Security Advisory

CVE-2024-10525: Mosquitto Client SUBACK Heap Overflow

Last updated:
  • CVE-2024-10525

CVE-2024-10525 is an out-of-bounds memory access in the Eclipse Mosquitto client library: a malicious or spoofed MQTT broker that answers a subscription with a SUBACK carrying no reason codes makes libmosquitto read past its buffer. NVD scores it 9.8, Mosquitto 2.0.19 fixes it, and CERT@VDE lists ADS-TEC IRF industrial firewalls as affected.

Vulnerability summary

CVSS base score

9.8

Critical severity

EPSS score (FIRST)

59.2%

Probability of exploitation in the wild in the next 30 days

CISA KEV catalog

No

Not listed

Published

Vulnerability ID: CVE-2024-10525 · Source: NVD

Affected components

ComponentCPEVersion range
Eclipse Mosquitto client library (libmosquitto, mosquitto_sub, mosquitto_rr)cpe:2.3:a:eclipse:mosquitto:*:*:*:*:*:*:*:*1.3.2 up to (excluding) 2.0.19; fixed in 2.0.19 (2 October 2024)
ADS-TEC IRF1000 industrial firewalls (DVG-IRF1401, DVG-IRF1421)—firmware before 2.1.0; fixed in 2.1.0 (CERT@VDE VDE-2025-033)
ADS-TEC IRF2000 industrial firewalls (DVG-IRF2100, DVG-IRF2200, DVG-IRF2220, DVG-IRF2601, DVG-IRF2621)—firmware before 6.1.0; fixed in 6.1.0 (CERT@VDE VDE-2025-033)
ADS-TEC IRF3000 industrial firewalls (DVG-IRF3401, DVG-IRF3421, DVG-IRF3801, DVG-IRF3821)—firmware before 2.1.0; fixed in 2.1.0 (CERT@VDE VDE-2025-033)

What does this mean for affected devices?

CVE-2024-10525 reverses the usual MQTT threat model: the victim is the client, and the attacker is the broker it connects to. NVD's description says that in Mosquitto 1.3.2 through 2.0.18, "if a malicious broker sends a crafted SUBACK packet with no reason codes, a client using libmosquitto may make out of bounds memory access when acting in its on_subscribe callback", and that this affects the mosquitto_sub and mosquitto_rr tools. NVD scores it 9.8 (critical), vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H, and maps it to CWE-787; Eclipse, the assigning CNA, filed CWE-122 and a CVSS v4.0 score of 7.2 (NVD).

The fix is commit 8ab20b4, "Don't allow SUBACK with missing reason codes", shipped in Mosquitto 2.0.19 on 2 October 2024. The 2.0.19 release notes list it as a client-library security fix next to the broker fixes for CVE-2024-8376 and CVE-2024-3935.

The device angle comes from CERT@VDE advisory VDE-2025-033, published 14 April 2025 by ADS-TEC Industrial IT. It states that the IRF1000, IRF2000 and IRF3000 firewall products "include Eclipse Mosquitto" and lists eleven models: DVG-IRF1401 and DVG-IRF1421, DVG-IRF2100, DVG-IRF2200, DVG-IRF2220, DVG-IRF2601 and DVG-IRF2621, and DVG-IRF3401, DVG-IRF3421, DVG-IRF3801 and DVG-IRF3821. The fix is firmware 2.1.0 for IRF1000 and IRF3000 and 6.1.0 for IRF2000. ADS-TEC rescores the CVE to 5.6 in its product context (AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:L) because "attacks on the product must be carried out via the MQTT server" and the client runs sandboxed with reduced privileges. Its interim mitigation is to disable MQTT publishing or connect only to trusted, TLS-secured brokers.

That mitigation is the general lesson. A gateway that authenticates its broker with TLS and a pinned CA cannot be fed a SUBACK by a spoofed server. Our MQTT protocol security guide covers broker authentication and lists this CVE among the Mosquitto library records, and the memory corruption fix guide covers bounds checks on length fields received from the network.

Debian fixed it in bullseye 2.0.11-1+deb11u2 (DLA-4059-1) and bookworm 2.0.11-1.2+deb12u2, so a 2.0.11 package string alone does not prove exposure (Debian Security Tracker).

CVE-2024-10525 is not in the CISA KEV catalog as of 2026-11-17 (catalog version 2026.11.13). FIRST EPSS gives it 0.5922, the 99th percentile, in the 2026-11-16 dataset, which is high for a client-side flaw that needs a hostile broker (CISA KEV, FIRST EPSS).

To fix it, update ADS-TEC IRF devices to firmware 2.1.0 (IRF1000, IRF3000) or 6.1.0 (IRF2000). In your own firmware, ship libmosquitto 2.0.19 or later and record the library version per release in a firmware SBOM.

How to fix this class of issueThis advisory's weakness class is Memory corruption. Our practical guide covers how to detect, fix, verify and disclose it: How to fix memory corruption in embedded firmware.

Evidence

  • Malicious broker's SUBACK with no reason codes causes out-of-bounds access in libmosquitto 1.3.2 through 2.0.18; affects mosquitto_sub and mosquitto_rr; NVD CVSS 3.1 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H), CWE-787; Eclipse CVSS 4.0 7.2, CWE-122. — NVD
  • Fixed by commit 8ab20b4 "Don't allow SUBACK with missing reason codes" in Mosquitto 2.0.19, released 2 October 2024. — Eclipse Mosquitto
  • ADS-TEC IRF1000/IRF3000 firmware before 2.1.0 and IRF2000 before 6.1.0 include Eclipse Mosquitto and are affected; product-context score 5.6; mitigation: disable MQTT publishing or use only trusted, TLS-secured brokers. — CERT@VDE VDE-2025-033
  • Debian fixed versions: bullseye 2.0.11-1+deb11u2 (DLA-4059-1), bookworm 2.0.11-1.2+deb12u2. — Debian Security Tracker
  • Not in the CISA KEV catalog as of 2026-11-17 (catalog version 2026.11.13). — CISA KEV
  • EPSS 0.59222 (99th percentile), dataset 2026-11-16, read 2026-11-17. — FIRST EPSS

Related advisories

Sources

  1. NVD
  2. CERT@VDE VDE-2025-033: Mosquitto MQTT Client Vulnerability in ADS-TEC IRF Products
  3. Eclipse Mosquitto: Version 2.0.19 released
  4. Mosquitto commit 8ab20b4: Don't allow SUBACK with missing reason codes
  5. Debian Security Tracker
  6. CISA KEV
  7. FIRST EPSS

Frequently asked questions

Who can exploit CVE-2024-10525?

The MQTT broker the client connects to, or anyone who can impersonate it. The bug is in how libmosquitto handles a SUBACK with no reason codes, so a device is at risk when it connects to a malicious or spoofed broker (per NVD and CERT@VDE VDE-2025-033).

Which devices are affected by CVE-2024-10525?

CERT@VDE lists ADS-TEC IRF1000 and IRF3000 firewalls before firmware 2.1.0 and IRF2000 firewalls before 6.1.0, eleven DVG-IRF models in total. Any firmware that ships libmosquitto 1.3.2 through 2.0.18 and subscribes to topics carries the same code (per CERT@VDE and NVD).

Why does NVD score CVE-2024-10525 9.8 while ADS-TEC scores it 5.6?

NVD rates it network-reachable with no privileges and high impact. ADS-TEC raises attack complexity because the attacker must control the MQTT server, and lowers impact because the client runs sandboxed with reduced privileges (per NVD and CERT@VDE VDE-2025-033).

Is CVE-2024-10525 in the CISA KEV catalog?

No, not as of 2026-11-17 (catalog version 2026.11.13). FIRST EPSS gives it 0.5922, the 99th percentile, in the 2026-11-16 dataset (per CISA KEV and FIRST EPSS).