Autonomous Unit Test Generation Agent powered by LangGraph FSM + E2B Sandbox + DeepSeek-V3
Autonomous Unit Test Generation Agent powered by LangGraph FSM + E2B Sandbox + DeepSeek-V3
基于 LangGraph 有限状态机 + E2B 沙箱 + DeepSeek-V3 的自主单元测试生成 Agent
auto-sdet test src/calculator.py
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ Auto-SDET │ target: calculator.py ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
✨ [Generator] Reading file context (MCP)...
✨ [Generator] Context loaded: 2 dependency file(s)
✨ [Generator] Calling DeepSeek-V3...
✓ [Generator] Generated test_calculator.py (42 lines)
⚡ [Executor] Creating E2B sandbox...
⚡ [Executor] Running pytest in sandbox...
✗ [Executor] Tests failed: ImportError: No module named 'numpy'
🧠 [Reflector] Analyzing error (attempt 1/3)...
✓ [Reflector] Patch applied, re-executing...
⚡ [Executor] Running pytest in sandbox...
✓ [Executor] All tests passed!
✨ Test file saved to: tests/test_calculator.py
Auto-SDET implements a deterministic FSM (Finite State Machine) using LangGraph. Unlike a simple LLM chain, the FSM makes every execution path predictable and traceable.
Auto-SDET 使用 LangGraph 实现确定性有限状态机(FSM)。与简单的 LLM 链不同,FSM 使每条执行路径都可预测、可追溯。
[CLI] → [Generator] → [Executor] ──→ END (success)
│
├──→ [Reflector] → [Executor] (self-healing loop / 自愈循环)
│
└──→ END (max retries exceeded)
| Node / 节点 | Responsibility / 职责 |
|---|---|
| Generator | Reads source via MCP → calls DeepSeek-V3 → produces pytest file<br>通过 MCP 读取源码 → 调用 DeepSeek-V3 → 生成 pytest 测试文件 |
| Executor | Spins up E2B micro-VM → uploads files → runs pytest → captures results<br>启动 E2B 微型虚拟机 → 上传文件 → 运行 pytest → 捕获结果 |
| Reflector | Uses Chain-of-Thought over full error history to patch test code<br>基于完整错误历史进行 CoT 推理,修复测试代码 |
| Router | Routes to next state based on exit_code and retry count< |
HAL 分层混合模型工作流 — 强模型(Claude)负责理解/拆解/验收,低成本模型(DeepSeek)负责检索/提取/清洗。Hermes Agent skill。
An LLM agent fine-tuned on DeepSeek for spaced repetition, dynamically integrating knowledge points based on the Ebbinghaus forgetting curve.
基于 STM32F103 构建的端到端 AI 智能手表生态。自研“零重定位”原生机器码动态加载引擎与页面栈式 UI 框架;集成生产级 OTA 回滚保护机制与高带宽(921600 baud)串口协议栈。通过 Node.js 中继实现 DeepSeek AI 语义控制及 ASRPRO 语音全双工交互,是一个集成了分布式计算、现代存储管理与 AI Agent 的嵌入式全栈工程。
A Meta-Agent-Driven Self-Evolving Multi-Agent System for UAV Detection and Tracking
One command to run Hermes AI Agent with a browser UI. Zero prerequisites. 一行命令,AI 就位。
网页应用Agent,接入DeepSeek、Mimo等模型
Workflows from the Neura Market marketplace related to this DeepSeek resource