The label is a note a client sticks on its message so it can tell which reply is its own. The server sends it back. UnrealIRCd sends it back without escaping it, and that is where the hole is.
Fact
The facts: on 8 November 2026 NVD published CVE-2026-114713, a vulnerability in UnrealIRCd's labeled-response module, with the record from MITRE. Per the description, the module does not escape the IRCv3 "label" tag when echoing it into replies, and any user on a network with two or more linked servers can inject server-to-server commands that are executed with server authority. The score is MITRE's as CNA: CVSS 3.1 base score 8.6 HIGH with the vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:N/I:H/A:N. NVD has not yet given its own. The weakness is CWE-140, improper neutralization of delimiters, named by MITRE. Affected are UnrealIRCd releases from 6.0.0 up to a version that is masked in NVD's text. According to NVD, CISA added CVE-2026-114713 to its catalog of known exploited vulnerabilities on 13 November 2026, with a due date of 16 November 2026.
Read it again: any user. Not an admin, not an operator, not someone with an account. It is enough to be on a network with two linked servers, and the command you push into the label travels between them with server authority. That is the difference between a nuisance and a hole.
The label exists so you can match a reply. Here it becomes a command.
The listing in CISA's catalog is the second half. It means the hole is being exploited, not that it is theory. The date added is 13 November, the due date is 16 November. Three days.
Update to the fixed release per the vendor's instructions. A network with two linked servers is enough for the command to go through.
Sources
NVD - CVE-2026-114713, 08.11.2026CISA - Known Exploited Vulnerabilities Catalog, CVE-2026-114713, 13.11.2026The visual is generated code-art, no third-party images.