Free check · no sign-up

What does an AI agent get from your MCP server?

Paste the URL your users connect to. We read what your server tells every client before sign-in, the way Claude or ChatGPT would, and show what a connected agent gets by default.

check.subcloak.comExample · fictional

GETmcp.example.com/.well-known/oauth-protected-resource200scopes_supported: 6
POSTmcp.example.com/mcp initialize401no scope in challenge
GETmcp.example.com/.well-known/oauth-authorization-server200PKCE S256 · CIMD

What a connected agent gets

Everything, by default

Clients are told to request all 6 scopes, including 3 that write.

What it reads

Three public documents. Nothing else.

The same reads as our State of MCP Scopes study, where 196 of 220 servers that use OAuth gave agents everything by default.

01
Protected-resource metadata
Where users sign in, and the scopes your server lists.
02
One unauthenticated connect
The sign-in challenge every client gets, and the scopes it names.
03
Authorization server metadata
PKCE, how clients register, and the scopes it offers.
No sign-in, no client registration, no tool calls. Every request names check.subcloak.com in its User-Agent, so you can find it in your logs. Read the study

What it can't see

The rest happens after sign-in.

Metadata shows what an agent is offered. Whether your server holds each tool to that, and what an attacker can do with it, takes a signed-in test by someone accountable for the result.

Whether each tool is held to its scope

A token for reading can still write if a tool doesn't check. That needs a signed-in test.

Whether tokens are bound to your server

Tokens issued for another service, or forwarded to other APIs, turn one leak into many.

Whether outside text can trigger risky tools

A sentence in a ticket or email making the agent delete, send or export.

A fix that keeps every client working

Splitting scopes and stepping up without breaking Claude, ChatGPT or Cursor.

What the free check can't see

Scope Snapshot

Two days of signed-in testing on one MCP server or agent: is each tool held to its scope, is each token held to your server, and which three changes matter most.

$490≈€440≈£370≈CA$700≈A$700≈S$630≈AED 1,800₹25,000

Fixed price · billed in USD

Approximate, at 9 Oct 2026 rates · invoiced in USD

For clients in India · plus 18% GST

2 working days

Details

Our main service

Agent Access Audit

An independent, signed-in test of what your MCP server or AI agent can really do, measured against what it needs, with the token, consent and prompt-injection paths around it. You get the fixes, a retest and a letter your customers can verify.

$2,900≈€2,600≈£2,200≈CA$4,100≈A$4,200≈S$3,700≈AED 10,700₹1,50,000per server

Fixed price · billed in USD

Approximate, at 9 Oct 2026 rates · invoiced in USD

For clients in India · plus 18% GST

Up to 4 servers: $6,500≈€5,800≈£4,900≈CA$9,200≈A$9,300≈S$8,300≈AED 23,900₹3,40,000

7–10 working days

See what's included