CVE-2026-62327: 9Router Information Disclosure Flaw

CVE-2026-62327 Overview

CVE-2026-62327 is an unauthenticated information disclosure vulnerability in 9Router through version 0.4.41. The flaw resides in the /api/usage/stats Next.js API route, which lacks authentication middleware. Remote attackers can send a single unauthenticated HTTP request to retrieve plaintext API keys for every connected artificial intelligence (AI) provider account. The response also exposes token counts, cost breakdowns, and request metadata. The vulnerability is classified under CWE-306: Missing Authentication for Critical Function.

Critical Impact
A single unauthenticated network request exposes plaintext API keys for all connected AI provider accounts, enabling account takeover, billing fraud, and quota exhaustion.

Affected Products

9Router through version 0.4.41
Deployments exposing the Next.js /api/usage/stats route
Any instance with connected third-party AI provider accounts

Discovery Timeline

2026-07-13 - CVE-2026-62327 published to the National Vulnerability Database (NVD)
2026-07-14 - Last updated in NVD database

Technical Details for CVE-2026-62327

Vulnerability Analysis

9Router is a router application built on the Next.js framework that proxies requests to multiple AI provider backends. The application exposes an internal telemetry endpoint at /api/usage/stats that aggregates usage data across all configured provider integrations. This endpoint returns the raw provider API keys alongside usage metrics rather than redacting or omitting them from the response payload.

The endpoint does not enforce authentication or session validation. Any remote client that can reach the HTTP interface receives the full response, including credentials for services such as commercial large language model (LLM) providers.

An attacker who obtains these keys can invoke third-party AI services directly, incurring charges against the victim's billing account. Attackers can also exhaust rate-limited quotas, disrupting legitimate application traffic. Because the leaked material is credential data rather than session tokens, revocation requires rotating keys at each upstream provider.

Root Cause

The root cause is missing authentication middleware on a Next.js API route that returns sensitive credential material. The handler was implemented without a guard that verifies caller identity before serialization. Compounding the flaw, the response schema embeds plaintext keys instead of masked identifiers, violating least-privilege design for telemetry surfaces.

Attack Vector

Exploitation requires only network reachability to the 9Router HTTP interface. The attacker issues an unauthenticated GET request to the /api/usage/stats path and parses the JSON response to extract provider credentials. No user interaction, prior authentication, or specialized tooling is required. See the GitHub Security Advisory GHSA-vjc7-jrh9-9j86 and the VulnCheck Security Advisory for additional technical detail.

Detection Methods for CVE-2026-62327

Indicators of Compromise

HTTP access log entries containing GET /api/usage/stats from unexpected source addresses or user agents
Unauthenticated 2xx responses to /api/usage/stats in reverse proxy or web application firewall (WAF) logs
Anomalous outbound API calls to AI provider endpoints originating from hosts other than the 9Router instance
Unexplained spikes in AI provider billing dashboards or sudden quota exhaustion notifications

Detection Strategies

Alert on any request to /api/usage/stats that lacks a valid session cookie or authorization header
Baseline normal request volume for the endpoint and flag deviations, particularly from external network ranges
Correlate 9Router access logs with AI provider audit logs to identify keys used from unfamiliar geolocations

Monitoring Recommendations

Enable verbose HTTP access logging on the reverse proxy fronting 9Router and forward logs to a centralized analytics platform
Enable API-side audit logging at each connected AI provider and monitor for use of keys from IP addresses outside the deployment
Configure billing threshold alerts at each AI provider to detect fraudulent consumption early

How to Mitigate CVE-2026-62327

Immediate Actions Required

Restrict network access to the 9Router HTTP interface using firewall rules, VPN, or a private network segment until patched
Rotate all API keys for every AI provider configured in 9Router, assuming compromise if the endpoint has been reachable from untrusted networks
Review AI provider billing and usage dashboards for unauthorized activity dating back to initial deployment
Upgrade 9Router to a fixed release once available from the maintainer

Patch Information

At the time of publication, the GitHub Security Advisory GHSA-vjc7-jrh9-9j86 is the authoritative source for patch availability. Administrators should track that advisory and the project repository for a release superseding version 0.4.41.

Workarounds

Place 9Router behind a reverse proxy that enforces authentication on all /api/* routes, including /api/usage/stats
Block external access to the /api/usage/stats path at the reverse proxy or WAF layer
Bind the 9Router service to 127.0.0.1 and require an authenticated tunnel for administrative access
Store provider API keys in a secrets manager and rotate them on a short schedule to limit exposure windows

bash

# Example nginx configuration to block external access to the vulnerable endpoint
location = /api/usage/stats {
    allow 127.0.0.1;
    deny all;
    return 403;
}