← 返回 Skills 市场
zsmj1994

finebi-skill

作者 刘荣歆 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
143
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install finebi-skill
功能描述
一个用于 OpenClaw 的 FineBI 技能,支持 npx 直接运行和插件体系,提供数据分析和可视化功能
使用说明 (SKILL.md)

\r \r

FineBI 技能\r

\r

技能介绍\r

\r 该技能集成了 FineBI,能够通过自然语言实现数据分析、生成报告以及查询数据集等功能。\r \r

工具列表\r

\r

get-dataset-info\r

通过名称查询 FineBI 数据集信息。\r \r 参数:\r

  • dataset (string, 必填): 要查询的数据集名称\r
  • filters (object, 选填): 用于过滤的键值对\r \r

preview-dataset-data\r

获取数据集的数据预览(限制最大不超过 10w 条数据)。\r \r 参数:\r

  • dataset (string, 必填): 数据表的原始名称\r
  • page (number, 选填): 页码,默认为 1\r
  • size (number, 选填): 每页数据量,最大支持 100000,默认为 100\r \r

search-datasets\r

根据关键字分⻚搜索 FineBI 数据集。\r \r 参数:\r

  • keyword (string, 必填): 要搜索的关键字\r
  • page (number, 选填): 页码,默认为 1\r \r

export-dashboard-excel\r

将仪表板导出为 Excel 格式。\r \r 参数:\r

  • reportId (string, 必填): 仪表板 ID\r
  • widgetId (string, 选填): 组件 ID,用于导出特定的组件\r \r

export-dashboard-pdf\r

将仪表板导出为 PDF 格式。响应中会返回文件路径,和文件名。\r \r 参数:\r

  • reportId (string, 必填): 仪表板 ID\r \r

export-dashboard-image\r

将仪表板导出为 PNG 格式的图片。\r \r 参数:\r

  • reportId (string, 必填): 仪表板 ID\r \r

get-dashboard-user-info\r

获取当前用户信息及其创建的仪表板。\r \r 参数: 无\r \r \r

get-dashboards-by-subject\r

获取特定主题下的仪表板列表。\r \r 参数:\r

  • subjectId (string, 必填): 主题 ID\r \r \r ---\r \r

🧩 扩展子技能(需手动加载)\r

\r 本技能目录下包含 4 个紧密关联的子技能,提供高级自动化场景支持。\r \r ⚠️ 重要:子技能不会自动触发,需要在使用时手动读取对应 SKILL.md 文件。\r \r | 子技能 | 功能描述 | 触发场景 | 文件路径 |\r |--------|---------|---------|----------|\r | bi-report-to-doc | 将 FineBI 仪表盘数据自动生成飞书云文档分析报告 | 经营周报/月报/异常分析报告自动生成 | bi-report-to-doc/SKILL.md |\r | skill-bi-briefing | 将 FineBI 核心指标压缩成飞书群播报卡片,支持定时任务 | 日报/周报/月报/异常播报定时群同步 | skill-bi-briefing/SKILL.md |\r | skill-bi-skill-alert-to-task | 监控 FineBI 指标阈值,自动创建飞书任务并通知负责人 | 库存/销售额/订单金额等阈值告警 | skill-bi-skill-alert-to-task/SKILL.md |\r | skill-bi-to-bitable | 将 FineBI 数据集同步到飞书多维表格,支持追加/覆盖/增量模式 | 数据归档/定时同步/二次协作 | skill-bi-to-bitable/SKILL.md |\r \r

如何加载子技能\r

\r 使用时通过 read 工具读取对应子技能的 SKILL.md:\r \r

{\r
  "action": "read",\r
  "path": "~/.openclaw/workspace/skills/finebi-skill/bi-report-to-doc/SKILL.md"\r
}\r
```\r
\r
或在主技能中根据用户意图自动推荐:\r
\r
> 检测到您需要生成经营分析报告,推荐加载 `bi-report-to-doc` 子技能。\r
\r
### 子技能快速索引\r
\r
#### 1. bi-report-to-doc(报告生成)\r
**典型场景**:\r
- "把经营看板整理成飞书文档"\r
- "生成上月异常分析报告"\r
- "数据刷新后自动写周报"\r
\r
**核心动作链**:\r
```\r
search-dashboards → get-dashboard-detail → export-dashboard-pdf/image \r
→ 结构化提取 → create-doc → write-content\r
```\r
\r
#### 2. skill-bi-briefing(群播报)\r
**典型场景**:\r
- "每天早上 9 点在群里发数据日报"\r
- "设置定时播报任务"\r
- "先发一条测试卡片看看"\r
\r
**核心动作链**:\r
```\r
search-dashboards → get-dashboard-detail → preview-dataset-data \r
→ 结构化播报对象 → send-message-card / create-cron-job\r
```\r
\r
#### 3. skill-bi-skill-alert-to-task(阈值告警)\r
**典型场景**:\r
- "库存低于 20 就创建任务通知负责人"\r
- "销售额环比下降超过 30% 自动报警"\r
- "监控订单金额异常"\r
\r
**核心动作链**:\r
```\r
search-datasets → get-dataset-info → preview-dataset-data \r
→ 阈值判定 → create-task → send-private-message\r
```\r
\r
#### 4. skill-bi-to-bitable(数据同步)\r
**典型场景**:\r
- "把 FineBI 数据同步到飞书多维表格"\r
- "每周自动追加新数据"\r
- "按主键增量更新"\r
\r
**核心动作链**:\r
```\r
search-datasets → get-dataset-info → preview-dataset-data \r
→ get-bitable-meta → 字段映射 → append/overwrite/upsert\r
```\r
\r
---\r
\r
## 常见工作流 (Common Workflows)\r
\r
### 使用/获取数据集数据\r
如果需要提取某个表的数据,标准的工作流程是:\r
1. **第一步**:先使用 `search-datasets` 命令,根据表名查到表信息,在返回结果中获取对应表的 ID(通常为 `name` 字段,即表的原始名称)。\r
2. **第二步**:调用 `preview-dataset-data` 命令,传入查到的 ID,从而获取该数据集的具体数据记录。\r
\r
### 导出 FineBI 仪表板为 PDF,然后分析导出的 PDF 文件生成报告 ⚠️\r
\r
⚠️ **重要警告**:分析 PDF 时**必须且只能**使用 OpenClaw 内置的 `pdf` 工具!禁止使用任何其他方式读取 PDF 内容!\r
\r
**工作流程**:\r
1. **导出 PDF**:调用 `export-dashboard-pdf` 命令,传入 `reportId` 参数,获取导出的 PDF 文件路径。\r
2. **复制到工作空间**:如果 PDF 在临时目录(如 `/tmp/` 或 `AppData/Local/Temp/`),必须先复制到工作空间目录(通常是 workspace)。\r
   - Windows: `Copy-Item \x3C源路径> -Destination "C:\Users\dailer\.openclaw\workspace\"`\r
   - Linux/Mac: `cp \x3C源路径> ~/workspace/`\r
3. **使用内置 pdf 工具分析** 📌:调用 `pdf` 工具读取 PDF 文件(传入本地文件路径),提取图表和数据内容。\r
4. **生成报告**:基于 PDF 内容生成 Markdown 格式的分析报告,包含核心主题、关键数据、趋势洞察和业务建议,最后发送给用户。\r
\r
\r
## 迁移的工具列表 (Migrated Tools)\r
\r
### `report-list`\r
Migrated tool: reportList\r
\r
**参数:**\r
- `subjectId`: 必填\r
\r
### `report-info`\r
Migrated tool: reportInfo\r
\r
**参数:**\r
- `reportIds`: 必填\r
\r
### `report-consanguinity`\r
Migrated tool: reportConsanguinity\r
\r
**参数:**\r
- `reportId`: 必填\r
\r
### `report-check`\r
Migrated tool: reportCheck\r
\r
**参数:**\r
- 无\r
\r
### `package-list`\r
Migrated tool: packageList\r
\r
**参数:**\r
- 无\r
\r
### `package-structure`\r
Migrated tool: packageStructure\r
\r
**参数:**\r
- `packId`: 必填\r
- `withTables`: 必填\r
\r
### `data-folders`\r
Migrated tool: dataFolders\r
\r
**参数:**\r
- 无\r
\r
### `data-folder-tree`\r
Migrated tool: dataFolderTree\r
\r
**参数:**\r
- `folderId`: 必填\r
\r
### `data-table-preview`\r
Migrated tool: dataTablePreview\r
\r
**参数:**\r
- `tableName`: 必填\r
\r
### `data-table-structure`\r
Migrated tool: dataTableStructure\r
\r
**参数:**\r
- `tableName`: 必填\r
\r
### `data-model`\r
Migrated tool: dataModel\r
\r
**参数:**\r
- `modelId`: 必填\r
\r
### `data-query`\r
Migrated tool: dataQuery\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `data-preview`\r
Migrated tool: dataPreview\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `data-search-tables`\r
Migrated tool: dataSearchTables\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `data-search-fields`\r
Migrated tool: dataSearchFields\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `data-field-data`\r
Migrated tool: dataFieldData\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `data-field-range`\r
Migrated tool: dataFieldRange\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `spider-status`\r
Migrated tool: spiderStatus\r
\r
**参数:**\r
- `taskInstanceId`: 必填\r
\r
### `subject-folders`\r
Migrated tool: subjectFolders\r
\r
**参数:**\r
- 无\r
\r
### `subject-tree-root`\r
Migrated tool: subjectTreeRoot\r
\r
**参数:**\r
- 无\r
\r
### `subject-folder`\r
Migrated tool: subjectFolder\r
\r
**参数:**\r
- `folderId`: 必填\r
\r
### `subject-tree`\r
Migrated tool: subjectTree\r
\r
**参数:**\r
- `folderId`: 必填\r
\r
### `subject-crumb`\r
Migrated tool: subjectCrumb\r
\r
**参数:**\r
- `folderId`: 必填\r
\r
### `subject-content`\r
Migrated tool: subjectContent\r
\r
**参数:**\r
- `subjectId`: 必填\r
\r
### `subject-reports`\r
Migrated tool: subjectReports\r
\r
**参数:**\r
- `subjectId`: 必填\r
\r
### `subject-get`\r
Migrated tool: subjectGet\r
\r
**参数:**\r
- `subjectIds`: 必填\r
\r
### `subject-search`\r
Migrated tool: subjectSearch\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `subject-groups-search`\r
Migrated tool: subjectGroupsSearch\r
\r
**参数:**\r
- `body`: 必填\r
\r
### `subject-consanguinity`\r
Migrated tool: subjectConsanguinity\r
\r
**参数:**\r
- `subjectId`: 必填\r
\r
安全使用建议
This skill appears coherent for integrating FineBI with reporting and Feishu workflows, but review before enabling: 1) Only provide FineBI credentials (use a least-privileged account or API user). 2) Confirm your OpenClaw environment has node and any finebi-cli the platform expects. 3) Be aware the skill may copy exported PDFs/images from temporary locations into the agent workspace — ensure you are comfortable with the agent reading files placed there and that paths are limited to the exported artifacts. 4) Platform actions the skill calls (create-task, send-message-card, create-doc, bitable writes) depend on the host/platform connectors and permissions; ensure those integrations are configured and scoped appropriately. 5) Because this is instruction-only (no shipped code), the runtime behavior depends entirely on the platform's tool implementations — review audit logs and test with non-production data and limited credentials first.
功能分析
Type: OpenClaw Skill Name: finebi-skill Version: 1.0.0 The finebi-skill bundle is a legitimate integration for automating FineBI data analysis and reporting via Feishu (Lark). It defines a structured set of tools and sub-skills (e.g., bi-report-to-doc, skill-bi-briefing) with detailed markdown instructions (SKILL.md) that implement strict Standard Operating Procedures (SOPs) to prevent AI hallucination and ensure data integrity. While the instructions use terms like 'Strong Instruction Injection Zone,' these are defensive prompt engineering techniques designed to enforce 'Hard Constraints' and prevent the agent from fabricating data or skipping validation steps. The required environment variables (FINEBI_USERNAME/PASSWORD) and file operations are consistent with the stated purpose of BI automation.
能力评估
Purpose & Capability
Name/description (FineBI integration, data analysis, export, sync) align with the declared requirements: node binary and FineBI credentials (FINEBI_BASE_URL, FINEBI_USERNAME, FINEBI_PASSWORD). There are no unrelated env vars or binaries requested.
Instruction Scope
The SKILL.md contains comprehensive, low‑freedom SOPs that restrict actions to FineBI data flows (search, preview, export) and downstream platform actions (create-task, send-message-card, create-doc, bitable writes). It instructs copying exported PDFs from temp directories into the workspace (examples using cp / Copy-Item) and to use the platform's built-in pdf tool for analysis only. This file-system copying is reasonable for processing exported artifacts but means the agent will operate on local file paths (examples reference user workspace paths). Verify that copying is limited to the expected exported files and not used to read arbitrary files.
Install Mechanism
Instruction-only skill with no install spec or downloaded code. Lowest installer risk; it relies on an existing node binary (reasonable if finebi-cli is node-based).
Credentials
Requested env vars (FINEBI_BASE_URL, FINEBI_USERNAME, FINEBI_PASSWORD) are appropriate and proportional for accessing a FineBI instance. No extra unrelated secret/env var requests are present. The skill references downstream actions to send messages/create tasks/create docs (Feishu) but does not request Feishu credentials — those likely come from platform connectors, which is expected.
Persistence & Privilege
always is false and the skill is user-invocable. There is no install step that persists code or modifies other skills; no elevated persistence requested.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install finebi-skill
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /finebi-skill 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
- Initial release of finebi-openclaw-skill (v0.1.2). - Provides FineBI data analysis and visualization tools for use with OpenClaw, supporting npx execution and plugin integration. - Includes commands for dataset query, dashboard export (Excel, PDF, image), user/dashboard info, and more. - Lists 4 subskills for advanced scenarios: report-to-doc, briefing, alert-to-task, and bi-to-bitable, with instructions for manual loading. - Details common workflows and best practices for data extraction and PDF dashboard analysis. - Documents a full list of migrated data/report/subject management tools.
元数据
Slug finebi-skill
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

finebi-skill 是什么?

一个用于 OpenClaw 的 FineBI 技能,支持 npx 直接运行和插件体系,提供数据分析和可视化功能. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 143 次。

如何安装 finebi-skill?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install finebi-skill」即可一键安装,无需额外配置。

finebi-skill 是免费的吗?

是的,finebi-skill 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

finebi-skill 支持哪些平台?

finebi-skill 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 finebi-skill?

由 刘荣歆(@zsmj1994)开发并维护,当前版本 v1.0.0。

💬 留言讨论