Advisory

CVE-2026-114713: UnrealIRCd Label-Tag Injection Lets Users Execute Server-Authority Commands

A delimiter-neutralization flaw in UnrealIRCd's labeled-response module allows any authenticated IRC user on a network with two or more linked servers to inject server-to-server commands that run with full server authority. CISA added this vulnerability to the Known Exploited Vulnerabilities catalog on 2026-11-13, making rapid patching an urgent priority.

TL;DR: Upgrade UnrealIRCd to version 6.2.7.1 or later immediately; if patching is not possible, follow CISA BOD 26-04 guidance and consider taking affected linked-server networks offline until a fix is applied.

What Is the Vulnerability?

CVE-2026-114713 is classified under CWE-140 – Improper Neutralization of Delimiters. The flaw resides in the labeled-response module of UnrealIRCd, which implements the IRCv3 label message tag feature. When a client sends a command bearing a label tag, the server echoes that label back inside its reply. The affected versions of UnrealIRCd do not sanitize or escape the label value before embedding it into server-to-server (S2S) protocol messages.

Because IRC's S2S wire format relies on specific delimiter characters to separate message components, an unsanitized label containing those delimiters is interpreted by receiving servers as additional, distinct protocol commands. Those injected commands are processed as if they originated from a trusted server peer — that is, they carry server authority, bypassing normal user-privilege checks entirely.

The CVSS 3.1 score is 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. Key attributes of that score are worth unpacking:

  • Network-accessible, low complexity: No special network position is required; any reachable IRC user can attempt the injection.
  • No privileges required, no user interaction: The attacker needs only a standard IRC connection — no operator status or social engineering of another party.
  • Changed scope: The impact extends beyond the attacker's own session into the broader server network.
  • High integrity impact, no confidentiality or availability impact: The primary risk is unauthorized modification of server state — for example, forcing channel-mode changes, kicking users, or altering server configuration — rather than data disclosure or denial of service.

Who and What Is Affected?

The vulnerability affects UnrealIRCd versions 6.0.0 up to but not including 6.2.7.1. Specifically:

  • The flaw is only exploitable when two or more servers are linked in an IRC network — single-server (standalone) deployments face a materially reduced risk because there is no S2S channel for injected commands to traverse.
  • Any client user account on a vulnerable linked network can trigger the injection; no elevated IRC privileges are required.
  • The labeled-response capability is part of the IRCv3 specification and is enabled in affected UnrealIRCd builds by default, so there is no non-default configuration needed to be exposed.

Organizations running IRC infrastructure — including gaming communities, open-source projects, hobbyist networks, and enterprises that use IRC for internal communication — should audit their UnrealIRCd version immediately.

What Is the Real-World Impact?

An attacker who successfully exploits this flaw can inject arbitrary server-to-server protocol commands that are executed with server authority. Practical consequences can include:

  • Privilege escalation on the IRC network: Granting IRC operator status to arbitrary accounts or revoking it from legitimate administrators.
  • Channel manipulation: Forcing mode changes, topic changes, kicks, or bans network-wide.
  • User impersonation: Spoofing server-originated messages that appear to come from other users or servers.
  • Network integrity compromise: Altering server-to-server trust relationships or injecting false routing information.

CISA's characterization in the KEV catalog — added 2026-11-13 — confirms that this vulnerability is being actively exploited in the wild. Organizations should treat this as requiring immediate remediation rather than routine patch-cycle scheduling.

While the CVSS vector shows no direct confidentiality impact, an attacker with server-level command authority may be able to chain this access with other techniques to read private messages or channel logs depending on the network's configuration; that risk should be considered during your own impact assessment.

What to Do Now

The following actions are listed in order of priority:

  1. Upgrade to UnrealIRCd 6.2.7.1 or later. This is the vendor-confirmed fix. Obtain the release from the official UnrealIRCd project. Refer to the vendor's own release notes and advisory for upgrade instructions specific to your platform.
  2. If immediate upgrading is not possible, isolate or shut down linked-server configurations. Because the exploit requires a multi-server network, temporarily de-linking servers removes the S2S attack surface while a maintenance window is arranged.
  3. Follow CISA BOD 26-04 patching timelines. CISA's required action explicitly references BOD 26-04 (Prioritizing Security Updates Based on Risk). Federal civilian agencies must comply with those timelines. Non-federal organizations are strongly encouraged to adopt the same urgency given KEV status.
  4. Review server logs for anomalous S2S traffic. CISA's required action also references forensics triage requirements. Examine server-to-server log data for unexpected command sequences, unusual mode changes, or privilege modifications that cannot be attributed to known administrators.
  5. If mitigations are unavailable and the product cannot be patched, CISA advises discontinuing use. This applies particularly to internet-exposed deployments.

Primary references: NVD entry for CVE-2026-114713; CISA Known Exploited Vulnerabilities Catalog; UnrealIRCd official project site for release downloads.

How External Security Testing Surfaces This Exposure

Identifying exposure to CVE-2026-114713 from an external vantage point involves several complementary techniques that a security assessment team would employ:

  • Version fingerprinting: IRC servers typically announce their software version in the 004 numeric reply during connection handshake. An assessor connecting to port 6667 (or the configured IRC port) can read the version string and compare it against the affected range (6.0.0 – 6.2.7.0).
  • Capability negotiation inspection: Using IRCv3 CAP LS negotiation, an assessor can confirm whether the labeled-response capability is advertised, establishing that the vulnerable module is active.
  • Network topology probing: Legitimate IRC commands such as LINKS or MAP (where permitted) reveal whether the server is part of a multi-server network, confirming the prerequisite for exploitability.
  • Configuration and patch audit: On the host side, reviewing installed package versions or binary signatures against the patched version confirms whether the fix has been applied, complementing network-layer checks.

These non-destructive checks establish whether a given UnrealIRCd deployment is in the vulnerable range and whether the network topology meets the exploitation precondition, without requiring any attempt to actually inject commands.

Sources & further reading: NVD CVE-2026-114713 and the CISA KEV catalog are the authority on affected and fixed versions. Vendor/independent analysis: forums.unrealircd.org, unrealircd.org, unrealircd.org, openwall.com. This advisory contains no exploit code.

Frequently asked questions

Is a standalone UnrealIRCd server (not linked to other servers) at risk?

The NVD description specifies that exploitation requires a network with two or more linked servers. A truly standalone, single-server deployment lacks the server-to-server channel that makes the injected commands dangerous. However, upgrading to 6.2.7.1 or later is still the correct action because network topology can change and the underlying code defect remains present in unpatched versions.

Does an attacker need IRC operator privileges to exploit this?

No. The CVSS vector records 'Privileges Required: None,' and the NVD description confirms that 'any user on a network' can trigger the flaw. A standard IRC user account is sufficient.

Is this vulnerability being actively exploited?

Yes. CISA added CVE-2026-114713 to the Known Exploited Vulnerabilities catalog on 2026-11-13, which indicates confirmed active exploitation. Organizations should not wait for a routine patch cycle.

What if disabling the labeled-response module is an option — does that mitigate the risk?

The vendor-recommended fix is upgrading to version 6.2.7.1 or later. While disabling the affected module could theoretically remove the specific attack vector, this approach has not been confirmed by the vendor as a supported mitigation in the available facts. Any configuration-only workaround should be validated against the vendor's own advisory before relying on it, and upgrading remains the definitive remediation.