mcpcheck · demo

@modelcontextprotocol/server-everything passing

npx -y @modelcontextprotocol/server-everything

free preview run · last run 5 d ago (2026-09-02 17:57 UTC)

Clean install pass

1/1 hosts

Clients pass

Claude, Cursor, VS Code, Python SDK

OAuth n/a

n/a (stdio)

Tool catalog pass

13 tools

Error ergonomics pass

70/100

Protocol n/a

n/a

Clean install pass

Fresh npm/uv cache, start over stdio (or connect for remote URLs), initialize + tools/list, on GitHub-hosted Linux, macOS and Windows runners.

Host	Status	Node	Server	Tools	Time	Result
local macOS (local probe)	pass	v22.23.2	mcp-servers/everything 2.0.0	13	2.0 s	13 tools in 2.0 s

Client compatibility pass

ClaudeCursor inferredVS Code inferredPython SDK

Cursor and VS Code embed the official TypeScript SDK client; their chips are inferred from the TypeScript SDK rows (no GUI automation). Claude Code is verified with a claude mcp list health check — a real handshake over its own transport.

Client	local macOS
TypeScript SDK client (current) v1.30.0 embedded by Cursor, VS Code, Claude Desktop, Windsurf	pass 13 tools via ts-sdk 1.30.0
TypeScript SDK client (previous minor) v1.29.0 embedded by Cursor, VS Code	pass 13 tools via ts-sdk 1.29.0
Python SDK client (current) v2.1.1	pass 13 tools via mcp 2.1.1
Python SDK client (previous minor) v2.0.1	pass 13 tools via mcp 2.0.1
Claude Code (`claude mcp list`) v2.1.196 embedded by Claude Code	pass Claude Code 2.1.196 connected (13 tools in the catalog)

OAuth conformance n/a

Package specs run over stdio; there is no HTTP endpoint to authorize.

Tool catalog pass

13 tools · hash 5477733993d8298d · captured 6 d ago · mcp-servers/everything 2.0.0

Tool	Description	Arguments
`echo`	Echoes back the input string	1 (1 required)
`get-annotated-message`	Demonstrates how annotations can be used to provide metadata about content.	2 (1 required)
`get-env`	Returns all environment variables, helpful for debugging MCP server configuration	0
`get-resource-links`	Returns up to ten resource links that reference different types of resources	1
`get-resource-reference`	Returns a resource reference that can be used by MCP clients	2
`get-structured-content`	Returns structured content along with an output schema for client data validation	1 (1 required)
`get-sum`	Returns the sum of two numbers	2 (2 required)
`get-tiny-image`	Returns a tiny MCP logo image.	0
`gzip-file-as-resource`	Compresses a single file using gzip compression. Depending upon the selected output type, returns either the compressed data as a gzipped re…	3
`simulate-research-query`	Simulates a deep research operation that gathers, analyzes, and synthesizes information. Demonstrates MCP task-based operations with progres…	2 (1 required)
`toggle-simulated-logging`	Toggles simulated, random-leveled logging on or off.	0
`toggle-subscriber-updates`	Toggles simulated resource subscription updates on or off.	0
`trigger-long-running-operation`	Demonstrates a long running operation with progress updates.	2

Error ergonomics pass 70/100

10 malformed inputs per tool (missing/mistyped/null/out-of-range/unknown fields …). Each error response is graded on code · cause · fix · retryable; 24 distinct error shapes graded by the judge ($0.0000), cached.

Tool	Score	Errors	Accepted	Cases
`echo`	75/100	5	5	10
`get-annotated-message`	75/100	9	1	10
`get-env`	75/100	5	5	10
`get-resource-links`	75/100	4	6	10
`get-resource-reference`	72/100	8	2	10
`get-structured-content`	75/100	9	1	10
`get-sum`	68/100	7	3	10
`get-tiny-image`	75/100	5	5	10
`gzip-file-as-resource`	68/100	7	3	10
`simulate-research-query`	50/100	10	0	10
`toggle-simulated-logging`	75/100	5	5	10
`toggle-subscriber-updates`	75/100	5	5	10
`trigger-long-running-operation`	63/100	4	2	10

Weakest errors

Tool · case	Grade	Response
`get-resource-reference` `resourceId` is negative	✗ code · ✓ cause · ✓ fix · ✗ retryable llm	Invalid resourceId: -1. Must be a finite positive integer.
`get-sum` required field `b` is null	✓ code · ✓ cause · ✗ fix · ✗ retryable heuristic	MCP error -32602: Input validation error: Invalid arguments for tool get-sum: Invalid input: expected number, received null at b
`get-sum` empty arguments {}	✓ code · ✓ cause · ✗ fix · ✗ retryable heuristic	MCP error -32602: Input validation error: Invalid arguments for tool get-sum: Invalid input: expected number, received undefined at a Invalid input: expected number, received undefined at b
`gzip-file-as-resource` `name` has the wrong type (12345)	✓ code · ✓ cause · ✗ fix · ✗ retryable heuristic	MCP error -32602: Input validation error: Invalid arguments for tool gzip-file-as-resource: Invalid input: expected string, received number at name
`gzip-file-as-resource` `data` has the wrong type (12345)	✓ code · ✓ cause · ✗ fix · ✗ retryable heuristic	MCP error -32602: Input validation error: Invalid arguments for tool gzip-file-as-resource: Invalid input: expected string, received number at data
`trigger-long-running-operation` `duration` has the wrong type ("twelve")	✓ code · ✓ cause · ✗ fix · ✗ retryable heuristic	MCP error -32602: Input validation error: Invalid arguments for tool trigger-long-running-operation: Invalid input: expected number, received string at duration
`trigger-long-running-operation` `steps` has the wrong type ("twelve")	✓ code · ✓ cause · ✗ fix · ✗ retryable heuristic	MCP error -32602: Input validation error: Invalid arguments for tool trigger-long-running-operation: Invalid input: expected number, received string at steps
`simulate-research-query` required field `topic` missing	✓ code · ✗ cause · ✓ fix · ✗ retryable llm	MCP error -32600: Tool "simulate-research-query" requires task-based execution. Use client.experimental.tasks.callToolStream() instead.

Protocol conformance n/a

Not applicable to stdio package specs.

Badge

[![mcpcheck](https://agentwares-agentcheck.vercel.app/mcpcheck/demo/badge/everything.svg)](https://agentwares-agentcheck.vercel.app/mcpcheck/demo/everything)

Machine-readable: report.json · trust.json