Language: 日本語English

datahub-skills Plugin

Author
DataHub
Category
Database
Topics
Data Engineering & Analytics · Databases & Storage · Code Quality, Review & Testing
First seen
2026-07-09
Last confirmed
2026-07-13
Explanation last updated
2026-07-12

The explanation below is AI-generated. Please verify it against the sources.

datahub-skills is a plugin that packages a set of agent skills for working with DataHub, covering catalog search, metadata enrichment, lineage tracing, data quality management, connection setup, and connector development (planning and PR review against 22 standards), according to the README. It is designed to work with Claude Code, Cortex Code, Cursor, Codex, Copilot, Gemini CLI, Windsurf, and other Agent Skills-compatible tools. Some capabilities, such as parallel multi-agent review, slash commands, and SessionStart hooks, are available only on Claude Code, while other platforms use natural-language equivalents for the same tasks. The homepage describes the underlying DataHub product as an 'AI & Data Context Management Platform' that unifies discovery, governance, observability, and lineage across a data estate.

Overview

According to the homepage, DataHub is an AI & Data Context Management Platform that provides AI-powered discovery, governance, and observability to unify an organization's data estate for data quality, compliance, and AI readiness. It is also positioned as a context management platform for AI agents, with MCP-native integration mentioned for tools such as Cortex, Genie, Cursor, Claude, LangChain, ADK, and Crew AI.

What you can do with datahub-skills

  • Search the DataHub catalog via keyword search, filtered browse, column-name search, structured property queries, and multi-step question answering (per the README)
  • Enrich metadata — descriptions, tags, glossary terms, ownership, deprecation — with a before/after plan requiring approval
  • Explore lineage: trace upstream sources and downstream consumers, run impact analysis, and map cross-platform data flows
  • Manage data quality: create/run assertions (freshness, volume, SQL, field, schema), set up AI-inferred assertions, raise/resolve incidents, and configure notification subscriptions (Open Source vs Cloud capabilities differ)
  • Set up the DataHub CLI, configure authentication, verify connectivity, and set default scopes/profiles
  • Plan a new connector through a four-step workflow: classify the source, research it, generate a _PLANNING.md, and get sign-off before coding
  • Review connector code against 22 standards, with parallel five-agent review available on Claude Code
  • Load all 22 connector standards into context before starting connector work

Sources

Original description (English)

DataHub development and interaction toolkit with connector planning, PR review, catalog search, metadata enrichment, lineage tracing, data quality management, and connection setup skills

History of datahub-skills

Back to list