knowledge-work-data

---
name: knowledge-work-data
description: Use when you need to write SQL, explore datasets, generate insights, and build visualizations or interactive dashboards across any data warehouse or dialect.
source: https://github.com/anthropics/knowledge-work-plugins/tree/main/data
generated: 2026-07-02T18:42:12.336Z
category: concept
audience: engineers
---

## When to use

- Answering a data question end-to-end — from a quick lookup to a full analysis
- Profiling a dataset to understand its shape, quality, and patterns
- Writing optimized SQL for a specific dialect like Snowflake or BigQuery
- Building interactive HTML dashboards or publication-quality charts from results

## Key concepts

### sql-queries

SQL best practices across dialects, common query patterns, and performance optimization, producing dialect-specific optimized SQL.

### data-exploration

Data profiling, quality assessment, and pattern discovery — reporting row counts, null rates, cardinality, and unexpected values.

### data-visualization

Chart selection, Python viz code patterns, and design principles for publication-quality charts.

### interactive-dashboard-builder

HTML/JS dashboard construction with Chart.js, filters, and styling for interactive, shareable dashboards.

### statistical-analysis

Descriptive stats, trend analysis, outlier detection, and hypothesis testing to back insights with rigor.

### data-validation

Pre-delivery QA — methodology, accuracy, and bias checks plus sanity checks and documentation standards before sharing.

## API reference

```
/plugin marketplace add anthropics/knowledge-work-plugins
```

Add Anthropic's knowledge-work plugin marketplace (one-time).

```
/plugin marketplace add anthropics/knowledge-work-plugins
```

```
/plugin install data@knowledge-work-plugins
```

Install the Data Analyst plugin.

```
/plugin install data@knowledge-work-plugins
```

## Gotchas

- Primarily designed for Cowork, Anthropic's agentic desktop app, though it also runs in Claude Code
- Best with a data-warehouse MCP connection; without one you paste SQL results or upload CSV/Excel for analysis
- /write-query needs the SQL dialect specified to produce correctly optimized queries
- Run /validate before sharing — the plugin's QA catches methodology, accuracy, and bias issues
- Works with any SQL-compatible warehouse and dialect, so answers depend on you naming your stack correctly

---
Generated by SkillMake from https://github.com/anthropics/knowledge-work-plugins/tree/main/data on 2026-07-02T18:42:12.336Z.
Verify against source before relying on details.

File: ~/.claude/skills/knowledge-work-data/SKILL.md