Data Analyst Skills

README

# OpenClaw Data Analyst Skills

A collection of data analyst skills for [OpenClaw](https://github.com/openclaw/openclaw), adapted from Anthropic's [knowledge-work-plugins](https://github.com/anthropics/knowledge-work-plugins/tree/main/data).

> **Original source**: [anthropics/knowledge-work-plugins](https://github.com/anthropics/knowledge-work-plugins) — Apache 2.0 License

---

## What's Included

| Skill | Description | Invocable |
|---|---|---|
| `analyze` | Answer data questions — from quick lookups to full analyses | ✅ |
| `explore-data` | Profile and explore a dataset: shape, quality, patterns | ✅ |
| `write-query` | Write optimized SQL for any dialect | ✅ |
| `create-viz` | Create publication-quality visualizations with Python | ✅ |
| `build-dashboard` | Build self-contained interactive HTML dashboards | ✅ |
| `validate-data` | QA an analysis before sharing with stakeholders | ✅ |
| `data-context-extractor` | Generate a company-specific data analysis skill from your warehouse schema | ✅ |
| `load-business-context` | Load a business knowledge file (data dictionary, field definitions, metric glossary) so all subsequent analysis uses correct business semantics | ✅ |
| `sql-queries` | SQL best practices across dialects (used internally by other skills) | 🔧 |
| `data-visualization` | Chart selection, Python viz patterns, design principles (used internally) | 🔧 |
| `statistical-analysis` | Descriptive stats, trend analysis, outlier detection (used internally) | 🔧 |

---

## Installation

### Option 1: Install individual skills

Copy the skill folder you want into your OpenClaw skills directory:

```bash
# Install all skills
cp -r analyze explore-data write-query create-viz build-dashboard \
      validate-data data-context-extractor \
      sql-queries data-visualization statistical-analysis \
      ~/.openclaw/skills/

# Or install just one
cp -r analyze ~/.openclaw/skills/
```

### Option 2: Clone and install

```bash
git clone https://github.com/chencanbiu04/openclaw-data-analyst-skills.git
cd openclaw-data-analyst-skills

# Install all
cp -r */ ~/.openclaw/skills/
```

---

## Usage

Once installed, the skills are automatically activated when relevant. You can also explicitly ask:

- **"帮我分析一下这个数据..."** → triggers `analyze`
- **"探查一下这张表"** → triggers `explore-data`
- **"帮我写一个 SQL 查询..."** → triggers `write-query`
- **"生成一个图表"** → triggers `create-viz`
- **"做一个 Dashboard"** → triggers `build-dashboard`
- **"帮我 QA 这个分析"** → triggers `validate-data`

---

## Connecting a Data Warehouse (Optional)

These skills work best when connected to a data warehouse via MCP. Supported platforms:

| Platform | MCP Server |
|---|---|
| BigQuery | `https://bigquery.googleapis.com/mcp` |
| Snowflake | Your Snowflake MCP endpoint |
| Databricks | Your Databricks MCP endpoint |
| Amplitude | `https://mcp.amplitude.com/mcp` |
| Hex | `https://app.hex.tech/mcp` |

**Without a data warehouse**: You can still paste CSV/SQL results or upload files, and the skills will analyze them directly.

---

## What Changed from the Original

This repo adapts the original Claude Code plugin format to OpenClaw:

- Removed `argument-hint` and `user-invocable` frontmatter fields (not supported by OpenClaw)
- Replaced `~~data warehouse` tool references with generic MCP tool language
- Updated `CONNECTORS.md` references to inline guidance
- All 10 skills are independent installable units

---

## License

Apache 2.0 — same as the original [knowledge-work-plugins](https://github.com/anthropics/knowledge-work-plugins).