Dbt

Please install the `dbt` MCP server into my current AI client (that's you).

Required prerequisites (do these first if not already done):
- **Install the dbt MCP package** — Add the MCP server package to your project or make it available so it can be run with npx. Run: `pnpm add -D @us-all/dbt-mcp`

Optional prerequisites:
- Install optional DQ backend dependency — Install the backend library only if you will query custom data-quality tables. Use @google-cloud/bigquery for BigQuery or pg for Postgres. Run: `pnpm add -D @google-cloud/bigquery OR pnpm add -D pg`

Canonical MCP server config (stdio transport):
- command: `npx`
- args: ["@us-all/dbt-mcp"]
- required environment variables:
  - `DBT_PROJECT_DIR`: Path to the dbt project root where dbt_project.yml lives. (example: `<path-to-dbt-project>`)
- optional environment variables:
  - `DBT_TARGET_DIR`: Path to the dbt target directory. Defaults to $DBT_PROJECT_DIR/target. (example: `<path-to-dbt-target-dir>`)
  - `DBT_RUN_HISTORY_DIR`: Optional directory containing archived run_results.json history. (example: `<path-to-run-history-dir>`)
  - `DQ_BACKEND`: Backend for custom DQ result tables: bigquery (default) or postgres. (example: `<bigquery-or-postgres>`)
  - `DQ_RESULTS_TABLE`: Fully qualified checks-results table. Required only for checks-based quality tools. (example: `<your-results-table>`)
  - `DQ_SCORE_TABLE`: Fully qualified score-daily table. Required for score-based quality tools. (example: `<your-score-table>`)
  - `GOOGLE_APPLICATION_CREDENTIALS`: Path to Google application credentials for BigQuery. ADC fallback is supported. (example: `<path-to-google-application-credentials-json>`)
  - `BQ_PROJECT_ID`: Explicit BigQuery project ID; otherwise inferred from ADC. (example: `<your-bigquery-project-id>`)
  - `PG_CONNECTION_STRING`: Postgres connection string when DQ_BACKEND=postgres. (example: `<your-postgres-connection-string>`)
  - `DQ_SCHEMA`: Schema preset for the quality category: generic (default) or us-all. (example: `<generic-or-us-all>`)
  - `DQ_COL_*`: Optional per-column overrides layered on top of DQ_SCHEMA. Includes variables such as DQ_COL_RUN_AT, DQ_COL_CHECK_TYPE, DQ_COL_STATUS, DQ_COL_DATASET, DQ_COL_TABLE_NAME, DQ_COL_SEVERITY, DQ_COL_FAILURE_COUNT, DQ_COL_MESSAGE, DQ_COL_CHECK_NAME, DQ_COL_SCORE_DATE, DQ_COL_SCOPE, and DQ_COL_TIER. (example: `<your-column-override>`)
  - `DQ_TIER1_TARGET_PCT`: Tier 1 SLA threshold used by dq-tier-status when no tier column is configured. Default is 99.5. (example: `<your-tier1-target-pct>`)
  - `DBT_SLA_CONFIG_PATH`: Path to an optional YAML file with tier_sla and dbt_sla blocks. (example: `<path-to-sla-config-yaml>`)
  - `DBT_ALLOW_WRITE`: Reserved for future write tools; none currently exist. (example: `<true-or-false>`)
  - `DBT_TOOLS`: Category allowlist toggle, for example to enable only dbt tools. (example: `<comma-separated-categories>`)
  - `DBT_DISABLE`: Category denylist toggle, for example to disable quality tools. (example: `<comma-separated-categories>`)

Note: Primary transport is stdio. The README also mentions an optional HTTP mode enabled by setting MCP_TRANSPORT=http, but it does not provide a canonical HTTP endpoint URL to configure, so stdio is the usable extracted configuration.

Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.