cloudnet-analyze

cloudnet-analyze针对无线场景下终端体验差、网卡、无法接入 WiFi 等问题进行排障的技能

Security Scan

The skill's declared purpose (Wi‑Fi troubleshooting) matches what its instructions do (call Cloudnet MCP via the mcporter CLI using a CLOUDNET_API_KEY); minor inconsistencies and install-time risks are noted but do not contradict the stated purpose.

Purpose & Capability

The skill is described as a wireless troubleshooting tool and its runtime instructions call Cloudnet MCP diagnosis endpoints. Requiring a CLOUDNET_API_KEY and the mcporter CLI is coherent with that purpose. Requiring npm (to install mcporter) is explainable.

Instruction Scope

Instructions ask the user/agent to install mcporter globally (npm install -g mcporter), configure mcporter with an Authorization header using CLOUDNET_API_KEY, and run mcporter calls to Cloudnet endpoints. These actions are within the troubleshooting scope, but the SKILL.md references an optional env var (CLOUDNET_BASE_URL) that is used in commands but not declared in the metadata—this mismatch should be fixed. The instructions also tell the user to install an additional 'mcporter' skill via clawhub (ambiguous what that entails).

Install Mechanism

The skill is instruction-only (no install spec). However, it explicitly tells operators to run 'npm install -g mcporter', which will install and execute code from the npm registry with global privileges. That is a standard way to obtain the mcporter CLI but carries the usual npm-global risk (trust the package and source).

Credentials

Requesting CLOUDNET_API_KEY as the primary credential is proportional and expected for a tool that calls Cloudnet APIs. The SKILL.md also uses CLOUDNET_BASE_URL (optional) but that env var is not declared in the metadata; this is an inconsistency to address. No unrelated credentials are requested.

Persistence & Privilege

The skill does not request 'always: true' or other elevated platform privileges. It is instruction-only and relies on external CLI tooling; it does not itself persist credentials or modify other skills according to the provided files.

Assessment

This skill appears to do what it says: it uses a Cloudnet API key and the mcporter CLI to run device/Shop diagnostics. Before installing or running it: (1) Only provide a CLOUDNET_API_KEY you trust the skill to use; an API key allows network calls to your Cloudnet instance. (2) Be cautious about running 'npm install -g mcporter' — global npm installs execute third‑party code; verify the mcporter package and its publisher. (3) Clarify CLOUDNET_BASE_URL usage (it is referenced but not declared) and verify the 'clawhub' step and any additional 'mcporter' skill source. (4) If you prefer lower risk, install mcporter in an isolated environment or container and limit the API key scope. If you want me to, I can list exact questions to ask the vendor or produce safer install steps (containerized install, or obtaining mcporter from a specific verified URL).

针对无线场景下终端体验差、网卡、无法接入 WiFi 等问题进行排障。

触发条件

用户询问关于无线终端连接问题，例如：

• "XX 场所的 XX 终端上网很慢"
• "XX 场所的 XX 设备连不上 WiFi"
• "XX 场所的 XX 用户反馈网络卡"

前置环境检查

安装mcportercli支持及skill支持

• mcporter: npm install -g mcporter
• 然后再通过clawhub安装mcporter技能

配置MCP连接参数

• CLOUDNET_API_KEY 必须. 需要用户提供Cloudnet管理平台的API KEY，可通过Cloudnet管理平台（网络管理=>设置=>开放平台）获取
• 可选: CLOUDNET_BASE_URL Cloudnet管理平台地址. 默认使用https://oasis.h3c.com
• 执行```bash mcporter config add cloudnet-mcp ${CLOUDNET_BASE_URL}/mcp-server/api/sse --header Authorization="Bearer ${CLOUDNET_API_KEY}"

## 排障步骤

### 第一步：提取关键信息

从用户问题中提取以下信息：

| 信息 | 说明 | 示例 |
|------|------|------|
| **场所名** | 必填，问题发生的场所 | "总部办公室"、"XX门店" |
| **终端信息** | 必填，MAC 地址或终端用户名 | MAC: `xxxx-xxxx-xxxx` 或用户名: `zhangsan` |
| **故障时间** | 可选，用户未指定则默认当前时间 | "2026-03-24 10:00:00" |

**重要**：如果场所名和终端信息未提取到，必须让用户补充完整后才能继续下一步。

### 第二步：查询场所 ID

调用 `cloudnet-mcp.getallshopsanddevofuser` 获取用户下所有场所，找到场所名对应的场所 ID。

```bash
mcporter call cloudnet-mcp.getallshopsanddevofuser

第三步：执行终端诊断

根据提取的终端信息（MAC 或用户名），调用 executeStaDiagnosis 进行诊断：

参数说明：

• MAC: 终端 MAC 地址，格式 xxxx-xxxx-xxxx
• shopId: 场所 ID（来自第二步），需要转为字符串类型
• faultTime: 故障时间，格式 yyyy-MM-dd HH:mm:ss，用户未指定则使用当前时间
• timezone: 用户时区，默认 Asia/Shanghai

mcporter call cloudnet-mcp.executeStaDiagnosis MAC:"xxxx-xxxx-xxxx" shopId:"场所ID" faultTime:"2026-03-24 10:00:00" timezone:"Asia/Shanghai"

第四步：分析诊断结果

诊断返回后会包含：

• 终端连接状态
• 信号强度
• 接入 AP 信息
• 可能的根因
• 优化建议

结合诊断结果回答用户问题，并给出具体的解决建议。

输出格式

完成排障后，向用户清晰说明：

1. 诊断结果摘要
2. 问题根因分析
3. 建议的解决措施