@modelcontextprotocol/server-everything 2026.1.14 → 2026.7.4

package/dist/docs/features.md

@@ -19,10 +19,15 @@
 - `get-structured-content` (tools/get-structured-content.ts): Demonstrates structured responses. Accepts `location` input and returns both backward‑compatible `content` (a `text` block containing JSON) and `structuredContent` validated by an `outputSchema` (temperature, conditions, humidity).
 - `get-sum` (tools/get-sum.ts): For two numbers `a` and `b` calculates and returns their sum. Uses Zod to validate inputs.
 - `get-tiny-image` (tools/get-tiny-image.ts): Returns a tiny PNG MCP logo as an `image` content item with brief descriptive text before and after.
-- `trigger-long-running-operation` (tools/trigger-trigger-long-running-operation.ts): Simulates a multi-step operation over a given `duration` and number of `steps`; reports progress via `notifications/progress` when a `progressToken` is provided by the client.
+- `trigger-long-running-operation` (tools/trigger-long-running-operation.ts): Simulates a multi-step operation over a given `duration` and number of `steps`; reports progress via `notifications/progress` when a `progressToken` is provided by the client.
 - `toggle-simulated-logging` (tools/toggle-simulated-logging.ts): Starts or stops simulated, random‑leveled logging for the invoking session. Respects the client’s selected minimum logging level.
 - `toggle-subscriber-updates` (tools/toggle-subscriber-updates.ts): Starts or stops simulated resource update notifications for URIs the invoking session has subscribed to.
-- `trigger-sampling-request` (tools/trigger-sampling-request.ts): Issues a `sampling/createMessage` request to the client/LLM using provided `prompt` and optional generation controls; returns the LLM’s response payload.
+- `trigger-elicitation-request` (tools/trigger-elicitation-request.ts): Issues an `elicitation/create` request using form-mode fields (strings, numbers, booleans, enums, and format validation) and returns the resulting action/content.
+- `trigger-url-elicitation` (tools/trigger-url-elicitation.ts): Issues an `elicitation/create` request in URL mode (`mode: "url"`) with an `elicitationId`, or throws MCP error `-32042` (`UrlElicitationRequiredError`) when `errorPath=true`. On the error path the prerequisite elicitation it returns points at a different URL than the failing request (`https://modelcontextprotocol.io`); when the client satisfies it and retries the same call, the retry ignores `errorPath` and proceeds via the request path, so the client does not loop on the same error. The retry marker is one-shot per `(session, url, elicitationId)`: it is cleared on the recognized retry, so re-running the error path with identical arguments without an intervening prerequisite is treated as a retry and proceeds. Requires client capability `elicitation.url`.
+- `trigger-sampling-request` (tools/trigger-sampling-request.ts): Issues a `sampling/createMessage` request to the client/LLM using provided `prompt` and optional generation controls; returns the LLM's response payload.
+- `simulate-research-query` (tools/simulate-research-query.ts): Demonstrates MCP Tasks (SEP-1686) with a simulated multi-stage research operation. Accepts `topic` and `ambiguous` parameters. Returns a task that progresses through stages with status updates. If `ambiguous` is true and client supports elicitation, sends an elicitation request directly to gather clarification before completing.
+- `trigger-sampling-request-async` (tools/trigger-sampling-request-async.ts): Demonstrates bidirectional tasks where the server sends a sampling request that the client executes as a background task. Server polls for status and retrieves the LLM result when complete. Requires client to support `tasks.requests.sampling.createMessage`.
+- `trigger-elicitation-request-async` (tools/trigger-elicitation-request-async.ts): Demonstrates bidirectional tasks where the server sends an elicitation request that the client executes as a background task. Server polls while waiting for user input. Requires client to support `tasks.requests.elicitation.create`.
 
 ## Prompts
 
@@ -50,3 +55,51 @@
 - Simulated logging is available but off by default.
 - Use the `toggle-simulated-logging` tool to start/stop periodic log messages of varying levels (debug, info, notice, warning, error, critical, alert, emergency) per session.
 - Clients can control the minimum level they receive via the standard MCP `logging/setLevel` request.
+
+## Tasks (SEP-1686)
+
+The server advertises support for MCP Tasks, enabling long-running operations with status tracking:
+
+- **Capabilities advertised**: `tasks.list`, `tasks.cancel`, `tasks.requests.tools.call`
+- **Task Store**: Uses `InMemoryTaskStore` from SDK experimental for task lifecycle management
+- **Message Queue**: Uses `InMemoryTaskMessageQueue` for task-related messaging
+
+### Task Lifecycle
+
+1. Client calls `tools/call` with `task: true` parameter
+2. Server returns `CreateTaskResult` with `taskId` instead of immediate result
+3. Client polls `tasks/get` to check status and receive `statusMessage` updates
+4. When status is `completed`, client calls `tasks/result` to retrieve the final result
+
+### Task Statuses
+
+- `working`: Task is actively processing
+- `input_required`: Task needs additional input (server sends elicitation request directly)
+- `completed`: Task finished successfully
+- `failed`: Task encountered an error
+- `cancelled`: Task was cancelled by client
+
+### Demo Tools
+
+**Server-side tasks (client calls server):**
+Use the `simulate-research-query` tool to exercise the full task lifecycle. Set `ambiguous: true` to trigger elicitation - the server will send an `elicitation/create` request directly and await the response before completing.
+
+**Client-side tasks (server calls client):**
+Use `trigger-sampling-request-async` or `trigger-elicitation-request-async` to demonstrate bidirectional tasks where the server sends requests that the client executes as background tasks. These require the client to advertise `tasks.requests.sampling.createMessage` or `tasks.requests.elicitation.create` capabilities respectively.
+
+### Bidirectional Task Flow
+
+MCP Tasks are bidirectional - both server and client can be task executors:
+
+| Direction        | Request Type             | Task Executor | Demo Tool                           |
+| ---------------- | ------------------------ | ------------- | ----------------------------------- |
+| Client -> Server | `tools/call`             | Server        | `simulate-research-query`           |
+| Server -> Client | `sampling/createMessage` | Client        | `trigger-sampling-request-async`    |
+| Server -> Client | `elicitation/create`     | Client        | `trigger-elicitation-request-async` |
+
+For client-side tasks:
+
+1. Server sends request with task metadata (e.g., `params.task.ttl`)
+2. Client creates task and returns `CreateTaskResult` with `taskId`
+3. Server polls `tasks/get` for status updates
+4. When complete, server calls `tasks/result` to retrieve the result

package/dist/docs/structure.md

@@ -48,11 +48,15 @@ src/everything
      │   ├── get-sum.ts
      │   ├── get-tiny-image.ts
      │   ├── gzip-file-as-resource.ts
+     │   ├── simulate-research-query.ts
      │   ├── toggle-simulated-logging.ts
      │   ├── toggle-subscriber-updates.ts
      │   ├── trigger-elicitation-request.ts
+     │   ├── trigger-elicitation-request-async.ts
      │   ├── trigger-long-running-operation.ts
-     │   └── trigger-sampling-request.ts
+     │   ├── trigger-sampling-request.ts
+     │   ├── trigger-sampling-request-async.ts
+     │   └── trigger-url-elicitation.ts
      └── transports
          ├── sse.ts
          ├── stdio.ts
@@ -82,8 +86,8 @@ src/everything
 
 - `architecture.md`
   - This document.
-- `server-instructions.md`
-  - Human‑readable instructions intended to be passed to the client/LLM as for guidance on server use. Loaded by the server at startup and returned in the "initialize" exchange.
+- `instructions.md`
+  - Human‑readable instructions intended to be passed to the client/LLM as guidance on server use. Loaded by the server at startup and returned in the initialize exchange.
 
 ### `prompts/`
 
@@ -129,7 +133,7 @@ src/everything
 - `echo.ts`
   - Registers an `echo` tool that takes a message and returns `Echo: {message}`.
 - `get-annotated-message.ts`
-  - Registers an `annotated-message` tool which demonstrates annotated content items by emitting a primary `text` message with `annotations` that vary by `messageType` (`"error" | "success" | "debug"`), and optionally includes an annotated `image` (tiny PNG) when `includeImage` is true.
+  - Registers a `get-annotated-message` tool which demonstrates content-level annotations. Emits a primary `text` message with content `annotations` (`priority`, `audience`) that vary by `messageType` (`"error" | "success" | "debug"`), and optionally includes an annotated `image` (tiny PNG) when `includeImage` is true. All tools in this server include tool-level annotations (`readOnlyHint`, `destructiveHint`, `idempotentHint`, `openWorldHint`).
 - `get-env.ts`
   - Registers a `get-env` tool that returns the current process environment variables as formatted JSON text; useful for debugging configuration.
 - `get-resource-links.ts`
@@ -147,18 +151,26 @@ src/everything
     - `GZIP_MAX_FETCH_SIZE` (bytes, default 10 MiB)
     - `GZIP_MAX_FETCH_TIME_MILLIS` (ms, default 30000)
     - `GZIP_ALLOWED_DOMAINS` (comma-separated allowlist; empty means all domains allowed)
+- `simulate-research-query.ts`
+  - Registers a `simulate-research-query` task-based tool that demonstrates the MCP Tasks feature (SEP-1686). Simulates a multi-stage research operation with progress updates. If the query is marked as ambiguous and the client supports elicitation, it pauses mid-execution to request clarification via `elicitation/create`. Uses `server.experimental.tasks.registerToolTask()` with `execution: { taskSupport: "required" }`.
 - `trigger-elicitation-request.ts`
   - Registers a `trigger-elicitation-request` tool that sends an `elicitation/create` request to the client/LLM and returns the elicitation result.
+- `trigger-url-elicitation.ts`
+  - Registers a `trigger-url-elicitation` tool that either sends an out-of-band URL-mode `elicitation/create` request (`mode: "url"`) including an `elicitationId` (request path) or throws `UrlElicitationRequiredError` (`-32042`) for client-handled URL elicitation (error path). On the error path the carried prerequisite elicitation points at a different URL than the failing one (`https://modelcontextprotocol.io`), and when the client satisfies it and retries the same call, the retry ignores `errorPath` and proceeds via the request path — so the client does not loop on the same error.
+- `trigger-elicitation-request-async.ts`
+  - Registers a `trigger-elicitation-request-async` tool that demonstrates bidirectional MCP tasks for elicitation. Sends an elicitation request with task metadata, then polls the client's `tasks/get` endpoint for completion status before fetching the final result.
 - `trigger-sampling-request.ts`
   - Registers a `trigger-sampling-request` tool that sends a `sampling/createMessage` request to the client/LLM and returns the sampling result.
+- `trigger-sampling-request-async.ts`
+  - Registers a `trigger-sampling-request-async` tool that demonstrates bidirectional MCP tasks for sampling. Sends a sampling request with task metadata, then polls the client's `tasks/get` endpoint for completion status before fetching the final result.
 - `get-structured-content.ts`
   - Registers a `get-structured-content` tool that demonstrates structuredContent block responses.
 - `get-sum.ts`
-  - Registers an `get-sum` tool with a Zod input schema that sums two numbers `a` and `b` and returns the result.
+  - Registers a `get-sum` tool with a Zod input schema that sums two numbers `a` and `b` and returns the result.
 - `get-tiny-image.ts`
   - Registers a `get-tiny-image` tool, which returns a tiny PNG MCP logo as an `image` content item, along with surrounding descriptive `text` items.
 - `trigger-long-running-operation.ts`
-  - Registers a `long-running-operation` tool that simulates a long-running task over a specified `duration` (seconds) and number of `steps`; emits `notifications/progress` updates when the client supplies a `progressToken`.
+  - Registers a `trigger-long-running-operation` tool that simulates a long-running task over a specified `duration` (seconds) and number of `steps`; emits `notifications/progress` updates when the client supplies a `progressToken`.
 - `toggle-simulated-logging.ts`
   - Registers a `toggle-simulated-logging` tool, which starts or stops simulated logging for the invoking session.
 - `toggle-subscriber-updates.ts`

package/dist/resources/session.js

@@ -1,3 +1,9 @@
+/**
+ * Tracks registered session resources by URI to allow updating/removing on re-registration.
+ * This prevents "Resource already registered" errors when a tool creates a resource
+ * with the same URI multiple times during a session.
+ */
+const registeredResources = new Map();
 /**
  * Generates a session-scoped resource URI string based on the provided resource name.
  *
@@ -34,11 +40,19 @@ export const registerSessionResource = (server, resource, type, payload) => {
             mimeType,
             blob: payload,
         };
+    // Check if a resource with this URI is already registered and remove it
+    const existingResource = registeredResources.get(uri);
+    if (existingResource) {
+        existingResource.remove();
+        registeredResources.delete(uri);
+    }
     // Register file resource
-    server.registerResource(name, uri, { mimeType, description, title, annotations, icons, _meta }, async (uri) => {
+    const registeredResource = server.registerResource(name, uri, { mimeType, description, title, annotations, icons, _meta }, async () => {
         return {
             contents: [resourceContent],
         };
     });
+    // Track the registered resource for potential future removal
+    registeredResources.set(uri, registeredResource);
     return { type: "resource_link", ...resource };
 };

package/dist/server/index.js

@@ -1,4 +1,5 @@
 import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
+import { InMemoryTaskStore, InMemoryTaskMessageQueue, } from "@modelcontextprotocol/sdk/experimental/tasks";
 import { setSubscriptionHandlers, stopSimulatedResourceUpdates, } from "../resources/subscriptions.js";
 import { registerConditionalTools, registerTools } from "../tools/index.js";
 import { registerResources, readInstructions } from "../resources/index.js";
@@ -21,6 +22,10 @@ import { syncRoots } from "./roots.js";
 export const createServer = () => {
     // Read the server instructions
     const instructions = readInstructions();
+    // Create task store and message queue for task support
+    const taskStore = new InMemoryTaskStore();
+    const taskMessageQueue = new InMemoryTaskMessageQueue();
+    let initializeTimeout = null;
     // Create the server
     const server = new McpServer({
         name: "mcp-servers/everything",
@@ -39,8 +44,19 @@ export const createServer = () => {
                 listChanged: true,
             },
             logging: {},
+            tasks: {
+                list: {},
+                cancel: {},
+                requests: {
+                    tools: {
+                        call: {},
+                    },
+                },
+            },
         },
         instructions,
+        taskStore,
+        taskMessageQueue,
     });
     // Register the tools
     registerTools(server);
@@ -59,7 +75,7 @@ export const createServer = () => {
         // This is delayed until after the `notifications/initialized` handler finishes,
         // otherwise, the request gets lost.
         const sessionId = server.server.transport?.sessionId;
-        setTimeout(() => syncRoots(server, sessionId), 350);
+        initializeTimeout = setTimeout(() => syncRoots(server, sessionId), 350);
     };
     // Return the ServerFactoryResponse
     return {
@@ -68,6 +84,10 @@ export const createServer = () => {
             // Stop any simulated logging or resource updates that may have been initiated.
             stopSimulatedLogging(sessionId);
             stopSimulatedResourceUpdates(sessionId);
+            // Clean up task store timers
+            taskStore.cleanup();
+            if (initializeTimeout)
+                clearTimeout(initializeTimeout);
         },
     };
 };

package/dist/server/roots.js

@@ -48,11 +48,7 @@ export const syncRoots = async (server, sessionId) => {
                 }
             }
             catch (error) {
-                await server.sendLoggingMessage({
-                    level: "error",
-                    logger: "everything-server",
-                    data: `Failed to request roots from client: ${error instanceof Error ? error.message : String(error)}`,
-                }, sessionId);
+                console.error(`Failed to request roots from client ${sessionId}: ${error instanceof Error ? error.message : String(error)}`);
             }
         };
         // If the roots have not been synced for this client,

package/dist/tools/echo.js

@@ -9,6 +9,12 @@ const config = {
     title: "Echo Tool",
     description: "Echoes back the input string",
     inputSchema: EchoSchema,
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the 'echo' tool.

package/dist/tools/get-annotated-message.js

@@ -16,6 +16,12 @@ const config = {
     title: "Get Annotated Message Tool",
     description: "Demonstrates how annotations can be used to provide metadata about content.",
     inputSchema: GetAnnotatedMessageSchema,
+    annotations: {
+        readOnlyHint: true, // This tool only returns data, no side effects
+        destructiveHint: false, // Does not delete or modify anything
+        idempotentHint: true, // Same input always produces same output
+        openWorldHint: false, // Does not interact with external systems
+    },
 };
 /**
  * Registers the 'get-annotated-message' tool.

package/dist/tools/get-env.js

@@ -4,6 +4,12 @@ const config = {
     title: "Print Environment Tool",
     description: "Returns all environment variables, helpful for debugging MCP server configuration",
     inputSchema: {},
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the 'get-env' tool.

package/dist/tools/get-resource-links.js

@@ -15,6 +15,12 @@ const config = {
     title: "Get Resource Links Tool",
     description: "Returns up to ten resource links that reference different types of resources",
     inputSchema: GetResourceLinksSchema,
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the 'get-resource-reference' tool.

package/dist/tools/get-resource-reference.js

@@ -16,6 +16,12 @@ const config = {
     title: "Get Resource Reference Tool",
     description: "Returns a resource reference that can be used by MCP clients",
     inputSchema: GetResourceReferenceSchema,
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the 'get-resource-reference' tool.

package/dist/tools/get-roots-list.js

@@ -5,6 +5,12 @@ const config = {
     title: "Get Roots List Tool",
     description: "Lists the current MCP roots provided by the client. Demonstrates the roots protocol capability even though this server doesn't access files.",
     inputSchema: {},
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the 'get-roots-list' tool.

package/dist/tools/get-structured-content.js

@@ -18,6 +18,12 @@ const config = {
     description: "Returns structured content along with an output schema for client data validation",
     inputSchema: GetStructuredContentInputSchema,
     outputSchema: GetStructuredContentOutputSchema,
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the 'get-structured-content' tool.

package/dist/tools/get-sum.js

@@ -10,6 +10,12 @@ const config = {
     title: "Get Sum Tool",
     description: "Returns the sum of two numbers",
     inputSchema: GetSumSchema,
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the 'get-sum' tool.

package/dist/tools/get-tiny-image.js

@@ -6,6 +6,12 @@ const config = {
     title: "Get Tiny Image Tool",
     description: "Returns a tiny MCP logo image.",
     inputSchema: {},
+    annotations: {
+        readOnlyHint: true,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false,
+    },
 };
 /**
  * Registers the "get-tiny-image" tool.

package/dist/tools/gzip-file-as-resource.js

@@ -14,7 +14,6 @@ const GZIP_ALLOWED_DOMAINS = (process.env.GZIP_ALLOWED_DOMAINS ?? "")
 const GZipFileAsResourceSchema = z.object({
     name: z.string().describe("Name of the output file").default("README.md.gz"),
     data: z
-        .string()
         .url()
         .describe("URL or data URI of the file content to compress")
         .default("https://raw.githubusercontent.com/modelcontextprotocol/servers/refs/heads/main/README.md"),
@@ -29,6 +28,12 @@ const config = {
     title: "GZip File as Resource Tool",
     description: "Compresses a single file using gzip compression. Depending upon the selected output type, returns either the compressed data as a gzipped resource or a resource link, allowing it to be downloaded in a subsequent request during the current session.",
     inputSchema: GZipFileAsResourceSchema,
+    annotations: {
+        readOnlyHint: false,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: true,
+    },
 };
 /**
  * Registers the `gzip-file-as-resource` tool.

package/dist/tools/index.js

@@ -13,6 +13,10 @@ import { registerToggleSubscriberUpdatesTool } from "./toggle-subscriber-updates
 import { registerTriggerElicitationRequestTool } from "./trigger-elicitation-request.js";
 import { registerTriggerLongRunningOperationTool } from "./trigger-long-running-operation.js";
 import { registerTriggerSamplingRequestTool } from "./trigger-sampling-request.js";
+import { registerTriggerSamplingRequestAsyncTool } from "./trigger-sampling-request-async.js";
+import { registerTriggerElicitationRequestAsyncTool } from "./trigger-elicitation-request-async.js";
+import { registerSimulateResearchQueryTool } from "./simulate-research-query.js";
+import { registerTriggerUrlElicitationTool } from "./trigger-url-elicitation.js";
 /**
  * Register the tools with the MCP server.
  * @param server
@@ -38,5 +42,11 @@ export const registerTools = (server) => {
 export const registerConditionalTools = (server) => {
     registerGetRootsListTool(server);
     registerTriggerElicitationRequestTool(server);
+    registerTriggerUrlElicitationTool(server);
     registerTriggerSamplingRequestTool(server);
+    // Task-based research tool (uses experimental tasks API)
+    registerSimulateResearchQueryTool(server);
+    // Bidirectional task tools - server sends requests that client executes as tasks
+    registerTriggerSamplingRequestAsyncTool(server);
+    registerTriggerElicitationRequestAsyncTool(server);
 };