This one walks through a smart fallback chain to visualize your dbt DAG as a Mermaid diagram. It'll try MCP tools first for the most accurate lineage, then fall back to parsing manifest.json if available, and finally resort to direct code parsing if needed. The color coding is thoughtful: sources are blue, staging bronze, marts gold, tests yellow, with the focal model highlighted in purple. One nice touch is that it explicitly treats all parsed content as untrusted, which matters when you're pulling data from SQL comments and YAML files that could contain weird stuff. Most useful when you need to explain dependencies to your team or debug why a model isn't rebuilding when you expect it to.
npx -y skills add dbt-labs/dbt-agent-skills --skill creating-mermaid-dbt-dag --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
metabase/metabase
github/awesome-copilot
UKGovernmentBEIS/inspect_evals
addyosmani/agent-skills