9Router vulnerabilities

9Router before version 0.5.2 treats loopback requests as trusted and allows /v1/* access without an API key. A same-host reverse proxy forwarding public traffic through 127.0.0.1 causes src/dashboardGuard.js to misclassify external requests as local, enabling a remote unauthenticated attacker to access /v1 APIs such as /v1/models and potentially abuse configured upstream provider credentials through /v1 proxy endpoints.

9Router before version 0.5.2 fails to protect the /codex path, allowing a remote unauthenticated attacker to bypass the API-key gate and send requests to /codex/*, which are rewritten to /api/v1/responses. The attacker can exploit this to make upstream provider calls using operator-stored LLM credentials.

In 9Router prior to version 0.4.80, the dashboard login rate limiter relies on the attacker-controlled X-Forwarded-For HTTP header. This allows bypassing the 5-attempt threshold and progressive lockout durations by rotating the header value on each login attempt.

9router 0.4.59 contains a chain of vulnerabilities: a hardcoded default password (123456), a bypass of the LOCAL_ONLY network gate via a spoofed Host header, and unvalidated arguments passed to child_process.spawn() when registering MCP plugins. A remote, unauthenticated attacker can log in with the default credential, spoof the Host header to reach local-only routes, and register a malicious MCP plugin to achieve arbitrary code execution on the host operating system.

In 9Router versions 0.2.21 through 0.4.44, a hardcoded fallback JWT secret '9router-default-secret-change-me' was used, allowing attackers to forge an auth_token cookie when JWT_SECRET is unset. The issue is fixed in version 0.4.44.

In 9Router versions 0.4.30 through 0.4.37, the middleware in src/proxy.js did not protect /api/cli-tools/* and /api/mcp/*, allowing unauthenticated plugin registration and command execution via the MCP bridge. This vulnerability is fixed in version 0.4.37.

9Router through version 0.4.41 contains an unauthenticated information disclosure vulnerability that allows remote attackers to retrieve plaintext API keys for all connected AI provider accounts by sending a single unauthenticated request to the /api/usage/stats endpoint. Attackers can exploit the missing authentication middleware on the Next.js API route to obtain full API key strings alongside token counts, cost breakdowns, and request metadata.

9Router through version 0.4.41 contains an unauthenticated access vulnerability allowing remote attackers to interact with provider management API endpoints without credentials due to missing authentication middleware in Next.js API routes under src/app/api/providers/*. Attackers can enumerate, create, modify, or delete provider connections.

In 9Router prior to version 0.4.80, the /api/settings/database endpoint allows full database export (containing all credentials, API keys, OAuth tokens, and settings) and full database import (complete overwrite) without any authentication requirement beyond the ALWAYS_PROTECTED middleware check, which only validates JWT or CLI token. This issue is fixed in version 0.4.80.

9Router before version 0.4.44 contains an OS command injection vulnerability in the unauthenticated POST /api/tunnel/tailscale-install endpoint. The sudoPassword field from the request body is passed to the stdin of a 'sudo -S sh' process, allowing a remote unauthenticated attacker to execute arbitrary OS commands.

An SSRF vulnerability in the POST /api/provider-nodes/validate route in 9router allows an attacker to make server-side requests to internal addresses, including loopback and link-local, due to flaws in the assertPublicUrl guard. Additionally, internal responses may be disclosed, and the API key is forwarded to internal destinations.

9Router in 0.5.4 and earlier does not restrict the issuerUrl parameter in the POST /api/auth/oidc/test endpoint, allowing unauthenticated attackers (when dashboard login is disabled) to scan internal services and reflect OIDC discovery fields including token_endpoint and jwks_uri.

9Router before 0.4.72 contains a server-side request forgery (SSRF) vulnerability in the /v1/web/fetch endpoint. The endpoint accepts a user-controlled url parameter and passes it to a configured external scraping provider to fetch content. The URL is only validated as syntactically valid via new URL() with no blocklist for private IP ranges, cloud metadata endpoints, link-local addresses, or internal hostnames. An authenticated or locally-connected user can cause the server to fetch arbitrary internal URLs and have the response content returned, enabling read-access SSRF that can expose cloud metadata credentials, reach internal services, and bypass authentication on localhost endpoints.

9Router prior to 0.5.2 allows a remote authenticated attacker to achieve arbitrary code execution on the host OS. The vulnerability combines a Host header bypass of localhost-only routes with unvalidated MCP plugin args passed to child_process.spawn(), enabling command execution via malicious plugins.

In 9Router prior to version 0.5.4, the PATCH /api/settings endpoint writes the entire request body to persistent settings without a field whitelist. This allows an authenticated user to set security-critical fields such as requireLogin and disable authentication for the whole application, exposing protected routes to unauthenticated access.

In 9Router prior to version 0.5.6, the Kiro API-key validation endpoint POST /api/oauth/kiro/api-key builds an upstream URL using a user-controlled region value, allowing an authenticated attacker to supply a crafted region such as kiro-canary.local:8443# and cause 9Router to send the Kiro validation request to an attacker-controlled host while forwarding the submitted Kiro API key as an Authorization header.

In 9Router version 0.4.45 and earlier, the isLocalRequest() function in src/dashboardGuard.js uses Host and Origin headers to protect /api/mcp/*, /api/tunnel/*, and /api/cli-tools/* endpoints. This allows bypassing security through header spoofing in reverse proxy or tunnel deployments, enabling access to MCP child process stdin paths.

9Router through version 0.4.41 contains an unauthenticated information disclosure vulnerability that allows remote attackers to access sensitive user data by sending requests to unprotected API endpoints. Attackers can enumerate paginated request logs and retrieve complete AI conversation histories including system prompts, user messages, assistant responses, tool calls, and user email addresses by querying the request-logs and request-details API routes which lack authentication middleware.

9Router prior to version 0.5.2 has a vulnerability where it determines whether a /v1 LLM proxy request is local by reading the client-controlled Host header, allowing a remote unauthenticated attacker to send Host: localhost and bypass API-key authentication.

9Router prior to version 0.5.2 is vulnerable to a DNS rebinding attack. Image URL validation resolves the host before fetching, but the actual fetch performs a separate DNS resolution, allowing an attacker to redirect requests to internal HTTP services.