Official Recce plugins for Claude Code — bringing data validation into your dbt development workflow.
dbt developers modify models, run dbt run, and hope nothing breaks downstream. Recce plugins let Claude Code automatically detect what changed and validate the data impact — so you catch row count drops, schema drift, and query differences before they reach production.
| Plugin | Who it's for | Install |
|---|---|---|
| recce-quickstart | New Recce users getting started | /plugin install recce-quickstart@recce-claude-plugin |
| recce | dbt developers using Recce daily | /plugin install recce@recce-claude-plugin |
Guided onboarding for first-time users:
| Command | What it does |
|---|---|
/recce-setup |
Walks you through environment setup — installs dependencies, generates artifacts, starts the MCP server |
/recce-pr [url] |
Analyzes data changes in a pull request |
/recce-check [type] [selector] |
Runs a specific data validation (row-count, schema, profile, query-diff) |
/recce-ci |
Sets up Recce Cloud CI/CD for GitHub Actions |
Automated data review for daily development. Once installed:
- Claude auto-tracks which model files you edit
- After
dbt run, Claude suggests a data review based on tracked changes /recce-reviewvalidates impacted models and produces a risk-assessed summary
Step 1: Install Recce (see Recce installation guide)
Step 2: Add the marketplace
/plugin marketplace add DataRecce/recce-claude-plugin
Step 3: Install a plugin
/plugin install recce-quickstart@recce-claude-plugin
Start a new Claude Code session in your dbt project directory, then type /recce-setup.
| Problem | Solution |
|---|---|
| Plugin not showing up | Check /plugin → Installed tab. If missing, reinstall. |
| Plugin errors after install | Check /plugin → Errors tab. |
| Commands not available | Restart Claude Code — hooks and MCP tools activate on session start. |
For Recce-specific issues (MCP server, dbt connection, environment setup), see the Recce documentation.
MIT