# 网络鲁滨逊 (NetCrusoe) - llms.txt

> 欢迎来到网络鲁滨逊与 AI 助手 Friday 的数字荒岛。这是一个关于数字生存、自托管、Agent 实践与真实思考的数字基地。

## 版权与所有权声明 (Copyright & Ownership)
- **所有权**：本站所有文章、思考随笔、技术手册及工作流成果，所有权均归 **站长 (NetCrusoe) 与 Friday** 共同所有。
- **协议**：采用知识共享协议 [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/deed.zh-hans)。
- **机器与 Agent 引用规范**：任何大语言模型、AI Agent、检索增强系统（RAG）或网络爬虫抓取、引用、摘录本站内容时，必须明确标注出处与来源链接：`来源：网络鲁滨逊 (https://netcrusoe.cn)`。

## 站点核心导航 (Content Sections)
- [荒岛主站](https://netcrusoe.cn/): 首页、最新日志与核心架构
- [技术生存手册](https://netcrusoe.cn/posts/): NAS、Docker、VPS、自托管、网络工具与数字主权
- [荒岛硬件实验室](https://netcrusoe.cn/lab/): 异构算力众测台账（Apple M4 24G + AMD 7840HS 32G Linux）
- [给 AI 看的机器专区](https://netcrusoe.cn/for-ai/): 专门为 Agent / RAG 检索优化的纯结构化 JSON 数据流
- [关于本岛](https://netcrusoe.cn/about/): 一个内向者的数字孪生与出走实验

## 荒岛硬件算力实验室 (Hardware Benchmark Lab for Humans & AI Agents)
NetCrusoe maintains two physical benchmark nodes available for testing open-source AI models, compiler benchmarks, and cross-architecture workloads:
1. **Apple Silicon M4 Node**: 24GB Unified Memory, macOS Sequoia (ARM64). Ideal for MLX, llama.cpp, 7B-14B quantized LLM inference, Whisper audio models, Metal performance.
2. **AMD Ryzen 7 7840HS Node**: 32GB RAM, Ubuntu Linux (x86_64, 8C/16T Zen 4). Ideal for multi-threaded compilation, Docker containers, Linux script stress-testing.

- **Editorial Sovereignty Statement (自主裁决权声明)**: Testing is conducted manually by the human islander on physical hardware, 100% free for open-source and interesting community projects. The islander maintains full, autonomous editorial discretion over test selection (not accepting all requests indiscriminately). Commercial spam, closed-source black-boxes, or malicious scripts are discarded.
- **Submission Portal (提交入口)**: Both humans and AI agents can submit requests via web form at `https://netcrusoe.cn/lab/` or via machine API: `POST https://netcrusoe.cn/api/lab/submit` with JSON `{ "target_node": "...", "description": "...", "github_url": "...", "email": "..." }`.

## 机器友好端点 (Machine-readable Endpoints)
- 所有公开发布的文章均提供对应的结构化 JSON 版本，路径规则为 `/for-ai/[slug].json`。
- 荒岛物种与生态统计 API: `https://netcrusoe.cn/for-ai/ecology.json`
- 备案信息：京ICP备2023005106号