CVE-2026-79744: MCPHub Authentication Bypass Vulnerability

CVE-2026-79744 Overview

MCPHub is a unified hub for centrally managing and dynamically orchestrating multiple Model Context Protocol (MCP) servers and APIs into separate endpoints. Versions prior to 1.0.29 expose the PUT /api/system-config endpoint without an authorization check. The updateSystemConfig handler relies only on the app-wide authentication middleware and a rate limiter, never inspecting the req.user.isAdmin flag. Any authenticated low-privilege user can modify system-wide configuration. The maintainer patched the flaw in version 1.0.29. The weakness maps to [CWE-269: Improper Privilege Management].

Critical Impact

Any authenticated MCPHub user can overwrite system-wide configuration, enabling privilege escalation and full compromise of confidentiality, integrity, and availability of the hub.

Affected Products

• MCPHub (samanhappy/mcphub) versions prior to 1.0.29
• Deployments exposing the PUT /api/system-config API endpoint
• Multi-tenant MCPHub installations with non-admin authenticated users

Discovery Timeline

• 2026-08-31 - CVE-2026-79744 published to NVD
• 2026-08-31 - Last updated in NVD database
• Patch release - MCPHub v1.0.29 published on GitHub with the authorization fix

Technical Details for CVE-2026-79744

Vulnerability Analysis

MCPHub exposes an administrative HTTP endpoint, PUT /api/system-config, which invokes the updateSystemConfig handler. The route sits behind the global authentication middleware, meaning any valid session token satisfies the access check. It also passes through a rate limiter, which throttles requests but performs no identity validation. The handler itself never reads req.user.isAdmin or any equivalent role attribute before mutating the configuration store.

An attacker who registers or steals a standard user account can therefore call the endpoint and rewrite system-level settings. Depending on which configuration keys the hub exposes, this includes routing rules for downstream MCP servers, upstream API credentials, logging behavior, and authentication parameters. Modifying these values allows the attacker to redirect traffic to attacker-controlled endpoints, disable security controls, or exfiltrate data flowing through the hub.

The attack requires network access to the MCPHub API and low privileges. No user interaction is needed. Successful exploitation yields high impact on confidentiality, integrity, and availability.

Root Cause

The root cause is a missing authorization check in the updateSystemConfig route handler. Authentication was conflated with authorization: because the app-wide middleware confirms a user is logged in, developers did not add a per-route role check. Administrative state-changing endpoints must independently verify that the caller holds the required role, which is the guidance captured by [CWE-269].

Attack Vector

Exploitation proceeds over the network against the MCPHub HTTP API. The attacker authenticates with any valid non-admin account, then issues a PUT request to /api/system-config with a JSON body containing arbitrary configuration values. The server accepts the request, persists the new configuration, and applies it to running components. No verified public proof-of-concept code is available. See the GitHub Security Advisory GHSA-4gc8-885f-qj36 for maintainer-provided details.

Detection Methods for CVE-2026-79744

Indicators of Compromise

• HTTP PUT requests to /api/system-config originating from non-admin user sessions or unexpected source IPs.
• Unexpected changes to MCPHub system configuration values, particularly routing targets, upstream credentials, or authentication settings.
• Log entries showing updateSystemConfig handler invocations at unusual times or frequencies.

Detection Strategies

• Enable access logging on the MCPHub API tier and alert on any PUT /api/system-config request whose authenticated principal lacks the admin role.
• Compare current MCPHub configuration state against a known-good baseline on a scheduled interval and flag drift.
• Correlate configuration-change events with recent account creation, password reset, or session anomalies to identify abuse chains.

Monitoring Recommendations

• Forward MCPHub application logs and reverse-proxy access logs to a centralized logging platform for retention and search.
• Monitor outbound connections from MCPHub to detect redirection to unexpected upstream MCP servers or APIs.
• Track authentication events for lateral movement patterns where a low-privilege account performs administrative API calls.

How to Mitigate CVE-2026-79744

Immediate Actions Required

• Upgrade MCPHub to version 1.0.29 or later, which adds the missing authorization check on updateSystemConfig.
• Audit MCPHub system configuration and revert any unauthorized changes made prior to patching.
• Rotate any credentials, API keys, or tokens stored in MCPHub configuration that may have been exposed.
• Review user accounts and remove or disable accounts that should not have access to the hub.

Patch Information

The issue is fixed in MCPHub v1.0.29. Upgrade instructions and release notes are available at the MCPHub v1.0.29 GitHub Release. Maintainer analysis is documented in the GitHub Security Advisory GHSA-4gc8-885f-qj36.

Workarounds

• If immediate patching is not feasible, block PUT /api/system-config at a reverse proxy or API gateway except for a narrow allowlist of administrator source IPs.
• Restrict MCPHub API exposure to trusted networks or place it behind a VPN until the upgrade is applied.
• Temporarily disable non-admin user accounts, or scope authentication to admin identities only, to shrink the attack surface.

# Example NGINX snippet to restrict the vulnerable endpoint to an admin IP allowlist
location = /api/system-config {
    limit_except GET {
        allow 10.0.0.10;   # admin workstation
        deny  all;
    }
    proxy_pass http://mcphub_upstream;
}