# 服务

| 服务名称 | 分支 | 描述 |
| --- | --- | --- |
| mis-debugging | qas_liushuang | SE智能体 |

## 新建的表

```sql
CREATE TABLE `mis-debugging`.`work_sheet` (
    `id`                  BIGINT        NOT NULL AUTO_INCREMENT COMMENT 'ID',
    `system_code`         VARCHAR(64)   NOT NULL COMMENT '系统编码',
    `work_code`           VARCHAR(512)   NOT NULL COMMENT '工单编码',
    `current_try_times`   INT           NOT NULL COMMENT '当前是第几次重试',
    `max_try_times`       INT           NOT NULL COMMENT '最多重试多少次',
    `task_id`             VARCHAR(512)   NOT NULL COMMENT '追踪标识（traceId、业务标识、时间戳等）',
    `state`               VARCHAR(32)   NOT NULL COMMENT '状态：wait-待执行，error-失败，success-成功',
    `json_param`          JSON          NULL     COMMENT '请求参数（JSON格式）',
    `str_param`           VARCHAR(512)  NULL     COMMENT '请求参数（简单字符串）',
    `response_param`      LONGTEXT      NULL     COMMENT '响应参数',
    `error`               LONGTEXT NULL     COMMENT '错误信息',
    `create_time`         DATETIME      NOT NULL COMMENT '创建时间',
    `start_time`          DATETIME      NULL     COMMENT '执行开始时间',
    `finish_time`         DATETIME      NULL     COMMENT '执行结束时间',
    PRIMARY KEY (`id`),
    KEY idx_task_id(`task_id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COMMENT='工单表': 
```

## Nacos

```yml
agents:
  filesystem:
    # SESSION/USER/AGENT/GLOBAL
    isolationScope: USER
    # 当上下文没有用户ID时使用的默认值
    anonymousUserId: "sys_uid"
  workspace: 
    localRootPath: ${user.home}/.se-agent
    shell: false
  # https://java.agentscope.io/v2/zh/docs/harness/compaction.md
  compaction:
    triggerMessages: 50
    triggerTokens: 0
    reserved: 20000
    keepMessages: 20
    keepTokens: -1
    keepTokensMin: 2000
    keepTokensMax: 8000
    keepTokensRatio: 0.25
    # 设置为default表示使用默认值
    summaryPrompt: "default"
    flushBeforeCompact: true
    offloadBeforeCompact: true
    prune:
      protectTokens: 40000
      protectTokens: 20000
      maxOutputChars: 2000
      excludedTools: 
        - "read_file"
        - "memory_search"
        - "memory_get"
        - "session_search"
    model:
      name: "gpt-oss-120b"
      stream: true
      temperature: 0.2
```