AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
MCP verified Apache-2.0 Self-run

LLM Dev Demo

mcp-a3rdlab-llm-dev-demo · by A3RDLab

Hands-on examples for LLM application development, including agents, RAG, MCP, structured output, and Text-to-SQL.

No reviews yet
0 installs
9 views
0.0% view→install

Install

$ agentstack add mcp-a3rdlab-llm-dev-demo

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets Used
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/mcp-a3rdlab-llm-dev-demo)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
1mo ago

Declared compatibility

Claude CodeClaude DesktopCursorWindsurf

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of LLM Dev Demo? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

LLM 应用开发教程实验合集

大模型应用开发的系列示例,覆盖从基础调用到 Agent 编排的完整学习路线。

A³ R&D Lab(读作 A-cubed R&D Lab / A立方研发实验室)维护。

> 📖 新手请从[实验手册.md](实验手册.md)开始——一步步引导式实验,全部命令经过实机验证。

环境准备

pip install -r requirements.txt

# 配置 API Key(复制模板并填入自己的 Key,.env 不会被提交)
cp .env.example .env
# 或直接 export:
export API_KEY=
export DASHSCOPE_API_KEY=

所有脚本支持 --model / --api_key / --base_url 参数切换模型与平台(默认硅基流动)。

Demo 索引

基础调用

| Demo | 说明 | |---|---| | first_llm_call.py | 最简单的单次调用 | | first_llm_app.py | 流式多轮对话,含推理模型思考过程展示 | | structured_output_demo.py | 结构化输出:json_schema 约束 + Pydantic 校验,信息抽取场景 |

Agent / Function Calling(推荐学习顺序)

| Demo | 说明 | |---|---| | function_call_and_agent_demo/demo-agent-with-tools-arg.py | 入门:原生 function calling + 手写调用循环(GitHub 仓库分析场景) | | function_call_and_agent_demo/langgraph_demo/ | 进阶:LangGraph 图编排,手搓 StateGraph vs create_react_agent 对比 | | function_call_and_agent_demo/agent_harness_demo/ | 进阶:Agent Harness(DeepAgents)+ Agent Skills 渐进式披露,接入 anthropics/skills 现成 Skill | | function_call_and_agent_demo/mcp_server_demo.py + mcp_client_agent_demo.py | MCP 协议:FastMCP 服务端 + Streamable HTTP 客户端 agent | | function_call_and_agent_demo/excel-process/ | Streamlit + 代码沙箱的 Excel 处理应用(独立依赖,见其 requirements.txt) | | function_call_and_agent_demo/archive/ | 已归档的历史写法(手写 ReAct prompt 等),附归档原因说明 |

RAG

| Demo | 说明 | |---|---| | mini_rag_demo.py | 自包含 RAG 全链路:HTML清洗→切块→embedding→余弦检索→生成(索引缓存,无需 Redis) | | AliyunQA_RAG_demo/ | 阿里云运维问答:Scrapy 采集 + Redis 向量库 + 向量召回 + rerank 精排(Streamlit 界面) | | PDF_RAG_demo/ | PDF 文档 RAG(notebook,含表格/图片处理) | | wiki_kb_agent_demo/ | LLM Wiki 知识库接入 Agent:从 index.md 出发的自然语言导航式检索(readpage/listpages/grep_kb 工具,零 embedding) |

Text2SQL

| Demo | 说明 | |---|---| | text2sql_demo.py | schema 自动内省 + function calling 执行 SQL + 错误自修复(Chinook 示例库) | | vanna_demo.ipynb | Vanna 框架 Text2SQL(对比路线) |

生产风格完整项目

| Demo | 说明 | |---|---| | finance_doc_review/ | 金融文档审核:LangGraph 编排 + YAML 规则引擎 + Pydantic AI 结构化输出(独立 pyproject,见其 README) |

推荐学习路线

  1. 基础first_llm_call.pyfirst_llm_app.pystructured_output_demo.py
  2. Agentdemo-agent-with-tools-arg.pylanggraph_demoagent_harness_demo(Agent Harness + Skills)→ MCP demo
  3. RAGmini_rag_demo.py(无依赖全链路)→ AliyunQA_RAG_demo(召回+重排)→ PDF_RAG_demo
  4. 数据text2sql_demo.pyexcel-process
  5. 综合finance_doc_review/

注意事项

  • 切勿在代码中硬编码 API Key,一律使用环境变量(见 .env.example
  • 过时写法统一移入各目录的 archive/,其中的 README 说明了归档原因与替代方案
  • 人工智能生成内容仅供参考

许可证与第三方材料

  • A³ R&D Lab 原创代码与文档采用 [Apache License 2.0](LICENSE)。
  • 第三方代码、模型文件、数据集、公开报告和示例语料不因收入本仓库而改变其原有权利状态,也不自动纳入 Apache-2.0。
  • 复用前请阅读 [THIRDPARTYNOTICES.md](THIRDPARTYNOTICES.md) 及相关子目录内的许可证或来源说明。
  • 仓库包含的 Anthropic xlsx Skill 受其目录内独立条款约束,不属于本仓库的 Apache-2.0 授权范围。

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.