# agent

**Repository Path**: SuNan202011/agent

## Basic Information

- **Project Name**: agent
- **Description**: No description available
- **Primary Language**: Java
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-04-13
- **Last Updated**: 2026-04-14

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Spring Boot + Agent + Skills Demo (LLM Function Calling)

This project is a runnable end-to-end demo with:

- Spring Boot REST API
- Agent orchestration (`plan -> choose skill -> execute`)
- Pluggable skills (`Skill` interface)
- Real LLM function-calling router (OpenAI or Tongyi)
- Automatic fallback to local routing rules

## 1. Requirements

- JDK 17+
- Maven 3.9+

## 2. Run

```bash
mvn spring-boot:run
```

Default port: `8080`

## 3. Modes

### 3.1 Local mode (default)

```yaml
agent:
  planner:
    mode: local
```

The agent selects skills by local rules only.

### 3.2 LLM function-calling mode

```yaml
agent:
  planner:
    mode: llm
  llm:
    provider: openai # openai or tongyi
```

Set API key env vars:

```bash
# OpenAI
set OPENAI_API_KEY=your_openai_key

# Tongyi (DashScope compatibility mode)
set TONGYI_API_KEY=your_tongyi_key
```

## 4. LLM Config

```yaml
agent:
  planner:
    mode: local # local | llm
  llm:
    provider: openai # openai | tongyi
    timeout-seconds: 30
    openai:
      base-url: https://api.openai.com/v1
      api-key: ${OPENAI_API_KEY:}
      model: gpt-4.1-mini
    tongyi:
      base-url: https://dashscope.aliyuncs.com/compatible-mode/v1
      api-key: ${TONGYI_API_KEY:}
      model: qwen-plus
```

## 5. API

### 5.1 List skills

```bash
curl http://localhost:8080/api/skills
```

### 5.2 Chat with agent

```bash
curl -X POST http://localhost:8080/api/agent/chat \
  -H "Content-Type: application/json" \
  -d "{\"message\":\"please calculate 12 * 8\"}"
```

Response includes:

- `answer`: final result
- `trace`: steps (`thought` / `llm` / `plan` / `observation` / `final`)

## 6. Built-in Skills

- `time_skill`: time query
- `math_skill`: arithmetic
- `weather_skill`: weather (demo data)
- `fallback_skill`: fallback response

## 7. Add a New Skill

1. Implement `com.demo.agent.skill.Skill`
2. Add `@Component`
3. Implement methods:
   - `name()`
   - `description()`
   - `supports(String message)`
   - `execute(String message)`

Spring auto-discovers it and injects it into the agent.

## 8. Test

```bash
mvn test
```