ys/GenericAgent
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

update the readme

Browse Source
This commit is contained in:
Jinyi Han
2026-03-14 01:45:14 +08:00
parent ab84bc686c
commit 80ced3081d
1 changed files with 18 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
README.md
View File

@@ -14,12 +14,12 @@
Its design philosophy: **don't preload skills — evolve them.** Its design philosophy: **don't preload skills — evolve them.**
Every time GenericAgent solves a new task, it automatically crystallizes the execution path into an SOP for direct reuse later. The longer you use it, the more skills accumulate — forming a skill tree that belongs entirely to you, grown from 3,300 lines of seed code. Every time GenericAgent solves a new task, it automatically crystallizes the execution path into an skill for direct reuse later. The longer you use it, the more skills accumulate — forming a skill tree that belongs entirely to you, grown from 3,300 lines of seed code.
> **🤖 Self-Bootstrap Proof** — Everything in this repository, from installing Git and running `git init` to every commit message, was completed autonomously by GenericAgent. The author never opened a terminal once. > **🤖 Self-Bootstrap Proof** — Everything in this repository, from installing Git and running `git init` to every commit message, was completed autonomously by GenericAgent. The author never opened a terminal once.
## 📋 Core Features ## 📋 Core Features
- **Self-Evolving**: Automatically crystallizes each task into an SOP. Capabilities grow with every use, forming your personal skill tree. - **Self-Evolving**: Automatically crystallizes each task into an skill. Capabilities grow with every use, forming your personal skill tree.
- **Minimal Architecture**: ~3,300 lines of core code. Agent Loop is just 92 lines. No complex dependencies, zero deployment overhead. - **Minimal Architecture**: ~3,300 lines of core code. Agent Loop is just 92 lines. No complex dependencies, zero deployment overhead.
- **Strong Execution**: Injects into a real browser (preserving login sessions). 7 atomic tools take direct control of the system. - **Strong Execution**: Injects into a real browser (preserving login sessions). 7 atomic tools take direct control of the system.
- **High Compatibility**: Supports Claude / Gemini / Kimi and other major models. Cross-platform. - **High Compatibility**: Supports Claude / Gemini / Kimi and other major models. Cross-platform.
@@ -31,14 +31,14 @@ This is what fundamentally distinguishes GenericAgent from every other agent fra
``` ```
[New Task] --> [Autonomous Exploration] (install deps, write scripts, debug & verify) --> [New Task] --> [Autonomous Exploration] (install deps, write scripts, debug & verify) -->
[Crystallize Execution Path into SOP] --> [Write to Memory Layer] --> [Direct Recall on Next Similar Task] [Crystallize Execution Path into skill] --> [Write to Memory Layer] --> [Direct Recall on Next Similar Task]
``` ```
| What you say | What the agent does the first time | Every time after | | What you say | What the agent does the first time | Every time after |
|---|---|---| |---|---|---|
| *"Read my WeChat messages"* | Install deps → reverse DB → write read script → save SOP | **one-line invoke** | | *"Read my WeChat messages"* | Install deps → reverse DB → write read script → save skill | **one-line invoke** |
| *"Monitor stocks and alert me"* | Install mootdx → build selection flow → configure cron → save SOP | **one-line start** | | *"Monitor stocks and alert me"* | Install mootdx → build selection flow → configure cron → save skill | **one-line start** |
| *"Send this file via Gmail"* | Configure OAuth → write send script → save SOP | **ready to use** | | *"Send this file via Gmail"* | Configure OAuth → write send script → save skill | **ready to use** |
After a few weeks, your agent instance will have a skill tree no one else in the world has — all grown from 3,300 lines of seed code. After a few weeks, your agent instance will have a skill tree no one else in the world has — all grown from 3,300 lines of seed code.
@@ -192,8 +192,8 @@ python tgapp.py
| **Deployment** | `pip install` + API Key | Multi-service orchestration | CLI + subscription | | **Deployment** | `pip install` + API Key | Multi-service orchestration | CLI + subscription |
| **Browser Control** | Real browser (session preserved) | Sandbox / headless browser | Via MCP plugin | | **Browser Control** | Real browser (session preserved) | Sandbox / headless browser | Via MCP plugin |
| **OS Control** | Mouse/kbd, vision, ADB | Multi-agent delegation | File + terminal | | **OS Control** | Mouse/kbd, vision, ADB | Multi-agent delegation | File + terminal |
| **Self-Evolution** | Autonomous SOP growth | Plugin ecosystem | Stateless between sessions | | **Self-Evolution** | Autonomous skill growth | Plugin ecosystem | Stateless between sessions |
| **Out of the Box** | 10 .py files + 5 SOPs | Hundreds of modules | Rich CLI toolset | | **Out of the Box** | 10 .py files + 5 skills | Hundreds of modules | Rich CLI toolset |
## 🧠 How It Works ## 🧠 How It Works
@@ -205,7 +205,7 @@ GenericAgent accomplishes complex tasks through **Layered Memory × Minimal Tool
- **L0 — Meta Rules**: Core behavioral rules and system constraints of the agent - **L0 — Meta Rules**: Core behavioral rules and system constraints of the agent
- **L2 — Global Facts**: Stable knowledge accumulated over long-term operation - **L2 — Global Facts**: Stable knowledge accumulated over long-term operation
- **L3 — Task SOPs**: Workflows for completing specific task types - **L3 — Task Skillss**: Workflows for completing specific task types
2️⃣ **Autonomous Execution Loop** 2️⃣ **Autonomous Execution Loop**
@@ -269,12 +269,12 @@ MIT License — see [LICENSE](LICENSE)
它的设计哲学是：**不预设技能，靠进化获得能力。** 它的设计哲学是：**不预设技能，靠进化获得能力。**
每解决一个新任务，GenericAgent 就将执行路径自动固化为 SOP，供后续直接调用。使用时间越长，沉淀的技能越多，形成一棵完全属于你、从 3,300 行种子代码生长出来的专属技能树。 每解决一个新任务，GenericAgent 就将执行路径自动固化为 Skill，供后续直接调用。使用时间越长，沉淀的技能越多，形成一棵完全属于你、从 3,300 行种子代码生长出来的专属技能树。
> **🤖 自举实证** — 本仓库的一切，从安装 Git、`git init` 到每一条 commit message，均由 GenericAgent 自主完成。作者全程未打开过一次终端。 > **🤖 自举实证** — 本仓库的一切，从安装 Git、`git init` 到每一条 commit message，均由 GenericAgent 自主完成。作者全程未打开过一次终端。
## 📋 核心特性 ## 📋 核心特性
- **自我进化**: 每次任务自动沉淀 SOP，能力随使用持续增长，形成专属技能树 - **自我进化**: 每次任务自动沉淀 Skill，能力随使用持续增长，形成专属技能树
- **极简架构**: ~3,300 行核心代码，Agent Loop 仅 92 行，无复杂依赖，部署零负担 - **极简架构**: ~3,300 行核心代码，Agent Loop 仅 92 行，无复杂依赖，部署零负担
- **强执行力**: 注入真实浏览器（保留登录态），7 个原子工具直接接管系统 - **强执行力**: 注入真实浏览器（保留登录态），7 个原子工具直接接管系统
- **高兼容性**: 支持 Claude / Gemini / Kimi 等主流模型，跨平台运行 - **高兼容性**: 支持 Claude / Gemini / Kimi 等主流模型，跨平台运行
@@ -285,17 +285,17 @@ MIT License — see [LICENSE](LICENSE)
``` ```
[遇到新任务]-->[自主摸索](安装依赖、编写脚本、调试验证)--> [遇到新任务]-->[自主摸索](安装依赖、编写脚本、调试验证)-->
[将执行路径固化为 SOP]-->[写入记忆层]-->[下次同类任务直接调用] [将执行路径固化为 Skill]-->[写入记忆层]-->[下次同类任务直接调用]
``` ```
| 你说的一句话 | Agent 第一次做了什么 | 之后每次 | | 你说的一句话 | Agent 第一次做了什么 | 之后每次 |
|---|---|---| |---|---|---|
| *"监控股票并提醒我"* | 安装 mootdx → 构建选股流程 → 配置定时任务 → 保存 SOP | **一句话启动** | | *"监控股票并提醒我"* | 安装 mootdx → 构建选股流程 → 配置定时任务 → 保存 Skill | **一句话启动** |
| *"用 Gmail 发这个文件"* | 配置 OAuth → 编写发送脚本 → 保存 SOP | **直接可用** | | *"用 Gmail 发这个文件"* | 配置 OAuth → 编写发送脚本 → 保存 Skill | **直接可用** |
用几周后，你的 Agent 实例将拥有一套任何人都没有的专属技能树，全部从 3,300 行种子代码中生长而来。 用几周后，你的 Agent 实例将拥有一套任何人都没有的专属技能树，全部从 3,300 行种子代码中生长而来。
<!-- | *"帮我读取微信消息"* | 安装依赖 → 逆向数据库 → 写读取脚本 → 保存 SOP | **一句话调用** | --> <!-- | *"帮我读取微信消息"* | 安装依赖 → 逆向数据库 → 写读取脚本 → 保存 Skill | **一句话调用** | -->
#### 🎯 实例展示 #### 🎯 实例展示
@@ -448,8 +448,8 @@ python tgapp.py
| **部署方式** | `pip install` + API Key | 多服务编排 | CLI + 订阅 | | **部署方式** | `pip install` + API Key | 多服务编排 | CLI + 订阅 |
| **浏览器控制** | 注入真实浏览器（保留登录态） | 沙箱 / 无头浏览器 | 通过 MCP 插件 | | **浏览器控制** | 注入真实浏览器（保留登录态） | 沙箱 / 无头浏览器 | 通过 MCP 插件 |
| **OS 控制** | 键鼠、视觉、ADB | 多 Agent 委派 | 文件 + 终端 | | **OS 控制** | 键鼠、视觉、ADB | 多 Agent 委派 | 文件 + 终端 |
| **自我进化** | 自主生长 SOP 和工具 | 插件生态 | 会话间无状态 | | **自我进化** | 自主生长 Skill 和工具 | 插件生态 | 会话间无状态 |
| **出厂配置** | 10 个 .py + 5 个 SOP | 数百模块 | 丰富 CLI 工具集 | | **出厂配置** | 10 个 .py + 5 个 Skills | 数百模块 | 丰富 CLI 工具集 |
## 🧠 工作机制 ## 🧠 工作机制
@@ -462,7 +462,7 @@ GenericAgent 通过**分层记忆 × 最小工具集 × 自主执行循环**完
- **L0 — 元规则（Meta Rules）**：Agent 的基础行为规则和系统约束 - **L0 — 元规则（Meta Rules）**：Agent 的基础行为规则和系统约束
- **L2 — 全局事实（Global Facts）**：在长期运行过程中积累的稳定知识 - **L2 — 全局事实（Global Facts）**：在长期运行过程中积累的稳定知识
- **L3 — 任务 SOP（Standard Operating Procedure）**：完成特定任务的操作流程 - **L3 — 任务 Skills（Standard Operating Procedure）**：完成特定任务的操作流程
2️⃣ **自主执行循环** 2️⃣ **自主执行循环**