test: add output schema, sampling constraint, roots error, and version-rejection interaction tests

Author: maxisbey

tests/interaction/_requirements.py

@@ -116,6 +116,13 @@ class Requirement:
             "requested version with its own latest supported version rather than an error."
         ),
     ),
+    "lifecycle:initialize:protocol-version:client-rejects": Requirement(
+        source=f"{SPEC_BASE_URL}/basic/lifecycle#version-negotiation",
+        behavior=(
+            "A client that receives an initialize response carrying a protocol version it does not "
+            "support fails initialization with an error rather than proceeding with the session."
+        ),
+    ),
     "lifecycle:requests-before-initialized": Requirement(
         source=f"{SPEC_BASE_URL}/basic/lifecycle#initialization",
         behavior="A request sent before the initialization handshake completes is rejected with an error.",
@@ -154,6 +161,19 @@ class Requirement:
             "A cancellation notification referencing an unknown or already-completed request is ignored without error."
         ),
     ),
+    "cancellation:server-to-client": Requirement(
+        source=f"{SPEC_BASE_URL}/basic/utilities/cancellation#behavior-requirements",
+        behavior=(
+            "A server that abandons an in-flight server-initiated request (sampling, elicitation, roots) "
+            "cancels it, and the client stops processing the cancelled request."
+        ),
+        deferred=(
+            "Not expressible through the public API: abandoning a server-side send_request emits no "
+            "cancellation notification (the same sender-side gap recorded on timeouts:per-request), and "
+            "the client could not act on one anyway because client callbacks run inline in the receive "
+            "loop, so a cancellation would not even be read until the callback had already finished."
+        ),
+    ),
     # ═══════════════════════════════════════════════════════════════════════════
     # Progress
     # ═══════════════════════════════════════════════════════════════════════════
@@ -325,6 +345,13 @@ class Requirement:
             "with the validation failure described in content), not a protocol error."
         ),
     ),
+    "tools:call:output-schema-validation": Requirement(
+        source=f"{SPEC_BASE_URL}/server/tools#tool-result",
+        behavior=(
+            "A tool result whose structuredContent does not conform to the tool's declared outputSchema "
+            "is rejected by the client: the call raises instead of returning the invalid result."
+        ),
+    ),
     # ═══════════════════════════════════════════════════════════════════════════
     # Completion
     # ═══════════════════════════════════════════════════════════════════════════
@@ -473,6 +500,17 @@ class Requirement:
         source=f"{SPEC_BASE_URL}/server/prompts#error-handling",
         behavior="prompts/get for an unknown prompt name returns a JSON-RPC error.",
     ),
+    "prompts:get:missing-arguments": Requirement(
+        source=f"{SPEC_BASE_URL}/server/prompts#error-handling",
+        behavior="prompts/get with a required argument missing returns a JSON-RPC error.",
+        divergence=Divergence(
+            note=(
+                "The spec says missing required arguments are answered with -32602 Invalid params; "
+                "MCPServer's prompt renderer raises a plain ValueError before the prompt function runs, "
+                "which the low-level server converts to error code 0 with the exception text as the message."
+            ),
+        ),
+    ),
     # ═══════════════════════════════════════════════════════════════════════════
     # Sampling (server → client)
     # ═══════════════════════════════════════════════════════════════════════════
@@ -487,7 +525,7 @@ class Requirement:
         source=f"{SPEC_BASE_URL}/client/sampling#creating-messages",
         behavior=(
             "The sampling parameters supplied by the server (messages, maxTokens, systemPrompt, "
-            "modelPreferences, temperature, stopSequences) reach the client callback intact."
+            "modelPreferences, temperature, stopSequences, includeContext) reach the client callback intact."
         ),
     ),
     "sampling:create-message:image-content": Requirement(
@@ -501,6 +539,13 @@ class Requirement:
             "by the server before anything reaches the wire, with an Invalid params error."
         ),
     ),
+    "sampling:create-message:tools:message-constraints": Requirement(
+        source=f"{SPEC_BASE_URL}/client/sampling#message-content-constraints",
+        behavior=(
+            "A sampling request whose messages violate the tool_use/tool_result pairing rules is rejected "
+            "by the server-side validator before anything reaches the wire."
+        ),
+    ),
     "sampling:create-message:tools:round-trip": Requirement(
         source=f"{SPEC_BASE_URL}/client/sampling#sampling-with-tools",
         behavior=(
@@ -587,6 +632,17 @@ class Requirement:
             ),
         ),
     ),
+    "elicitation:url:not-supported": Requirement(
+        source=f"{SPEC_BASE_URL}/client/elicitation#capabilities",
+        behavior=(
+            "A URL-mode elicitation to a client that declared only form-mode support is rejected with an "
+            "Invalid params error."
+        ),
+        deferred=(
+            "Not expressible through the public API: a Client with an elicitation callback always declares "
+            "both the form and url sub-capabilities, so a form-only client cannot be constructed."
+        ),
+    ),
     # ═══════════════════════════════════════════════════════════════════════════
     # Roots (server → client)
     # ═══════════════════════════════════════════════════════════════════════════
@@ -614,6 +670,10 @@ class Requirement:
             ),
         ),
     ),
+    "roots:list:client-error": Requirement(
+        source=f"{SPEC_BASE_URL}/client/roots#error-handling",
+        behavior="A roots callback that answers with an error surfaces to the requesting handler as an MCPError.",
+    ),
     "roots:list-changed": Requirement(
         source=f"{SPEC_BASE_URL}/client/roots#root-list-changes",
         behavior="A roots/list_changed notification sent by the client is delivered to the server's handler.",
@@ -671,6 +731,22 @@ class Requirement:
             "tests/shared/test_streamable_http.py."
         ),
     ),
+    "transport:streamable-http:resumability": Requirement(
+        source=f"{SPEC_BASE_URL}/basic/transports#streamable-http",
+        behavior="A client that reconnects with Last-Event-ID receives the events it missed.",
+        deferred=(
+            "Replay requires dropping and re-establishing the SSE connection, which the in-process ASGI "
+            "client cannot express. Covered over a real socket by tests/shared/test_streamable_http.py."
+        ),
+    ),
+    "transport:streamable-http:origin-validation": Requirement(
+        source=f"{SPEC_BASE_URL}/basic/transports#streamable-http",
+        behavior="Requests with a disallowed Origin or Host header are rejected before reaching the session.",
+        deferred=(
+            "The in-process fixture disables DNS-rebinding protection because no network attack surface "
+            "exists in-process. Covered by tests/server/test_streamable_http_security.py."
+        ),
+    ),
     "transport:stdio": Requirement(
         source=f"{SPEC_BASE_URL}/basic/transports#stdio",
         behavior="The interaction round trip works over a stdio subprocess.",

tests/interaction/lowlevel/test_initialize.py

@@ -1,8 +1,10 @@
 """Initialization handshake against the low-level Server, driven through the public Client API.
 
-The last two tests drive a bare ClientSession over an InMemoryTransport instead: Client always
+The later tests drive a bare ClientSession over an InMemoryTransport instead: Client always
 performs the full handshake with the latest protocol version, so skipping initialization or
-requesting a different version can only be expressed one level down.
+requesting a different version can only be expressed one level down. The final test goes one step
+further and plays the server's side of the wire by hand, because no real Server can be made to
+answer initialize with an unsupported protocol version.
 """
 
 import anyio
@@ -14,6 +16,8 @@
 from mcp.client._memory import InMemoryTransport
 from mcp.client.client import Client
 from mcp.server import Server, ServerRequestContext
+from mcp.shared.memory import create_client_server_memory_streams
+from mcp.shared.message import SessionMessage
 from mcp.types import (
     INVALID_PARAMS,
     CallToolResult,
@@ -26,6 +30,8 @@
     InitializeRequest,
     InitializeRequestParams,
     InitializeResult,
+    JSONRPCRequest,
+    JSONRPCResponse,
     ListToolsRequest,
     ListToolsResult,
     LoggingCapability,
@@ -195,10 +201,11 @@ async def call_tool(ctx: ServerRequestContext, params: types.CallToolRequestPara
             for name, value in (
                 ("sampling", capabilities.sampling),
                 ("elicitation", capabilities.elicitation),
-                ("roots", capabilities.roots),
             )
             if value is not None
         ]
+        if capabilities.roots is not None:
+            declared.append(f"roots(list_changed={capabilities.roots.list_changed})")
         return CallToolResult(content=[TextContent(text=",".join(declared) or "none")])
 
     async def list_roots(context: ClientRequestContext) -> types.ListRootsResult:
@@ -213,7 +220,7 @@ async def list_roots(context: ClientRequestContext) -> types.ListRootsResult:
 
     async with Client(server, list_roots_callback=list_roots) as client:
         result = await client.call_tool("abilities", {})
-    assert result == snapshot(CallToolResult(content=[TextContent(text="roots")]))
+    assert result == snapshot(CallToolResult(content=[TextContent(text="roots(list_changed=True)")]))
 
 
 @requirement("lifecycle:requests-before-initialized")
@@ -275,3 +282,48 @@ def initialize_request(protocol_version: str) -> InitializeRequest:
             with anyio.fail_after(5):
                 result = await session.send_request(initialize_request("1999-01-01"), InitializeResult)
     assert result.protocol_version == snapshot("2025-11-25")
+
+
+@requirement("lifecycle:initialize:protocol-version:client-rejects")
+async def test_unsupported_server_protocol_version_fails_initialization() -> None:
+    """An initialize response carrying a protocol version the client does not support fails initialization.
+
+    A real Server only ever answers with a version it supports, so this test alone plays the
+    server's side of the wire by hand: it reads the initialize request off the raw stream and
+    answers it with a hand-built result. Reserve this pattern for behaviour no real server can
+    be made to produce.
+    """
+    async with create_client_server_memory_streams() as (client_streams, server_streams):
+        client_read, client_write = client_streams
+        server_read, server_write = server_streams
+
+        async def scripted_server() -> None:
+            message = await server_read.receive()
+            assert isinstance(message, SessionMessage)
+            request = message.message
+            assert isinstance(request, JSONRPCRequest)
+            assert request.method == "initialize"
+            result = InitializeResult(
+                protocol_version="1991-08-06",
+                capabilities=ServerCapabilities(),
+                server_info=Implementation(name="relic", version="0.0.1"),
+            )
+            await server_write.send(
+                SessionMessage(
+                    JSONRPCResponse(
+                        jsonrpc="2.0",
+                        id=request.id,
+                        # Serialized exactly as a real server serializes results onto the wire.
+                        result=result.model_dump(by_alias=True, mode="json", exclude_none=True),
+                    )
+                )
+            )
+
+        async with anyio.create_task_group() as tg:
+            tg.start_soon(scripted_server)
+            async with ClientSession(client_read, client_write) as session:
+                with anyio.fail_after(5):
+                    with pytest.raises(RuntimeError) as exc_info:
+                        await session.initialize()
+
+            assert str(exc_info.value) == snapshot("Unsupported protocol version from the server: 1991-08-06")

tests/interaction/lowlevel/test_prompts.py

@@ -10,6 +10,7 @@
     INVALID_PARAMS,
     ErrorData,
     GetPromptResult,
+    Icon,
     ListPromptsResult,
     Prompt,
     PromptArgument,
@@ -35,6 +36,7 @@ async def list_prompts(ctx: ServerRequestContext, params: types.PaginatedRequest
                         PromptArgument(name="code", description="The code to review.", required=True),
                         PromptArgument(name="style_guide", description="Optional style guide to apply."),
                     ],
+                    icons=[Icon(src="https://example.com/review.png", mime_type="image/png", sizes=["48x48"])],
                 ),
                 Prompt(name="daily_standup"),
             ]
@@ -55,6 +57,7 @@ async def list_prompts(ctx: ServerRequestContext, params: types.PaginatedRequest
                         PromptArgument(name="code", description="The code to review.", required=True),
                         PromptArgument(name="style_guide", description="Optional style guide to apply."),
                     ],
+                    icons=[Icon(src="https://example.com/review.png", mime_type="image/png", sizes=["48x48"])],
                 ),
                 Prompt(name="daily_standup"),
             ]

tests/interaction/lowlevel/test_resources.py

@@ -14,6 +14,7 @@
     CallToolResult,
     EmptyResult,
     ErrorData,
+    Icon,
     ListResourcesResult,
     ListResourceTemplatesResult,
     ReadResourceResult,
@@ -48,6 +49,7 @@ async def list_resources(
                     mime_type="text/markdown",
                     size=1024,
                     annotations=Annotations(audience=["user", "assistant"], priority=0.8),
+                    icons=[Icon(src="https://example.com/readme.png", mime_type="image/png", sizes=["48x48"])],
                 ),
             ]
         )
@@ -69,6 +71,7 @@ async def list_resources(
                     mime_type="text/markdown",
                     size=1024,
                     annotations=Annotations(audience=["user", "assistant"], priority=0.8),
+                    icons=[Icon(src="https://example.com/readme.png", mime_type="image/png", sizes=["48x48"])],
                 ),
             ]
         )
@@ -159,6 +162,7 @@ async def list_resource_templates(
                     title="Service logs",
                     description="One day of logs for one service.",
                     mime_type="text/plain",
+                    icons=[Icon(src="https://example.com/logs.png", mime_type="image/png", sizes=["48x48"])],
                 ),
             ]
         )
@@ -178,6 +182,7 @@ async def list_resource_templates(
                     title="Service logs",
                     description="One day of logs for one service.",
                     mime_type="text/plain",
+                    icons=[Icon(src="https://example.com/logs.png", mime_type="image/png", sizes=["48x48"])],
                 ),
             ]
         )

tests/interaction/lowlevel/test_roots.py

@@ -9,7 +9,7 @@
 from mcp.client import ClientRequestContext
 from mcp.client.client import Client
 from mcp.server import Server, ServerRequestContext
-from mcp.types import CallToolResult, ListRootsResult, Root, TextContent
+from mcp.types import INTERNAL_ERROR, CallToolResult, ErrorData, ListRootsResult, Root, TextContent
 from tests.interaction._requirements import requirement
 
 pytestmark = pytest.mark.anyio
@@ -107,6 +107,37 @@ async def call_tool(ctx: ServerRequestContext, params: types.CallToolRequestPara
     assert result == snapshot(CallToolResult(content=[TextContent(text="-32600: List roots not supported")]))
 
 
+@requirement("roots:list:client-error")
+async def test_list_roots_callback_error_surfaces_to_the_handler() -> None:
+    """A roots callback that answers with an error fails the roots/list request with that exact error.
+
+    The callback's code and message reach the requesting handler verbatim as an MCPError.
+    """
+
+    async def list_tools(
+        ctx: ServerRequestContext, params: types.PaginatedRequestParams | None
+    ) -> types.ListToolsResult:
+        return types.ListToolsResult(tools=[types.Tool(name="show_roots", input_schema={"type": "object"})])
+
+    async def call_tool(ctx: ServerRequestContext, params: types.CallToolRequestParams) -> CallToolResult:
+        assert params.name == "show_roots"
+        try:
+            await ctx.session.list_roots()
+        except MCPError as exc:
+            return CallToolResult(content=[TextContent(text=f"{exc.error.code}: {exc.error.message}")])
+        raise NotImplementedError  # the callback always answers with an error
+
+    server = Server("rooted", on_list_tools=list_tools, on_call_tool=call_tool)
+
+    async def list_roots(context: ClientRequestContext) -> ErrorData:
+        return ErrorData(code=INTERNAL_ERROR, message="roots provider crashed")
+
+    async with Client(server, list_roots_callback=list_roots) as client:
+        result = await client.call_tool("show_roots", {})
+
+    assert result == snapshot(CallToolResult(content=[TextContent(text="-32603: roots provider crashed")]))
+
+
 @requirement("roots:list-changed")
 async def test_roots_list_changed_reaches_server_handler() -> None:
     """A roots/list_changed notification from the client is delivered to the server's handler.