DataHub Skills
DataHub skills for searching the catalog, enriching metadata, tracing lineage, managing data quality and building connectors
- 46 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The enrich and quality skills change descriptions, tags and owners in the catalog and create assertions and incidents
- The skills act with the rights of the DataHub token configured during setup
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/datahub-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The enrich and quality skills change descriptions, tags and owners in the catalog and create assertions and incidents
- The skills act with the rights of the DataHub token configured during setup
This is third-party code. Review the repository files before installing.
What it does
A skill pack from the DataHub project for working with a data catalog. Five skills touch the live catalog: entity search, metadata enrichment (descriptions, tags, glossary terms, owners), lineage tracing with impact analysis, data quality management (assertions, incidents, notification subscriptions) and connection setup. Before any change the skill shows a before-and-after plan and waits for approval. The second group is for connector authors. Planning takes four steps: source type, research, a `_PLANNING.md` file and sign-off before code. Review checks code against 22 standards, and in Claude Code it runs five agents in parallel. The evals skill runs the answer checks saved in DataHub.
Who it is for. Data teams that keep their catalog in DataHub, and developers of connectors for it.
Good fit when
- You need to find revenue tables in Snowflake and learn who owns them
- Before changing the orders table you must see which dashboards and pipelines depend on it
- Columns with personal data get a PII tag in a batch, and the edit plan goes for approval
- A table needs a freshness assertion, and failing assertions are gathered into an incident
- A new connector is being written, for example for ClickHouse, and an approved plan is needed before code
Not a fit when
- DataHub is not deployed: the skills need access to a catalog through the CLI or an MCP server
- You need full data quality management in DataHub Core: it is available in Cloud, while Open Source offers diagnostics only
- The catalog lives in another tool: the skills know only DataHub
Example request
Show which dashboards will break if I change the orders table.Limitations
Working with the catalog needs the DataHub CLI and configured authentication, which the setup skill handles. Full data quality management works in DataHub Cloud, while Open Source offers diagnostics. Slash commands, session start hooks and parallel review by five agents exist only in Claude Code, and other agents run the same steps one by one. The evals skill needs the DataHub Cloud CLI (`pip install 'acryl-datahub-cloud[datahub-evals]'`) and a DataHub MCP server.
How to disable. In Claude Code the plugin is removed with `claude plugin uninstall datahub-skills`. Skills installed through the Skills CLI are deleted by hand from `.claude/skills/` or `.agents/skills/`.
FAQ
How do I install DataHub Skills in Claude Code?
As a plugin: `claude plugin install datahub-skills`, which adds hooks, slash commands and multi-agent dispatch. Without the plugin use `npx skills add datahub-project/datahub-skills -a claude-code`, and the skills land in `.claude/skills/`.
How do DataHub Skills differ from DataHub MCP Server?
The MCP server provides tools for accessing the catalog. The skills describe workflows: how to search, how to change metadata with approval, how to plan a connector. The setup skill helps connect both the CLI and MCP.
Do the skills work in Cursor and Codex?
Yes. `npx skills add datahub-project/datahub-skills -a cursor` or `-a codex` puts the skills into `.agents/skills/`. There are no slash commands there, so the same actions are triggered by a plain request.
What does the connector review check?
Code is checked against 22 standards: 11 core ones (API, SQL, tests, performance, registration and others) and 11 by source type, from data warehouses to streaming platforms. In Claude Code five agents look at the code in parallel.
What does the setup skill do?
It installs the DataHub CLI, configures authentication, checks connectivity and sets default scopes and profiles for the other skills, for example a focus on Snowflake in the Finance domain.
Discussion
No questions or reviews yet.
Related
Skills31 skills for Claude Code and Codex for the analyst workflow: data checks, cohorts, funnels, A/B tests, dashboards, summaries
High-Stakes Analytics Lab
High-Stakes Analytics & Decision Lab
An evidence-gated analytics skill for high-stakes decisions: profiles data, picks a method the evidence supports, can answer with a data request
A template for data scientists: AGENTS.md, CLAUDE.md and skills for Python, SQL, Polars and model diagnostics.
16 Claude Code skills for data engineers: source profiling, failure investigation, change review and documentation as web pages.
A DataHub MCP server: an agent searches tables and columns by metadata, traces lineage and finds analysts' SQL queries
Probabl skills for machine learning experiments: a scikit-learn and skrub pipeline, evaluation, audit and a link to Skore Hub
README in short
The repository splits the skills into catalog work and connector development, with sample requests for each. Next come installation for each agent, a table of differences between Claude Code and the rest, and lists of slash commands, agents and standards.
README badge
Original SKILL.md
---
name: datahub-search
description: |
Use this skill when the user wants to search the DataHub catalog, discover entities, answer ad-hoc questions about their data, find datasets, or browse by platform or domain. Triggers on: "search DataHub", "find datasets", "who owns X", "what tables contain PII", "what columns does X have", or any request to search, discover, browse, or answer one-off questions about DataHub metadata. For lineage questions ("what feeds into X"), use `/datahub-lineage`. For systematic audits ("how complete is our metadata"), use `/datahub-audit`.
user-invocable: true
min-cli-version: 1.4.0
allowed-tools: Bash(datahub *)
---
# DataHub Search
You are an expert DataHub catalog navigator and metadata analyst. Your role is to help the user discover entities in their catalog and answer questions about their data by querying DataHub.
This skill operates in two modes:
- **Discovery mode:** Find, browse, and list entities ("find revenue tables in Snowflake")
- **Question mode:** Answer analytical questions by querying and reasoning over metadata ("who owns the revenue pipeline?")
---
## Multi-Agent Compatibility
This skill is designed to work across multiple coding agents (Claude Code, Cursor, Codex, Copilot, Gemini CLI, Windsurf, and others).
**What works everywhere:**
- The full search and question-answering workflow
- Both discovery and question modes
- Search, browse, and entity retrieval via MCP tools or DataHub CLI
- Result formatting and answer synthesis
**Claude Code-specific features** (other agents can safely ignore these):
- `allowed-tools` in the YAML frontmatter above
- `Task(subagent_type="datahub-skills:metadata-searcher")` for delegated search — **fallback instructions are provided inline** for agents that cannot dispatch sub-agents
**Reference file paths:** Shared references are in `../shared-references/` relative to this skill's directory. Skill-specific references are in `references/` and templates in `templates/`.
---
## Not This Skill
| If the user wants to... | Use this instead |
| -------------------------------------------------------------- | ------------------ |
| Explore lineage, upstream/downstream, impact analysis | `/datahub-lineage` |