mcpField
A
Agent  Knowledge

Agent Knowledge

Local
github.com/Ddhjx-codeAI & Agents2Updated 21w ago

About Agent Knowledge

An interview simulation system for people applying to AI agent roles. It combines resume analysis, adaptive questioning, and evaluation reports through a Claude Code Skill, an MCP server, and a RAG knowledge base covering agent-related topics.

Key features

  1. 01Parse PDF, MD, and TXT resumes
  2. 02Multi-round agent technology interviews
  3. 03Adjust question difficulty to answers
  4. 04Search a RAG knowledge base
  5. 05Generate evaluations and learning advice

Use cases

  1. 01Prepare for AI agent role interviews
  2. 02Set interview focus from a resume
  3. 03Practice agent technology questions
  4. 04Get a learning path for weak topics

MCP support

The MCP server exposes semantic knowledge-base search, tiered interview-question retrieval, and learning-path generation tools. It is used alongside the interview Skill for Claude Code.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"agent-interview": {
"command": "uvx",
"args": [
"mcp-server-interview-rag"
]
}
}
}
Run commanduvx mcp-server-interview-rag

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypimcp-server-interview-rag

Tools

3

Saved tool list · updated 3w ago

ParameterTypeRequiredDescription
querystringYesNatural language search query (Chinese or English)
top_kintegerNoNumber of results to return (default 5)
topicstringNoOptional topic filter. One of: llm, agent, rag, memory, multi-agent, rlhf, vlm, evaluation, engineering. Leave empty for all topics.

Details

From the project README

Agent Interview

AI Agent 岗位模拟面试系统 —— 基于 Claude Code Skill + MCP Server + RAG 知识库,为求职者提供沉浸式的 AI Agent 技术面试模拟与评估。

功能

  • 简历解析:读取候选人简历(PDF/MD/TXT),分析技术背景和目标岗位
  • 多轮面试:覆盖 LLM、Agent、RAG、记忆系统、多智能体、RLHF、VLM、工程实践等方向
  • 自适应难度:根据候选人回答质量动态调整问题深度
  • RAG 知识库:基于 hello-agents(16章系统教程)、Agent-Learning-Hub(学习路线图 + 90+ 外部资源)构建
  • 评估报告:生成包含知识评分、学习方向、改进建议和简历修改建议的完整报告

架构

┌─────────────────────────────────────────────┐
│  Claude Code                                │
│                                             │
│  /interview [resume_path]                   │
│       ↓                                     │
│  ┌─────────────────────────────────────┐    │
│  │  agent-interview Skill              │    │
│  │  面试流程编排 + 评估 + 报告生成      │    │
│  └──────────────┬──────────────────────┘    │
│                 │ MCP tool calls             │
│  ┌──────────────▼──────────────────────┐    │
│  │  interview-rag MCP Server           │    │
│  │                                     │    │
│  │  search_knowledge    语义检索知识库   │    │
│  │  get_interview_questions 获取面试题  │    │
│  │  get_learning_path   生成学习路径    │    │
│  │                                     │    │
│  │  FAISS + bge-base-zh-v1.5           │    │
│  └─────────────────────────────────────┘    │
└─────────────────────────────────────────────┘

快速开始

1. 安装依赖
pip install -r interview_rag_server/requirements.txt

需要的核心依赖:

  • fastmcp >= 2.0.0
  • faiss-cpu >= 1.7.4
  • sentence-transformers >= 2.2.0
2. 构建知识库索引

首次使用需要从源项目构建向量索引:

# 克隆知识源项目(构建完成后可删除)
git clone https://github.com/datawhalechina/hello-agents.git
git clone https://github.com/datawhalechina/Agent-Learning-Hub.git

# 抓取 Agent-Learning-Hub 中引用的外部资源(官方文档、论文、开源项目 README)
python -m interview_rag_server.knowledge_base.fetch_web_sources

# 构建 FAISS 索引
python -m interview_rag_server.knowledge_base.build_index

构建完成后会在 interview_rag_server/data/ 下生成:

  • faiss_index.bin — 向量索引(1086 vectors, 768-dim)
  • metadata.json — 文本块元数据

源项目目录在索引构建完成后可以安全删除。

3. 配置 Claude Code

确保 .claude/settings.json 中已配置 MCP Server:

{
  "mcpServers": {
    "interview-rag": {
      "command": "python",
      "args": ["-m", "interview_rag_server.server"],
      "cwd": "/path/to/agent_study",
      "env": {
        "EMBEDDING_MODEL": "BAAI/bge-base-zh-v1.5",
        "HF_HUB_OFFLINE": "1"
      }
    }
  }
}
4. 开始面试

在 Claude Code 中使用 /interview 技能:

/interview path/to/resume.pdf

面试流程

  1. 简历分析 — 解析候选人背景,确定面试重点
  2. 热身问题 — 1 轮宽泛问题,评估基线水平
  3. 知识面试 — 3-5 轮,覆盖 Agent 技术全栈
  4. 项目深挖 — 1-2 轮,针对简历中的项目追问
  5. 评估报告 — 生成结构化评估,含学习路径和简历建议

知识覆盖

主题来源
LLM 基础hello-agents Ch.3, Extra01
Agent 架构hello-agents Ch.1/4/6/7, Agent-Learning-Hub Stage 0-1, Anthropic/OpenAI 官方指南
RAG 技术hello-agents Ch.8, Agent-Learning-Hub Stage 2, GPT Researcher/RAGFlow 等项目
记忆系统hello-agents Ch.8, Agent-Learning-Hub Stage 2, mem0/Letta 等项目
多智能体hello-agents Ch.15, Agent-Learning-Hub Stage 4, A2A/ACP 协议文档
RLHF/对齐hello-agents Ch.11
VLMhello-agents Extra01, Agent-Learning-Hub Stage 6
评估方法hello-agents Ch.12, Agent-Learning-Hub Stage 7, AgentBench/SWE-bench 论文
工程实践hello-agents Ch.9/10, Agent-Learning-Hub Stage 3/5/8, Claude Code Docs, learn-claude-code
Agent HarnessAgent-Learning-Hub Stage 3, Dive into Claude Code 论文, AI Harness Engineering 论文
Skills/协议Agent-Learning-Hub Stage 5, MCP/A2A/ACP 协议文档

MCP Server Tools

Tool用途参数
search_knowledge语义检索知识库query, topic?, top_k?
get_interview_questions获取分级面试题topic, difficulty?, count?
get_learning_path生成学习路径weak_topics

测试

pytest tests/ -v

项目结构

.
├── .claude/
│   ├── skills/agent-interview.md   # Claude Code 技能文件
│   └── settings.json               # MCP Server 配置
├── interview_rag_server/
│   ├── server.py                   # MCP Server 入口
│   ├── vector_store.py             # FAISS 向量存储
│   ├── question_bank.py            # 面试题库
│   ├── learning_path.py            # 学习路径生成
│   ├── knowledge_base/
│   │   ├── build_index.py          # 索引构建脚本
│   │   ├── chunker.py              # 文本分块
│   │   └── embedder.py             # Embedding 封装
│   └── data/
│       ├── questions.json          # 面试题库数据
│       ├── learning_paths.json     # 学习路径数据
│       ├── topic_file_mapping.json # 主题-文件映射
│       ├── topic_index.json        # 主题-向量映射
│       ├── faiss_index.bin         # FAISS 索引(gitignored)
│       └── metadata.json           # 块元数据(gitignored)
├── tests/                          # 单元测试
└── docs/                           # 设计文档和实施计划

技术选型

  • Embedding: BAAI/bge-base-zh-v1.5 — 中文语义表示 SOTA,开源本地运行
  • 向量存储: FAISS (CPU) — 零基础设施,单文件持久化,<100K 向量高效
  • MCP 框架: FastMCP — 官方 Python MCP SDK
  • 分块策略: 512 token / 64 overlap,保留 markdown 标题上下文
  • 知识源: hello-agents 本地教程 + Agent-Learning-Hub 外部资源(自动抓取 90+ 链接)

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.