Lanqiao-Luca/ai-java-workflow
☕ 用 AI 辅助 Java / Spring Boot 研发(AI for Java / AI for Spring)|需求、实现、测试、审查与交付|Graph → Loop → Workflow → Agent / Skill、虚构可运行案例及 worktree / Docker 工具|面向业务开发,非 Spring AI 框架实现
What it solves
This project provides a structured framework for managing AI-assisted Java and Spring Boot development. It solves the problem of disorganized AI workflows by offering a standardized method to handle requirements, implementation, testing, review, and delivery, ensuring that AI agents follow consistent rules and maintain project integrity.
How it works
The framework is built on a hierarchical architecture: Standards (rules) and Tools (execution) are combined into Skills, which are then used by Agents. These agents are organized into Workflows (stages), Loops (repeated tasks), and Graphs (complex dependencies). A central "Brain" component routes tasks to the appropriate layer, ensuring that AI agents only perform necessary actions while preserving existing code behavior and project evidence.
Who it’s for
It is designed for developers and AI assistants who want to apply structured, repeatable engineering methods to Java and Spring Boot projects. It is particularly useful for those who need to manage complex, multi-stage development tasks while maintaining strict control over code changes and verification.
Highlights
- Structured Roles: Includes 20 defined Agent roles with clear responsibilities, inputs, and delivery criteria.
- Evidence-Based: Emphasizes the collection of verifiable evidence, such as test results and code diffs, to ensure changes meet requirements.
- Modular Design: Allows users to pick and choose components—from simple agent tasks to complex, multi-loop graphs—without requiring the use of the entire framework.
- Safety-First: Supports sandboxed execution environments (like Docker or macOS sandboxes) to isolate code execution and prevent unauthorized system access.
- Proven Examples: Provides concrete, runnable examples for common tasks like Spring Boot batch archiving and SQL query review.
Related
- Project
- Project
- Project
- Project