UpSnap flaws allowed superuser takeover and remote code execution
Vulnerability Disclosure · 3 sources · Updated 3mo ago · First seen Aug 13, 2026

UpSnap disclosed and patched two high-severity vulnerabilities in version 5.4.0 affecting the wake-on-LAN web application. One flaw, tracked as CVE-2026-49819, affected fresh installations from versions 4.4.1 through 5.3.5 and allowed an unauthenticated network-adjacent attacker to complete the initial setup flow first, claim the first superuser account, and obtain a long-lived JWT. The issue could then be chained into root-level remote code execution through the application's wake functionality, yielding a CVSS 9.8 rating.
A second flaw, CVE-2026-49481, was an authenticated OS command injection bug in device management that let a low-privileged user inject shell metacharacters into IP or MAC fields used by wake_cmd and shutdown_cmd, leading to command execution via /bin/sh -c on Linux or cmd /C on Windows. UpSnap said the initial setup issue does not affect instances that were already fully initialized, while the command injection risk primarily impacts deployments where untrusted users can create or edit devices. The project mitigated the bugs by moving first superuser creation to a one-time PocketBase link shown only in server console logs, validating IP and MAC values in the backend, adding PocketBase regex constraints, and enforcing browser-side input patterns.


