"
    },
    "username": {
      "type": "string"
    },
    "cpf": { 
      "type": "string", 
      "pattern": "^\\d{3}\\.\\d{3}\\.\\d{3}-\\d{2}$" 
    },
    "email": {
      "type": "string",
      "format": "email"
    },
    "email": {
      "type": "string",
      "format": "email"
    },
    "phone": {
      "type": "string"
    },
    "password": {
      "type": "string"
    },
    "createdAt": {
      "type": "string",
      "format": "date-time"
    },
    "updatedAt": {
      "type": "string",
      "format": "date-time"
    },
    "deletedAt": {
      "type": [
        "string",
        "null"
      ],,
    "default": null
    "format": "date-time"
    }
  },
  "required": [
    "name",
    "username",
    "email",
    "cpf",
    "password",
  ]
}

Enter fullscreen mode 退出全屏模式

汝当自问:

验证身份证之功能安在?

吾答曰:

以SemanticTypes观之,万物皆备一validate,此乃常法。

然今论常例:汝当立一文书。/src/entities/user/cpf.validate.ts

何時生成《諸子》之範式與其ORM/ODM之範式,凡有屬檔名含「.validate.ts」之屬性,皆可簡以require/import,無需於每層使用處重寫此函數。吾未知君意,然最少之驗證有二:路由與儲存庫。吾之架構中,《語義類型》之函數「驗證」遇一值即自動執行,開發者無需寫一字。若值途經中間被改,則終至流程末尾方知,引入此函數毫無損失,僅需喚之。

// src/routes/user.routes.ts
import { RoutesActionsFactory } from "../shared/factories/RoutesActionsFactory.js";

export async function userRoutes(app: any) {
// 动态生成洁净之路由,以Factory为用
app.post("/users", RoutesActionsFactory.forge("User", "create"));
app置"/users",RoutesActionsFactory造("User","updateBy")
app删"/users",RoutesActionsFactory造("User","deleteBy")
app取"/users",RoutesActionsFactory造("User","findAll")
app取"/users/:identifier",RoutesActionsFactory造("User","findBy")
}


  1. 思辨之系非徒行脚本者

常制之构,恃固性之控,循既定之序。若A生,则行B。然于全能之栈,此制易为“诸使”之协舞。每使司一域(价、诉、宣),互议相协,以成繁务,遂启“自生之态”。

思此高阶之境:一管理者请于系统曰

"增诸商品之折,其于近四十八时内,所受之诉言逾十则也。"

非惟一脚本,乃次第之变与协动而生焉:

  1. 投诉之使(ComplaintAgent)察商品之诉率甚高,乃布一"高诉率"之变(HighComplaintRate(product_id))。
  2. O 定價之代理者,啟此事件,察他因(若銷售之趨勢),遂決施新折。
  3. 其決記之,啟營銷之代理者,以新編網上目錄,或訊諸客。

无中枢者统御其步。终局之果,生于专司之剂相与交感,若"自适之舞",随境迁化,赖以 RabbitMQ、NATS、Kafka 诸强健之事件总线为之赋能。


  1. 开发者由编程者蜕为认知流之舞者。

是此新架构之需,须为开发者立一新名。其变也,自“全栈开发者”而转“全能开发者”。

其一主司前后端之层,其二主司"认知之流"。其任之要易矣。非复惟书代码以施固定之能,乃须谋制、训习、理智能体相与之际。其注移自代码之版,若吾所识之语义版(SemVer),转至行为之版("Behavioral Versioning"),务使智能体之交相演进而有度可循。


  1. O futuro é declarativo: sistemas que se autocompõem

Full Agentic Stack之终局,指范式之变,由"如何"转"为何"。不复编程系统须循之步履(如何),但由开发者宣示所期之果(为何),使自主之能者,于运行时自组微服务。

此乃于声明式文件中描述一整系统之可能,如 .tyflow.yaml。譬如:

代理人:

  • 名:客 事:[客成,客新] 存:postgres
  • 名:訂單 儲存:MongoDB 消費:[客戶創建] 生產:[訂單已置]
  • 名:Analytics 存储:weaviate 监听:[OrderPlaced]

Ao executar tyflow run, o sistema cria filas, APIs, eventos e dashboards automaticamente — sem código manual. Isso aponta para um futuro de "engenharia sem código" — não no sentido de simplicidade, mas de alta abstração, onde a arquitetura se torna auto-adaptativa, reconfigurando fluxos com base no contexto. __JHSNS_SEG_e1ef8c7a_105__行tyflow run,系统自创队列、API、事件及仪表盘,无需手写代码。此乃“无码工程”之未来所向——非指简易,实乃高阶抽象,使架构能自适,依情境重构流变。


O "全能栈"非仅新工之集;乃软件发展之逻辑演进也。其认智能非复附属,实为根本且结构之要件。犹云与响应式设计已成不可或缺之范,原生智能当为现代软件之次柱。

吾辈正入一新时代,其终极之旨骤变。盖未来之发展,非惟编程智能之系统,亦乃系统自编程之系统也。

吾辈可备此新世之发展乎?