[MCP] Initialize metadata providers in stdio mode

[Christos Despotakis (despotak)]

Why make this change?

Closes #3783.

Summary of the linked issue: in every 2.1.x build, dab start --mcp-stdio registers entities but never infers their database objects, so every MCP tool call fails with Database object for entity '<name>' has not been inferred. while the identical config serves the same entity correctly over REST. 2.0.12 is unaffected.

version	published	--mcp-stdio	REST
2.0.12 (GA)	2026-08-20	PASS	PASS
2.1.0-rc	2026-08-11	FAIL	PASS
2.1.3-rc	2026-08-25	FAIL	PASS

Related: #3676, #3675 (the change this regressed from), and #3430 — see the note at the bottom.

What is this change?

Schema inference is reachable only through the ASP.NET Core startup path, which stdio mode skips:

1. Program.cs — StartEngine returns before the host is started:

   IHost host = CreateHostBuilder(args, runMcpStdio, mcpRole).Build();
   
   if (runMcpStdio)
   {
       return McpStdioHelper.RunMcpStdioHost(host);   // returns here
   }
   
   host.Run();   // normal web mode

2. McpStdioHelper.RunMcpStdioHost — initialises the tool registry and nothing else, so Startup.Configure never runs.
3. Startup.cs:822 — Configure is the only place that calls PerformOnConfigChangeAsync(app) (line 866).
4. Startup.cs:1431 — PerformOnConfigChangeAsync is the only caller of IMetadataProviderFactory.InitializeAsync().

So entity names reach the tool registry from config, while IMetadataProviderFactory is never initialised and no entity receives a database object. That is exactly the observed split: tools/list succeeds, every tool call fails.

This is a side effect of #3676 "Avoid starting web host in MCP stdio mode". That change was right — stdio mode should not bind an HTTP port — but PerformOnConfigChangeAsync did more than serve HTTP, and nothing took over its metadata-initialisation duty on the stdio path.

The fix is 10 lines: RunMcpStdioHost resolves IMetadataProviderFactory from DI and initialises it before registering tools. Program.cs is untouched, and #3676's behaviour is preserved — the existing assertions that StartAsync/StopAsync are never called still pass.

How was this tested?

• Integration Tests
• Unit Tests

RunMcpStdioHost_DoesNotStartWebHost gains a stub IMetadataProviderFactory and an assertion that InitializeAsync is called exactly once. Its original assertions are unchanged and still pass, so this cannot silently re-introduce the web host.

Also verified by hand against SQL Server, on main built from source with the pinned SDK (10.0.302), before and after the patch:

	before	after
describe_entities, 1-entity config	has not been inferred	success (1 entity)
describe_entities, 28-entity config	has not been inferred	success (28 entities)
read_records on a table	error	success, rows returned
REST GET /api/<entity>	HTTP 200	HTTP 200 (unchanged)

dotnet format --verify-no-changes on both touched files exits 0.

Sample Request(s)

Driving the stdio server by hand, so no MCP client is involved:

printf '%s\n' \
 '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"probe","version":"1"}}}' \
 '{"jsonrpc":"2.0","method":"notifications/initialized","params":{}}' \
 '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"describe_entities","arguments":{}}}' \
 | dab start --mcp-stdio

Before:

{"toolName":"describe_entities","status":"error",
 "error":{"type":"DataApiBuilderError",
          "message":"Database object for entity 'MyTable' has not been inferred."}}

After:

{"entities":[{"name":"MyTable","description":"","fields":[],"permissions":["READ"]}],
 "count":1,"status":"success"}

One note for reviewers

#3430 reports that stdio blocks the initialize response until introspection finishes (~17 s against a remote instance with 53 entities). Restoring inference on this path necessarily brings that latency back — it was only absent because inference was not happening at all. If you would prefer inference to run asynchronously after initialize returns, that is a larger change and I am happy to rework this accordingly; this PR deliberately restores correctness first.

---

Investigated and written with Claude Code (Νύξ) 🌑 — the version bracket, the REST/stdio control and the root-cause trace were worked out together. Reviewed and submitted by me.

[azure-pipelines]

Azure Pipelines:
There may be pipelines that require an authorized user to comment /azp run to run.

[Christos Despotakis (despotak)]

@microsoft-github-policy-service agree

[Copilot]

Pull request overview

Restores schema inference for MCP stdio mode by explicitly initializing the metadata providers on the stdio startup path (which bypasses ASP.NET Core Startup.Configure). This addresses the 2.1.x regression where tools were registered from config but all tool calls failed because database objects were never inferred.

Changes:

• Initialize IMetadataProviderFactory inside McpStdioHelper.RunMcpStdioHost before registering MCP tools.
• Extend the existing unit test to assert IMetadataProviderFactory.InitializeAsync() is invoked exactly once in stdio mode.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/Service/Utilities/McpStdioHelper.cs	Initializes metadata providers during stdio startup so entities have inferred database objects before tool calls.
src/Service.Tests/UnitTests/McpStdioHelperTests.cs	Adds a stub IMetadataProviderFactory and asserts InitializeAsync() is called once without starting the web host.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Souvik Ghosh (souvikghosh04)]

nit- just import using Azure.DataApiBuilder.Core.Services.MetadataProviders instead of calling the same multiple times.

[Souvik Ghosh (souvikghosh04)]

since this inside a try/finally that only disposes the host, so any initialization exception propagates raw out of RunMcpStdioHost (which otherwise returns bool). Consider wrapping it so a metadata-inference failure produces a clear, logged error over the stdio channel rather than an unhandled exception

[Copilot]

🟢 Approval recommended

The change is small, directly addresses the reported regression in stdio mode, and is covered by a focused unit test that preserves the existing “does not start web host” guarantees.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 0 new
• Review effort level: Lite