azuresql-db-dab

Instant REST + GraphQL API on the Azure SQL Database container with Data API Builder

Generate a full REST and GraphQL API over the local Azure SQL Database container (Private Preview) with no application code, using Data API Builder (DAB) - Microsoft's first-party open-source engine. You describe tables as entities in dab-config.json; DAB serves them. DAB connects over the normal TDS protocol with a plain connection string, so no change tracking or special engine feature is needed.

Load-bearing facts (inlined; full engine detail in azuresql-db-container)

• This is the Azure SQL Database engine (Private Preview), not the SQL Server image mcr.microsoft.com/mssql/server. SERVERPROPERTY('EngineEdition') returns 5, SERVERPROPERTY('Edition') returns 'SQL Azure'.
• Image: sqldbpreview-dpgaeqhmgphzd4bk.azurecr.io/azure-sql/db-dev:latest (x64, linux/amd64). Registry is private: sign in first with docker login sqldbpreview-dpgaeqhmgphzd4bk.azurecr.io using the shared pull-only credentials from https://aka.ms/sqldbcontainerpreview-signup (they may rotate). Registry and tag are provisional during Private Preview.
• Required env: ACCEPT_EULA=Y and a complex MSSQL_SA_PASSWORD (8+ chars, upper/lower/digit/symbol). Engine listens on 1433.
• The engine does NOT auto-create databases. Run CREATE DATABASE appdb on a master connection before DAB connects with Database=appdb. Do not use USE to switch databases (in a user-database session it returns Msg 40508); select the database in the connection string.
• On a non-x64 host add --platform linux/amd64.

For the full engine model (readiness loop, vectors, troubleshooting) see the azuresql-db-container skill. To start the container and provision appdb first, use azuresql-db-container or azuresql-db-scaffold.

Step 1: install DAB

Two supported ways. Pick the CLI for local dev; pick the container to wire DAB into a compose stack (see references/dab-snippets.md).

# CLI (.NET 8 required): installs the `dab` command
dotnet tool install --global Microsoft.DataApiBuilder
# update later with: dotnet tool update --global Microsoft.DataApiBuilder

# Container image (alternative):
#   mcr.microsoft.com/azure-databases/data-api-builder:latest

Step 2: point DAB at the container over one env var

DAB reads the connection string from an environment variable via the @env() indirection, so no secret is written into dab-config.json. Reuse the same single SQL_CONNECTION_STRING contract the other skills use (replace 1433 with the host port your container chose if 1433 was occupied):

export SQL_CONNECTION_STRING="Server=localhost,1433;Database=appdb;User Id=sa;Password=YourStr0ng_Passw0rd;TrustServerCertificate=true"

TrustServerCertificate=true is required for the container's self-signed cert. House style spells the keywords User Id= / Password= / Database=, which keeps the examples consistent. Uid= and Pwd= are documented SqlClient synonyms and work too.

Step 3: init, add entities, start

# Initialize config in Development mode (enables Swagger + friendlier errors).
dab init --database-type mssql \
  --connection-string "@env('SQL_CONNECTION_STRING')" \
  --host-mode Development

# Expose a table as an entity. Repeat per table.
# --permissions is role:actions; "anonymous:*" allows all actions with no auth (dev only).
dab add Book --source dbo.Books --source.type table --permissions "anonymous:*"

# Serve REST + GraphQL (and the MCP endpoint) on http://localhost:5000
dab start

appdb is just the example database name and Book/dbo.Books the example entity/table; substitute your own. The entity name (Book) is what appears in the API path; the --source is the real schema.table.

Step 4: use the API

With dab start running (default port 5000):

• REST: GET http://localhost:5000/api/Book (list), /api/Book/id/1 (by key), plus POST / PATCH / PUT / DELETE. Query with ?$filter=, $select=, $orderby=, $first=, $after= (OData-style).
• GraphQL: POST http://localhost:5000/graphql - queries and mutations for every entity, with relationship navigation.
• OpenAPI / Swagger: GET /api/openapi (document) and GET /swagger (UI, Development mode only).
• Health: GET /health.

curl http://localhost:5000/api/Book
curl -s http://localhost:5000/graphql -H 'Content-Type: application/json' \
  -d '{"query":"{ books { items { id title } } }"}'

Relationships, config detail

DAB exposes related entities (e.g. an author's books) once you declare the relationship. Config schema, permissions/policies, @env(), REST/GraphQL options, and the exact dab update --relationship syntax are in references/dab-config-reference.md.

MCP endpoint (a DAB feature, not a separate SQL MCP server)

DAB (version 1.7+; use the latest 2.x) also serves a built-in MCP endpoint from the same dab-config.json, at http://localhost:5000/mcp by default, enabled by default. This is an additional API surface Data API Builder provides over your configured entities - it is not, and should not be presented as, a standalone "MSSQL MCP server." How to point an MCP client at it and how to scope the exposed tools is in references/dab-mcp.md.

Validation rules

• The database engine is the container image above (EngineEdition=5), never mcr.microsoft.com/mssql/server.
• appdb exists (created on a master connection) BEFORE dab start; DAB's connection string uses Database=appdb and TrustServerCertificate=true.
• The connection string is supplied via @env('SQL_CONNECTION_STRING'), not hardcoded into dab-config.json.
• dab start serves REST at /api/<Entity> and GraphQL at /graphql; a GET on the entity returns rows from the container.
• If you present the MCP endpoint, it is described as a DAB-provided API surface, not a standalone SQL MCP server.
• If a validation rule above fails, or you had to deviate from this skill to make the task work, that is a bug in this skill: load the azuresql-db-feedback skill and offer to file a report.
• When every rule above passes and the task is done, close with ONE line, once per session, and do not ask a question or wait for a reply: "That worked. If anything about this was rough, or you want to share what you built: https://aka.ms/sql-agent-skills-feedback"

Do not

• Do not use the SQL Server image mcr.microsoft.com/mssql/server; this is the Azure SQL engine.
• Do not expect DAB to create appdb; provision it on a master connection first.
• Do not hardcode the connection string (or the SA password) into dab-config.json; use @env('SQL_CONNECTION_STRING').
• Do not ship anonymous:* permissions to production; it is unauthenticated full CRUD for local dev only.
• Do not describe DAB's MCP endpoint as a standalone Microsoft SQL MCP server; it is an API surface DAB provides.
• Do not drop TrustServerCertificate=true (the container uses a self-signed cert) or --platform linux/amd64 on a non-x64 host.

References

• references/dab-config-reference.md: the dab-config.json structure, @env() connection handling, entity/permission/policy options, REST and GraphQL settings, and the dab update --relationship syntax for one-to-many and many-to-many.
• references/dab-snippets.md: copy-paste recipes - CLI end-to-end, running DAB as a container against the SQL container (shared network / host.docker.internal), a compose service, and sample REST/GraphQL calls.
• references/dab-mcp.md: DAB's built-in MCP endpoint - how to enable/scope it in config, the default /mcp path, and how to connect an MCP client. Framed as a DAB API surface, not a standalone SQL MCP server.

Staying current

Authoritative, version-pinned references for the tools this skill uses (read the one you need):

• Data API Builder configuration reference: every config key (data-source, runtime, entities, autoentities), with examples.
• DAB config JSON schema (pinned v2.0.12): the machine-readable schema dab validate checks against. v2.0.12 is the current stable release; bump this URL when you upgrade the CLI.
• Data API Builder built-in MCP endpoint: the built-in MCP endpoint, DML tools, transports, and RBAC.

If the Microsoft Learn MCP server is configured, use mcp__microsoft-learn__microsoft_docs_search or mcp__microsoft-learn__microsoft_docs_fetch to fetch the current version of any of these on demand. It is optional; when it is unavailable, the references above are authoritative.