mcp-best-practices

Build, harden, and debug production MCP servers with the TypeScript SDK. Use when writing or reviewing an MCP server - transports, tool schemas, errors, OAuth, token bloat, SDK migrations, MCP Apps, Registry. Assumes a server already exists.

Decision reference for building production MCP servers with the TypeScript SDK. Not a tutorial - assumes you already have a working server and need to make it correct, fast, and secure.

Quick Reference

Component	Current	Notes
Spec (released)	2026-07-28 (specification)	Stateless/sessionless overhaul - see "Spec 2026-07-28" below and references/spec-2026-07-28.md
Spec (still deployed)	2025-11-25	What most shipped clients and servers actually speak today; the v2 SDK's default
TS SDK (current)	v2.0.0 (2026-07-27), nine packages in lockstep: /server, /client, /core, /hono, /express, /node, /fastify, /codemod, /server-legacy	Speaks 2025-era by default; 2026-07-28 is opt-in
TS SDK (legacy)	v1.30.0 (@modelcontextprotocol/sdk)	Bug + security fixes for >=6 months after v2 GA; source on the v1.x branch
JSON Schema	2020-12 default (2019-09 / draft-07 accepted since v2.0.0)	-
Transport	Streamable HTTP (remote), stdio (local)	SSE + WebSocket removed in v2
Extensions	MCP Apps (Stable, SEP-1865), Auth Extensions (official), Tasks (ext-tasks)	Domain-specific WGs
Registry	Preview with v0.1 API freeze since 2025-10-24 (registry)	GA pending

v2 imports (current):

typescript

import { McpServer } from "@modelcontextprotocol/server";
import { WebStandardStreamableHTTPServerTransport } from "@modelcontextprotocol/server";
import { ProtocolError, ProtocolErrorCode } from "@modelcontextprotocol/core";

v1 imports (legacy line, still widely deployed):

typescript

import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { WebStandardStreamableHTTPServerTransport } from "@modelcontextprotocol/sdk/server/webStandardStreamableHttp.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

The Two Eras

The most decision-relevant fact after the 2026-07-28 release: upgrading to SDK v2.0.0 does not move you to the new spec. A hand-constructed Client/Server/McpServer keeps speaking the 2025-era protocol it was written for.

Every revision from 2024-10-07 through 2025-11-25 opens with initialize and shares one wire behavior - the SDK calls that family legacy. 2026-07-28 starts the modern era: no initialize, a server/discover advertisement instead, a _meta envelope on every request. Selection is explicit:

versionNegotiation.mode	Behavior
absent / 'legacy'	The 2025 initialize handshake, byte for byte. No probe. This is the default.
'auto'	Probe with server/discover; fall back to initialize against a 2025-only server
{ pin: '2026-07-28' }	That revision or nothing - a pin never falls back

Build new servers on the 2025-era wire unless you control both ends. The stateless design guidance throughout this skill is what makes the eventual era switch cheap.

Tooling: SDK docs (v2); MCP Inspector, which connects as legacy by default (see "Testing Against Each Era" in references/spec-2026-07-28.md); the conformance suite; and the mcp-server-dev plugin for scaffolding.

Server Setup

Transport Decision

Scenario	Transport	Key Config
Remote, stateless (K8s, CF Workers)	WebStandardStreamableHTTPServerTransport	sessionIdGenerator: undefined, enableJsonResponse: true
Remote, stateful (long tasks, SSE)	WebStandardStreamableHTTPServerTransport	sessionIdGenerator: () => randomUUID()
Local CLI / Claude Desktop	StdioServerTransport	Default
Legacy SSE clients	SSE removed in v2 - migrate to Streamable HTTP	-

Stateless Pattern (recommended for remote deployment)

Per-request server+transport creation is the canonical pattern. Maintainer @ihrpr confirms: "each transport should have an instance of MCPServer" (#343). Sharing instances leaks cross-client data (GHSA-345p-7cg4-v4c7).

typescript

app.post("/mcp", async (c) => {
  const server = new McpServer({ name: "my-server", version: "1.0.0" });
  // Register tools, resources, prompts...
  registerTools(server);

  const transport = new WebStandardStreamableHTTPServerTransport({
    sessionIdGenerator: undefined,   // stateless - no session tracking
    enableJsonResponse: true,        // JSON responses, no SSE streaming
    // Origin/Host checking is OFF unless you turn it on: the SDK defaults
    // enableDnsRebindingProtection to false and leaves both lists unset.
    enableDnsRebindingProtection: true,
    allowedOrigins: ["https://app.example.com"],
    allowedHosts: ["mcp.example.com"],
  });

  // All tools/resources must be registered before connect() (#893)
  try {
    await server.connect(transport);
    return transport.handleRequest(c.req.raw);
  } finally {
    await transport.close();
    await server.close();
  }
});

The McpServer must be per-request, but its constant inputs must not be. Hoist to module level: Zod schemas, annotation objects ({ readOnlyHint: true, ... }), tool description strings, payment configs, upstream API clients.

If you only route POST (the common stateless layout), answer GET /mcp with an explicit 405 Method Not Allowed - the spec requires it when no SSE stream is offered, and the official TS client reads 405 as the benign no-stream signal, while an empty 200 sends it into a reconnect storm.

For transports, sessions, HTTP/2 gotchas, and K8s deployment: see references/transport-patterns.md

Framework Integration

The transport is web-standard, so Hono and the Workers runtime need no adapter; v2 also ships @modelcontextprotocol/hono (createMcpHonoApp()) and @modelcontextprotocol/express (wrapping NodeStreamableHTTPServerTransport for IncomingMessage/ServerResponse). On Cloudflare Workers call preloadSchemas() at module scope - v2's workerd build does it automatically. Examples: references/transport-patterns.md.

Tool Design

Registration API

v1 (legacy line) - server.tool(name, description, zodShape, annotations, handler). Positional overloads are ambiguous; same fields as v2 below minus outputSchema. Removed entirely in v2.

v2 (current) - registerTool() with config object:

typescript

server.registerTool("search_docs", {
  title: "Document Search",
  description: "Search documents by keyword or phrase",
  inputSchema: z.object({
    query: z.string().describe("Search query"),
    max_results: z.number().optional().describe("Max results (default 20)"),
  }),
  outputSchema: z.object({
    results: z.array(z.object({ id: z.string(), text: z.string() })),
    has_more: z.boolean(),
  }),
  annotations: { readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: true },
}, async ({ query, max_results }) => {
  const result = await fetchDocs(query, max_results);
  return {
    // Both channels carry IDENTICAL bytes. Divergent payloads = the text block
    // silently vanishes on Claude Code/Codex/Copilot. See "Tool Result Delivery" below.
    structuredContent: result,
    content: [{ type: "text", text: JSON.stringify(result) }],
  };
});

Naming

Spec 2025-11-25 (SHOULD, not MUST): 1-128 chars, case-sensitive, A-Za-z0-9_-. only. DO: search_docs, get_user_profile, admin.tools.list. DON'T: search (generic names collide across servers), Search Docs (spaces disallowed). Service-prefix (github_*, jira_*) when multiple servers are active - LLMs confuse generic names. Bake the prefix into the tool name itself: the spec is explicit that "The server name (from serverInfo) is not guaranteed to be unique across servers and SHOULD NOT be relied upon for disambiguation", so an aggregator cannot derive a safe prefix for you.

Schema Rules

.describe() on every field - this is what LLMs use for argument generation. Three constructs break silently (z.union(), raw JSON Schema, z.transform()), as does client-side AJV strict validation - see "Known SDK Bugs" below.

Pagination is the primitive most servers hit first: a tools/list or resources/list with 50+ entries should paginate. The protocol cursor is opaque - never parse or synthesize it; loop until nextCursor is absent. It is distinct from in-tool offset/limit args.

Zod-to-JSON-Schema conversion rules, outputSchema/structuredContent patterns, non-text content types, the other tool-definition fields (icons, listChanged, execution.taskSupport), and the remaining primitives (prompts, resources, resource templates, completions, cancellation): see references/tool-schema-guide.md

Annotations

All are optional hints (untrusted from untrusted servers per spec):

Annotation	Default	Meaning
readOnlyHint	false	Tool doesn't modify its environment
destructiveHint	true	May perform destructive updates (only when readOnly=false)
idempotentHint	false	Repeated calls with same args have no additional effect
openWorldHint	true	Interacts with external entities (APIs, web)

Set them accurately - clients use them for consent prompts and auto-approval decisions.

The "Lethal Trifecta": private-data access + exposure to untrusted content + external communication in one agent creates data-theft conditions (demonstrated with a malicious calendar event, an MCP calendar server, and a code-execution tool). Design tool sets so no single agent holds all three.

Stateful Tools

With no protocol-level session on 2026-07-28, cross-call state uses server-minted handles passed as ordinary tool arguments: a creation tool returns { basket_id: "bsk_a1b2c3" }, later tools take basket_id as an argument, and the model carries it forward. A handle is a name, not a capability - validate the caller against it on every call, keep it opaque with real entropy, and state its retention policy in the creation tool's description. Expired or unknown handles return a tool execution error so the model can recover by creating new state. Full rules: references/spec-2026-07-28.md.

Tool Result Delivery: content vs structuredContent

The footgun: when a tool returns BOTH a text content block and structuredContent, several major clients (Claude Code, Codex CLI, VS Code Copilot, Goose) silently drop the text block and forward only structuredContent to the model. If the two payloads differ, the human-readable one vanishes. This is client behavior the spec does not constrain - not an SDK transform. Don't return both channels expecting both to reach the model.

Empirically tested - Claude Code 2.1.165 (MCP 2025-11-25)

Measured with claude -p --output-format=stream-json, reading the exact tool_result the model received:

Tool returns	What the model receives
One text block, no structuredContent	text verbatim
content: [] + structuredContent	JSON.stringify(structuredContent) as a string in the content slot - works
text block + structuredContent	text block silently dropped; structuredContent wins
text + structuredContent + outputSchema	same - outputSchema makes zero difference
two text blocks, no structuredContent	both preserved verbatim

structuredContent is not a separate typed channel to the model on Claude Code - it is stringified into the standard tool_result content slot, so it costs the same tokens as the equivalent JSON-as-text. It does not buy cheaper or out-of-band structured data.

Intentional, per Anthropic maintainer (anthropics/claude-code#9962): structuredContent support landed in Claude Code v2.0.21 and "we made structuredContent the default when both formats are present... optimizing for agent performance." Reproduced across unrelated servers (Laravel, Roblox Studio, YouTube) - host-side precedence, not a server bug.

What the spec actually says (2025-11-25)

There is no precedence rule - the spec never says which field a client should prefer when both are present (Discussion #1563), and that gap is the documented root cause of client divergence. The only relevant normative line is a backwards-compat SHOULD: "a tool that returns structured content SHOULD also return the serialized JSON in a TextContent block." The official TypeScript SDK passes both fields through verbatim; any stringify-into-content you observe is the host harness, not the SDK.

Cross-client behavior (the matrix above is Claude Code only)

Client	When both content + structuredContent present
Claude Code CLI, OpenAI Codex CLI, VS Code Copilot, Goose	shadow - only structuredContent reaches the model (text dropped)
Cursor, Claude.ai web, ChatGPT MCP connector	prefer content / surface both to the model
Google ADK (framework)	forwards both by default; content-only is opt-in

(Non-Claude-Code rows come from issue trackers and maintainer statements, not the stream-json harness - treat exact delivery as client-version-dependent.)

The rule for server authors

• DON'T return divergent content and structuredContent (e.g. a rendered ASCII table as text + different JSON as structured). On shadowing clients the text silently disappears and only the JSON reaches the model.
• DO, if you emit structuredContent, mirror the same bytes into a text block: content: [{ type: "text", text: JSON.stringify(payload) }]. This is the spec's backwards-compat SHOULD. Shadowing clients use the structured copy; others fall back to the identical text - either way the model gets the data. Mirroring does not double tokens on shadowing clients (they drop the text).
• PREFER one channel per tool / per mode. For a human-readable rendering (table, summary) to reach the model, return it as text only, no structuredContent - or expose a format: "table" | "json" arg (table -> text-only; json -> JSON mirrored into both channels). Both are empirically valid on Claude Code and keep one channel per call.
• outputSchema gates client-side validation only; it does not make the text block survive on shadowing clients.

content blocks are not text-only - image, audio, resource_link, and embedded resource blocks all exist, with annotations (audience, priority, lastModified); for those and the image preview + URL pattern see references/tool-schema-guide.md.

Error Handling

Two distinct mechanisms with different LLM visibility:

Type	LLM Sees It?	Use For
Tool error (isError: true in CallToolResult)	Yes - enables self-correction	Input validation, API failures, business logic errors
Protocol error (JSON-RPC error response)	Maybe - clients MAY expose	Unknown tool, malformed request, server crash

Per SEP-1303 (merged into spec 2025-11-25): input validation errors MUST be tool execution errors, not protocol errors. The LLM needs to see "date must be in the future" to self-correct.

typescript

// DO: Tool execution error - LLM can self-correct
return {
  isError: true,
  content: [{ type: "text", text: "Date must be in the future. Current date: 2026-03-25" }],
};

// DON'T: Protocol error for validation - LLM can't see this
throw new McpError(ErrorCode.InvalidParams, "Invalid date");

Known SDK behavior: converting an McpError thrown from a tool handler into a CallToolResult drops the error.data field, so structured data embedded there may never reach the client. The x402/MPP ecosystem standardized on isError: true results with structuredContent for this reason.

For full error taxonomy, code examples, payment error patterns, and why -32042 is not available as a "Payment Required" code: see references/error-handling.md

Resources and Instructions

Set instructions in the server constructor - a system-level hint to the LLM about how to use your server:

typescript

const server = new McpServer({
  name: "docs-api",
  version: "1.0.0",
  instructions: "Knowledge base API. Use search_docs for full-text search, get_doc for retrieval by ID. All tools are read-only.",
});

Ship guides and structured data as resources under a docs:// URI scheme (server.resource(...)) - see "Other Server Primitives" in references/tool-schema-guide.md.

Performance

Token Bloat Mitigation

Tool definitions consume context window before any conversation starts. GitHub MCP: 20,444 tokens for 80 tools (SEP-1576).

Strategies:

1. 5-15 tools per server - community sweet spot. Split beyond that.
2. Outcome-oriented tools - bundle multi-step operations into single tools (e.g., track_order(email) not get_user + list_orders + get_status).
3. Response granularity - return curated results, not raw API dumps. 800-token user object vs 20-token summary.
4. outputSchema + structuredContent - typed output for programmatic/PTC clients. Caveat: on shadowing clients structuredContent is stringified into the model's context at the same token cost as text - not a free out-of-band channel (see "Tool Result Delivery").
5. Dynamic tool loading - register only relevant tool subsets per request context (e.g. a ?tools=search,fetch query param). Pair with listChanged if the set changes mid-session. Vary the set per connection, not mid-conversation: tool definitions sit in the prompt prefix, and "Adding or removing tool definitions mid-conversation invalidates that cache, and the resulting miss can cost more tokens than the definitions you removed." A client must also treat a cached list as stale the moment list_changed arrives, even before the ttlMs you advertised.
6. Progressive tool discovery / code mode - large-catalog clients increasingly use a search_tools meta-tool and programmatic tool calling, where structuredContent is consumed outside the model context (client best practices). Curated, well-described tools make these flows work.

Result-Size Budgets (per-client caps)

Clients silently truncate large tool results. Budget for the strictest client you target:

Client	Default cap	Configurable
Claude Code	25,000 tokens (warning at 10k)	MAX_MCP_OUTPUT_TOKENS env; per-tool _meta["anthropic/maxResultSizeChars"] up to 500,000 chars, which replaces the token cap for text rather than being bounded by it
OpenAI Codex CLI	10,000 tokens on every current model (~40KB); bytes-mode 10,000 survives only on legacy gpt-5.2 and as the unknown-model fallback	tool_output_token_limit config
Gemini CLI	40,000 chars (head 20% / tail 80% trim; full output saved to a file)	settings; 0 or negative disables

Enforce your own cap server-side - see "Result-Size Budgets and Truncation" in references/tool-schema-guide.md. Two rules worth stating here: never truncate isError results (payment/auth challenges must survive intact), and treat client budgets as per-connection properties - accept them as URL query params (?max_chars=, alongside ?tools=) rather than growing every tool schema with override args.

Long-Running Tools

A client timeout is a wall clock, not an idle timer. Claude Code's per-server tool-call timeout is documented as a "Hard wall-clock limit per call; progress notifications do not extend it" - so the common instinct (emit notifications/progress to keep a slow call alive) does not work there. Progress is for the human watching, not for buying time.

Design past the cap instead: return quickly with a server-minted handle and let the caller poll (see "Stateful Tools"), or adopt the io.modelcontextprotocol/tasks extension, which is built for exactly this and returns a CreateTaskResult the client polls via tasks/get. Tasks is per-request opt-in - a server that cannot service a call synchronously for a client that did not declare the tasks capability MUST return -32021 (Missing Required Client Capability) naming the extension, not silently block.

No-Parameter Tools

For tools with no inputs, use an explicit empty schema - not undefined or omission:

typescript

inputSchema: { type: "object" as const, additionalProperties: false }

Security

Top Threats (real-world incidents, 2025-2026)

Attack	Example	Mitigation
Tool poisoning	Hidden instructions in descriptions (WhatsApp MCP, Apr 2025)	Review tool descriptions; clients should display them
Supply chain	Malicious npm packages (Smithery breach, Oct 2025)	Pin versions, audit dependencies
Stdio config injection	User-controlled input reaches StdioServerParameters unsanitized (OX Security, 2026-04-15)	Sanitize stdio config in client code; prefer first-party servers. Treated as "by design" - not patched in the SDK
Cross-server shadowing	Malicious server overrides legitimate tool names	Service-prefix tool names; validate tool sources
Token theft	Over-privileged PATs with broad scopes	Minimal scopes; OAuth 2.1 Resource Indicators (RFC 8707)
Token passthrough	Server accepts/forwards tokens not issued for it	Validate audience claim; never transit client tokens to upstream APIs
Confused deputy	Proxy server consent cookies exploited via DCR	Per-client consent before forwarding to third-party auth
Session hijacking	Stolen/guessed session IDs for impersonation	Cryptographically random IDs, bind to user identity, never use for auth
Cross-client response leak	Shared McpServer/transport reused across clients (CVE-2026-25536, affects v1.10.0-1.25.3)	Require SDK >= v1.26.0; per-request server+transport
UriTemplate ReDoS	Malicious URI patterns (CVE-2026-0621)	Upgrade to v1.25.2+ / v2.0.0-alpha.1+

Generic hygiene still applies: validate inputs at tool boundaries, enforce per-user access control, rate limit, never interpolate tool input into shell commands, block private IPs on outbound fetches, bind local servers to 127.0.0.1.

Server-Side Requirements (spec normative)

• Validate the Origin header - but only reject when it is present and invalid: "If the Origin header is present and invalid, servers MUST respond" with 403. Shipping clients exist that send no Origin at all; a blanket 403-on-missing locks them out.
• Turn the checks on. WebStandardStreamableHTTPServerTransport defaults enableDnsRebindingProtection to false and leaves allowedOrigins/allowedHosts unset, so the stock stateless constructor validates nothing. The @modelcontextprotocol/express and /hono factories enable Host validation for localhost by default; the raw transport does not.
• MCP-Protocol-Version is not optional on a modern wire. The header survived the sessionless overhaul: "Every POST request to the MCP endpoint MUST include an MCP-Protocol-Version header", and its value MUST match io.modelcontextprotocol/protocolVersion in the body's _meta or the server MUST answer 400 Bad Request with a HeaderMismatch error. The version rides _meta and the header, redundantly and on purpose - intermediaries route on the header while the server executes on the body, so both must agree.
• Be lenient about which version, not about whether it is declared. On 2025-era wires accept a range of declared versions rather than enforcing one - clients advertising 2024-11-05 are still in the wild, and a server supporting pre-2025-06-18 clients MAY treat a header-less request as 2025-03-26. A server that does not support those clients MUST reject a header-less request.

Auth (OAuth 2.1)

MCP normatively requires OAuth 2.1 (draft-ietf-oauth-v2-1-13), not 2.0 - PKCE mandatory, implicit flow removed. Servers are Resource Servers; clients MUST send Resource Indicators (RFC 8707) binding tokens to your server.

• Validate audience - reject tokens not issued for your server (passthrough is forbidden). PKCE S256, short-lived tokens, minimal scopes (elevate via WWW-Authenticate challenges).
• Use a tested validation library (Keycloak, Auth0, ...) - don't roll your own; never log Authorization headers/tokens/secrets.
• RFC 9207 iss interop footgun: advertising authorization_response_iss_parameter_supported: true makes strict clients MUST-validate a callback iss that some of them drop. Advertise the flag as false while still sending iss - see references/security-auth.md.

For full security attack/mitigation patterns and auth implementation details: see references/security-auth.md

Known SDK Bugs

Must-know as of sdk@1.30.0 / server@2.0.0:

• z.union()/z.discriminatedUnion() silently produce empty schemas on every released v1, v1.30.0 included (#1643, backport still open) - use flat z.object() + z.enum().
• Require SDK >= v1.26.0 - shared instances leaked cross-client data below that (CVE-2026-25536).
• Register everything before connect() - later registration throws; open on both main and v1.x (#893).
• Client AJV strict rejects unstripped structuredContent extras - .parse() upstream data first, or .passthrough() for intentional extras.
• v1.30.0 stamps every tool schema "$schema": "http://json-schema.org/draft-07/schema#", and a strict 2020-12 client rejects the whole tool: "JSON Schema declares an unsupported dialect ... The default validator supports JSON Schema 2020-12 only." One bad schema can take the server's other tools down with it in clients that drop the whole tools/list. v2 emits 2020-12. Open (#2721, #2677); @modelcontextprotocol/inspector >= 2.4.0 flags it for you.
• Don't reuse one McpServer across createMcpHandler requests on v2. Each request wraps onclose, the chain grows unbounded, and it dies with RangeError: Maximum call stack size exceeded at roughly 19-25k accumulated sessions - affects released server@2.0.0 (#2607). The per-request pattern above is the fix.

Full table (statuses, zod 3->4 dropping additionalProperties, refine/superRefine never running, transport-closure stack overflow, HTTP/2, raw JSON Schema, z.transform(), ReDoS): see references/sdk-bugs.md

V2 Migration

For comprehensive migration guide with all breaking changes and before/after code: see references/v2-migration.md

Key breaking changes:

1. Package split: @modelcontextprotocol/sdk -> @modelcontextprotocol/server + /client + /core
2. ESM-first (CJS builds restored in beta.2), Node.js 20+ (Bun/Deno supported)
3. Zod v4 required (or any Standard Schema library)
4. McpError -> ProtocolError (from @modelcontextprotocol/core)
5. extra parameter -> structured ctx with ctx.mcpReq
6. server.tool() -> registerTool() (config object, not positional args)
7. SSE server transport removed (clients can still connect to legacy SSE servers)
8. @modelcontextprotocol/hono and @modelcontextprotocol/express middleware packages
9. DNS rebinding protection enabled by default for localhost servers

v1.x gets 6 more months of support after v2 stable ships. No rush, but write new code with v2 patterns in mind.

Spec 2026-07-28 (released)

Published 2026-07-28 (release announcement, changelog) - now the latest revision. Remember it is opt-in on the SDK (see "The Two Eras"): 2025-11-25 remains what most deployed software speaks.

Four shifts that change a decision you make today:

• MCP is stateless and sessionless. The initialize handshake and Mcp-Session-Id are gone (SEP-2575, SEP-2567); every request carries its protocol version, client identity, and capabilities in _meta, and cross-call state uses handles (see "Stateful Tools"). Do not build new servers on session affinity.
• server/discover is a server MUST - it advertises versions/capabilities/identity; clients MAY skip it and handle UnsupportedProtocolVersionError inline.
• Roots, Sampling, Logging, and the HTTP+SSE transport are Deprecated under a formal feature lifecycle (12-month minimum window, SEP-2577/SEP-2596). They still work; design new servers without them.
• Allocate application-defined error codes outside -32768..-32000 - -32020..-32099 is reserved for the spec and -32000..-32019 is legacy that new implementations SHOULD NOT use at all (PR #2907).

The content vs structuredContent dual-delivery footgun is unchanged - no precedence rule landed, so the guidance above still holds.

Everything else - MRTR, subscriptions/listen, _meta identity keys, requestState, Mcp-Method/Mcp-Name, cacheable results, per-request log level, auth changes, the removals (SSE resumability, ping, execution.taskSupport), era testing, working groups: see references/spec-2026-07-28.md

Extensions

Optional, strictly additive capabilities named {vendor-prefix}/{extension-name} (official: io.modelcontextprotocol/*; third-party: reversed domain). Negotiated in initialize capabilities on 2025-era wires; on 2026-07-28 clients advertise support per request in _meta["io.modelcontextprotocol/clientCapabilities"]. Official ones: MCP Apps (/ui, interactive HTML UIs, Stable, widely supported; ext-apps 2.0.0 since 2026-09-08 - breaking on the TypeScript side only, the wire protocol is unchanged), OAuth Client Credentials (Draft), Enterprise-Managed Authorization (Stable 2026-06-18), Tasks (official since 2026-08-19) - client matrix.

Server capabilities beyond tools, all 2025-era APIs (the SDK default):

Capability	Purpose	v2 API
Elicitation	Request structured user input mid-tool	ctx.mcpReq.elicitInput()
Sampling	Request LLM completion from client	ctx.mcpReq.requestSampling()
Tasks	Long-running ops with lifecycle management	Official extension (SEP-2663)
Progress	Incremental progress on requests	ctx.mcpReq.sendProgress()

On 2026-07-28 servers cannot send requests to clients at all: elicitation and sampling go through MRTR (return an InputRequiredResult, read inputResponses on the retry). Tasks moved out of core into the polled io.modelcontextprotocol/tasks extension (ext-tasks).

For MCP Apps architecture, ext-apps SDK, and build patterns: see references/mcp-apps.md For the extensions system, auth extensions, elicitation/sampling/tasks detail, and the MCP Registry: see references/extensions-registry.md