# ericlee4191/vllm-doc-skill architecture

> Architecture diagram of the GitHub repository ericlee4191/vllm-doc-skill, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-04). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/ericlee4191/vllm-doc-skill
- Repository: https://github.com/ericlee4191/vllm-doc-skill

## Overview

这是一个 Claude Skill 知识库，不是 vLLM 推理引擎：用户提出架构或部署问题后，Skill 按章节 tags 定位知识，再据此回答并提供配置建议。图中保留了 README 列出的各知识领域；具体调用实现未采样，因此只展示 README 明确说明的流程与检索关系。

Main entry points: [`knowledge/INDEX.md`](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/INDEX.md) (Tags索引), [`SKILL.md`](https://github.com/ericlee4191/vllm-doc-skill/blob/main/SKILL.md) (Skill入口), [`knowledge/08-deployment-optimization.md`](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/08-deployment-optimization.md) (部署与调优), [`knowledge/00-overview.md`](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/00-overview.md) (架构总览)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_entry["Skill入口"]
  node_skill["Skill入口<br/>[SKILL.md]"]
end

subgraph group_retrieval["知识检索"]
  node_index["Tags索引<br/>[INDEX.md]"]
end

subgraph group_domains["知识领域"]
  node_overview["架构总览<br/>[00-overview.md]"]
  node_core["核心架构"]
  node_scheduler["调度与缓存"]
  node_execution["模型执行"]
  node_attention["Attention算子"]
  node_distributed["分布式并行<br/>[05-distributed.md]"]
  node_hardware["量化与硬件"]
  node_advanced["高级推理"]
  node_deployment["部署与调优"]
end

node_user(("提问者"))
node_answer["架构或部署建议"]

node_user -->|"提交问题"| node_skill
node_skill -->|"定位主题"| node_overview
node_skill -->|"检索标签"| node_index
node_index -->|"定位章节"| node_core
node_index -->|"定位章节"| node_scheduler
node_index -->|"定位章节"| node_execution
node_index -->|"定位章节"| node_attention
node_index -->|"定位章节"| node_distributed
node_index -->|"定位章节"| node_hardware
node_index -->|"定位章节"| node_advanced
node_index -->|"定位章节"| node_deployment
node_overview -->|"引导查阅"| node_deployment
node_core -->|"支撑回答"| node_answer
node_scheduler -->|"支撑回答"| node_answer
node_execution -->|"支撑回答"| node_answer
node_attention -->|"支撑回答"| node_answer
node_distributed -->|"支撑回答"| node_answer
node_hardware -->|"支撑回答"| node_answer
node_advanced -->|"支撑回答"| node_answer
node_deployment -->|"支撑建议"| node_answer

click node_skill "https://github.com/ericlee4191/vllm-doc-skill/blob/main/SKILL.md"
click node_overview "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/00-overview.md"
click node_index "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/INDEX.md"
click node_core "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/01-architecture-core.md"
click node_scheduler "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/02-scheduling-kv-cache.md"
click node_execution "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/03-model-execution.md"
click node_attention "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/04-attention-kernels.md"
click node_distributed "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/05-distributed.md"
click node_hardware "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/06-quantization-hardware.md"
click node_advanced "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/07-advanced-features.md"
click node_deployment "https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/08-deployment-optimization.md"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_skill toneBlue
class node_index toneAmber
class node_overview,node_core,node_scheduler,node_execution,node_attention,node_distributed,node_hardware,node_advanced,node_deployment toneMint
class node_user,node_answer toneIndigo
```

## Components

### Skill入口

- [Skill入口](https://github.com/ericlee4191/vllm-doc-skill/blob/main/SKILL.md): `SKILL.md`

### 知识检索

- [Tags索引](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/INDEX.md): `knowledge/INDEX.md`

### 知识领域

- [架构总览](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/00-overview.md): `knowledge/00-overview.md`
- [核心架构](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/01-architecture-core.md): `knowledge/01-architecture-core.md`
- [调度与缓存](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/02-scheduling-kv-cache.md): `knowledge/02-scheduling-kv-cache.md`
- [模型执行](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/03-model-execution.md): `knowledge/03-model-execution.md`
- [Attention算子](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/04-attention-kernels.md): `knowledge/04-attention-kernels.md`
- [分布式并行](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/05-distributed.md): `knowledge/05-distributed.md`
- [量化与硬件](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/06-quantization-hardware.md): `knowledge/06-quantization-hardware.md`
- [高级推理](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/07-advanced-features.md): `knowledge/07-advanced-features.md`
- [部署与调优](https://github.com/ericlee4191/vllm-doc-skill/blob/main/knowledge/08-deployment-optimization.md): `knowledge/08-deployment-optimization.md`

### Other components

- **提问者**
- **架构或部署建议**

## Connections

- 提问者 → Skill入口: 提交问题 (evidence: `README.md`)
- Skill入口 → 架构总览: 定位主题 (evidence: `README.md`)
- Skill入口 → Tags索引: 检索标签 (evidence: `README.md`)
- Tags索引 → 核心架构: 定位章节 (evidence: `README.md`)
- Tags索引 → 调度与缓存: 定位章节 (evidence: `README.md`)
- Tags索引 → 模型执行: 定位章节 (evidence: `README.md`)
- Tags索引 → Attention算子: 定位章节 (evidence: `README.md`)
- Tags索引 → 分布式并行: 定位章节 (evidence: `README.md`)
- Tags索引 → 量化与硬件: 定位章节 (evidence: `README.md`)
- Tags索引 → 高级推理: 定位章节 (evidence: `README.md`)
- Tags索引 → 部署与调优: 定位章节 (evidence: `README.md`)
- 架构总览 → 部署与调优: 引导查阅 (evidence: `README.md`)
- 核心架构 → 架构或部署建议: 支撑回答 (evidence: `README.md`)
- 调度与缓存 → 架构或部署建议: 支撑回答 (evidence: `README.md`)
- 模型执行 → 架构或部署建议: 支撑回答 (evidence: `README.md`)
- Attention算子 → 架构或部署建议: 支撑回答 (evidence: `README.md`)
- 分布式并行 → 架构或部署建议: 支撑回答 (evidence: `README.md`)
- 量化与硬件 → 架构或部署建议: 支撑回答 (evidence: `README.md`)
- 高级推理 → 架构或部署建议: 支撑回答 (evidence: `README.md`)
- 部署与调优 → 架构或部署建议: 支撑建议 (evidence: `README.md`)
