← Back to Skills Marketplace
diluke1600

grafana-insepction

by ken · GitHub ↗ · v0.0.3 · MIT-0
cross-platform ⚠ suspicious
331
Downloads
0
Stars
0
Active Installs
3
Versions
Install in OpenClaw
/install grafana-inspection
Description
Grafana 自动化巡检技能。支持浏览器截图 + API 数据巡检,多仪表盘批量巡检,自动发现仪表盘。
README (SKILL.md)

\r \r

Grafana 自动化巡检技能\r

\r

快速开始\r

\r

# 1. 配置\r
cd skills/grafana-inspector/scripts\r
cp config.example.json config.json\r
# 编辑 config.json\r
\r
# 2. 执行\r
python main.py config.json\r
\r
# 或在 OpenClaw 中输入\r
执行 Grafana 巡检\r
```\r
\r
## 配置说明\r
\r
```json\r
{\r
  "grafana_url": "http://localhost:3000",\r
  "api_key": "",\r
  "dashboard_uids": [],\r
  "auto_discover": true,\r
  "discover_limit": 10,\r
  "inspection_mode": "hybrid",\r
  "screenshot_dir": "./screenshots"\r
}\r
```\r
\r
| 参数 | 说明 | 必填 |\r
|------|------|------|\r
| `grafana_url` | Grafana 地址 | ✅ |\r
| `api_key` | API Key(API 巡检用) | ✅ |\r
| `dashboard_uids` | 仪表盘 UID 列表 | ❌ |\r
| `auto_discover` | 自动发现仪表盘 | ❌ |\r
| `discover_limit` | 最多巡检数量 | ❌ |\r
\r
## 使用方式\r
\r
### 方式 1:命令行\r
\r
```bash\r
cd skills/grafana-inspector/scripts\r
python inspect.py config.json\r
```\r
\r
### 方式 2:OpenClaw\r
\r
```\r
执行 Grafana 巡检\r
```\r
\r
## 输出\r
\r
- `inspection_*.json` - JSON 结果\r
- `inspection_*.md` - Markdown 报告\r
\r
## 获取 Dashboard UID\r
\r
1. 打开 Grafana 仪表盘\r
2. 查看 URL: `https://grafana/d/{UID}/{name}`\r
3. 复制 UID 部分\r
\r
## 获取 API Key\r
\r
1. Grafana → Configuration → API keys\r
2. 创建 Viewer 权限的 Key\r
3. 复制到配置\r
\r
## 文件结构\r
\r
```\r
grafana-inspector/\r
├── scripts/\r
│   ├── api_inspect.py    # API 巡检\r
│   ├── main.py        # 主脚本\r
│   └── config.json       # 配置\r
├── screenshots/          # 截图\r
├── SKILL.md             # 技能定义\r
└── README.md            # 说明文档\r
```\r
\r
## 故障排查\r
\r
### API 连接失败\r
- 检查 API Key 是否正确\r
- 验证网络连接\r
- 确认 Grafana 服务正常\r
- 检查飞书授权状态\r
Usage Guidance
This package appears to be a partly-implemented or poorly packaged Grafana inspector rather than something obviously malicious. Before installing or running it in a production environment: 1) review the code (api_inspect.py and main.py) to confirm there is no hidden network exfiltration or unexpected endpoints; 2) verify that the screenshot/browser functionality you expect actually exists (it does not in the visible code); 3) correct or clarify config file fields (dashboard_uids vs dashboard_uid, unused username/password) and the SKILL.md commands (inspect.py vs main.py); 4) run the tool with a read-only Viewer API key and in an isolated/test environment first; and 5) prefer a skill with an authoritative source or homepage — the package has no upstream/homepage and contains conflicting docs which increases the risk of mistakes or malicious edits. If you need this functionality, request the maintainer to provide a corrected release (consistent docs, implement browser mode or remove claims, and add a clear changelog/source).
Capability Analysis
Type: OpenClaw Skill Name: grafana-inspection Version: 0.0.3 The skill bundle is a legitimate tool for automated Grafana inspection, utilizing the Grafana API to check dashboard status, alerts, and datasource health. The Python scripts (api_inspect.py and main.py) follow standard practices for API interaction, using user-provided configuration for the Grafana URL and API key without any evidence of hardcoded credentials, data exfiltration, or malicious execution patterns.
Capability Assessment
Purpose & Capability
The README and SKILL.md claim a hybrid inspection (browser screenshots + API checks). The code (api_inspect.py and main.py) implements only API-based checks; there is no headless-browser, screenshot, or screenshot-processing code. Config examples include fields (username/password, wait_for_load_ms, screenshot_dir) that imply browser use but those fields are unused in the visible code. This mismatch means the delivered implementation does not match the advertised capability.
Instruction Scope
SKILL.md and README contain commands that don't match the files: SKILL.md suggests running 'python inspect.py config.json' but the repo has 'main.py' and 'api_inspect.py' (no inspect.py). Troubleshooting references '飞书授权状态' (Feishu) though no Feishu/Lark integration appears in code. The instructions tell the agent to copy/edit config.json and run the script which is expected, but they also reference non-existent or inconsistent components — granting the agent unclear discretion and risk of user confusion or misconfiguration.
Install Mechanism
There is no install spec and no external downloads or package installs declared. All code is bundled in the skill. From an install-mechanism perspective this is low risk (nothing fetched from arbitrary URLs).
Credentials
The skill requests no environment variables or platform credentials, which is proportional. It does require a Grafana API key via config, which is appropriate for the stated functionality. However, config files include unused credential fields (username/password) and inconsistent key names across example/config.json (dashboard_uids vs dashboard_uid), which is confusing and could lead to misconfiguration or accidentally placing sensitive credentials in the wrong file.
Persistence & Privilege
The skill does not request always:true, does not modify other skills, and has no declared system-level persistence. It writes local report files (inspection_*.json / *.md) in the working directory — expected behavior for a reporting tool.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install grafana-inspection
  3. After installation, invoke the skill by name or use /grafana-inspection
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.0.3
Changelog for grafana-inspection v0.0.3: - Major refactor: consolidated all inspection logic into new scripts; simplified skill interface and codebase. - Added support for hybrid inspection mode (browser screenshots + API data). - Introduced automatic dashboard discovery with configurable inspection limit. - Deprecated individual script modules (health check, screenshot, report); replaced with unified entry points. - Updated configuration structure; streamlined parameters for easier setup. - Skill documentation revised for concise usage and troubleshooting.
v0.0.2
- No code or documentation changes detected in this release. - Version bump only; all features and documentation remain unchanged.
v0.0.1
grafana-inspector 0.0.1 初始发布 - 首次发布支持自动化系统巡检,集成 Grafana API。 - 支持检查主机与中间件状态,捕获监控面板截图,计算健康评分,并生成 Markdown 格式报告。 - 可通过飞书 API 自动发送巡检报告至云文档。 - 提供命令行、关键词触发及定时任务三种运行方式。 - 内含详细配置说明和常见故障排查指南。
Metadata
Slug grafana-inspection
Version 0.0.3
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 3
Frequently Asked Questions

What is grafana-insepction?

Grafana 自动化巡检技能。支持浏览器截图 + API 数据巡检,多仪表盘批量巡检,自动发现仪表盘。 It is an AI Agent Skill for Claude Code / OpenClaw, with 331 downloads so far.

How do I install grafana-insepction?

Run "/install grafana-inspection" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is grafana-insepction free?

Yes, grafana-insepction is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does grafana-insepction support?

grafana-insepction is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created grafana-insepction?

It is built and maintained by ken (@diluke1600); the current version is v0.0.3.

💬 Comments