spring-ai-community/spring-ai-agent-utils

What it solves

Spring AI Agent Utils provides a set of pre-built tools and capabilities inspired by Claude Code, allowing Java developers to build sophisticated AI agents that can interact with the local environment, manage tasks, and maintain long-term memory. It bridges the gap between high-level agentic frameworks and the Spring AI ecosystem, providing the necessary "skills" for an agent to act as a coding assistant or a general-purpose autonomous agent.

How it works

The library implements a suite of tools that can be registered with a Spring AI ChatClient. These tools allow the model to perform actions such as reading/writing files, executing shell commands, searching the web, and managing a structured todo list. It also introduces a sub-agent system for delegating complex tasks to specialized agents and a memory system that uses sandboxed file-based storage to persist information across conversations.

Who it’s for

Java developers using Spring AI who want to build agentic applications, such as AI coding assistants, autonomous task managers, or specialized AI agents with extensible knowledge modules.

Highlights

• Core Environment Tools: Includes FileSystemTools, ShellTools, LgrepTool, and GlobTool for deep interaction with the local file system and shell.
• Web Access: Integrated BraveWebSearchTool and SmartWebFetchTool for real-world information retrieval.
• Long-term Memory: AutoMemoryTools and AutoMemoryToolsAdvisor provide persistent, file-based memory that survives across sessions.
• Agent Orchestration: TaskTools and TodoWriteTool enable structured task management and delegation to specialized sub-agents.
• Extensible Skills: SkillsTool allows adding reusable knowledge modules defined in Markdown with YAML front-matter.
• Cancellable Turns: InterruptAdvisor allows for cooperative cancellation of agent loops (e.g., via a stop button or budget guard).
• Observability: ToolCallListener provides audit logs and metrics for every tool invocation.