**CVE-2023-50257** is an eProsima Fast DDS flaw, published by NVD on 2024-02-19, that lets an attacker on the same network forge the unencrypted disconnect message for a publisher and drop every subscriber attached to it, even with SROS2 security enabled. NVD scores it 8.1. The fix shipped in Fast DDS 2.13.0, 2.12.2, 2.11.3, 2.10.3 and 2.6.7.
Vulnerability summary
CVSS base score
8.1
High severity
EPSS score (FIRST)
0.5%
Probability of exploitation in the wild in the next 30 days
CISA KEV catalog
No
Not listed
Published
Vulnerability ID: CVE-2023-50257 · Source: NVD
Affected components
| Component | CPE | Version range |
|---|---|---|
| eProsima Fast DDS | cpe:2.3:a:eprosima:fast_dds:*:*:*:*:*:*:*:* | Before 2.13.0, 2.12.2, 2.11.3, 2.10.3 and 2.6.7 (CVE record and GHSA-v5r6-8mvh-cp98); NVD's CPE ranges cover before 2.6.7 and the 2.10, 2.11 and 2.12 branches only |
What does this mean for affected devices?
NVD's description says that "even with the application of SROS2", the data (p[UD]) and guid values used to disconnect nodes are not encrypted, so an attacker who sends the disconnect packet to the Global Data Space (239.255.0.1:7400) with a publisher's ID makes all of its subscribers stop receiving data, and repeated packets keep new subscribers from connecting. NVD's Primary score is 8.1, vector CVSS:3.1/AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:H: adjacent network, no privileges, high integrity and availability impact. GitHub, the CNA, scores it 9.6 with a changed scope. NVD records the weakness as NVD-CWE-noinfo; the CNA lists CWE-284, improper access control (NVD).
eProsima's advisory GHSA-v5r6-8mvh-cp98 shows why security did not help. The researchers built ROS 2 Iron with -DSECURITY=ON and found that Data(p[UD]) calls remove_participant, which deletes the publisher's writers and readers, while the SecurityManager keeps that participant's tokens. Because packets sent to the Global Data Space were not encrypted, "an attacker within the same Global Data Space can eavesdrop on the Publisher's Participant ID" and then impersonate it. NVD's description dates the flaw to the first SecurityManager.cpp commit of 2016-11-08.
The fixed releases are 2.13.0 (2023-12-19), 2.6.7 (2024-01-17), 2.10.3 (2024-01-30), 2.11.3 (2024-02-19) and 2.12.2 (2024-03-22); each release note lists CVE-2023-50257. The advisory also names the ROS 2 middleware packages that carried vulnerable builds (rmw_fastrtps 7.1.1 with Fast DDS 2.10.2, and 6.2.3 with 2.6.6) and says ROS 2 Galactic (Fast DDS 2.3.6) and Foxy (2.1.4) get no fix because both distributions reached end of life. NVD's CPE ranges stop at four branches: before 2.6.7, then 2.10, 2.11 and 2.12. No range covers 2.7.0 to 2.9.x. Debian's security tracker lists bookworm's fastdds 2.9.1+ds-1+deb12u2 as vulnerable (no DSA, minor issue), and trixie (3.1.2) and sid as fixed, so a scanner that relies only on NVD's CPE match misses a Fast DDS 2.9 build.
The device angle comes from where the library runs. eProsima's platform support table lists QNX 7.1 on arm64 and arm32 next to Linux and Android, and Fast DDS is the DDS layer in many ROS 2 robots, controllers and gateways. As of 2026-11-16 no device vendor advisory names products affected by this CVE, and a search of CISA's CSAF repository finds no ICS advisory for it, so the Fast DDS version in your own firmware is what decides exposure.
CVE-2023-50257 is not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). FIRST EPSS gives it 0.0048, the 41.8th percentile, in the 2026-11-16 dataset. CISA's SSVC entry on the NVD record reads exploitation "poc", automatable "no", technical impact "total" (CISA KEV, FIRST EPSS).
For a robot or controller that ships Fast DDS, read the exact version from the firmware SBOM, move to a fixed release on your branch (or a supported branch if you are on 2.7 to 2.9 or an end-of-life ROS 2 distribution), and keep the DDS domain on a segment where untrusted hosts cannot reach the discovery multicast group. Our DDS security guide explains SROS2, governance files and RTPS discovery defaults, and the authentication bypass fix guide covers message-origin checks. Other Fast DDS advisories: CVE-2024-28231, CVE-2025-62799 and the 2021 amplification flaw CVE-2021-38425.
Evidence
- p[UD] and guid not encrypted even with SROS2; disconnect via 239.255.0.1:7400; present since SecurityManager.cpp commit of 2016-11-08; fixed in 2.13.0, 2.12.2, 2.11.3, 2.10.3, 2.6.7; NVD Primary 8.1 (AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:H); GitHub CNA 9.6; NVD-CWE-noinfo; CNA CWE-284; CPE ranges <2.6.7, 2.10.0-<2.10.3, 2.11.0-<2.11.3, 2.12.0-<2.12.2; CISA ADP SSVC poc/no/total. — NVD
- ROS 2 Iron built with -DSECURITY=ON; Data(p[UD]) invokes remove_participant; SecurityManager keeps tokens; attacker in the same Global Data Space can eavesdrop the Publisher's Participant ID; rmw_fastrtps 7.1.1 (Fast DDS 2.10.2) and 6.2.3 (2.6.6) affected; Galactic (2.3.6) and Foxy (2.1.4) no fix, end of life. — eProsima GHSA-v5r6-8mvh-cp98
- Release dates from GitHub releases API: 2.13.0 2023-12-19, 2.6.7 2024-01-17, 2.10.3 2024-01-30, 2.11.3 2024-02-19, 2.12.2 2024-03-22; each body lists CVE-2023-50257. — eProsima Fast DDS releases
- bookworm 2.9.1+ds-1+deb12u2 vulnerable (no-dsa, minor issue); trixie 3.1.2+ds-1 fixed; sid 3.3.0+ds-3 fixed; unstable fixed in 2.14.0+ds-2 (bug 1064515). — Debian security tracker
- Platform table lists QNX 7.1 (amd64, arm64 Tier 3) and arm32/arm64 Linux and Android. — Fast DDS PLATFORM_SUPPORT.md
- No CISA ICS advisory in the cisagov/CSAF repository mentions CVE-2023-50257 (checked 2026-11-13). — cisagov/CSAF
- Not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). — CISA KEV
- EPSS 0.00478 (41.8th percentile), dataset 2026-11-16. — FIRST EPSS

Related advisories
Sources
Frequently asked questions
Does SROS2 protect against CVE-2023-50257?
No. NVD's description says the disconnect data and GUID values are not encrypted "even with the application of SROS2", and the researchers reproduced the attack on a ROS 2 Iron build with security turned on (per NVD and GHSA-v5r6-8mvh-cp98).
Which Fast DDS versions fix CVE-2023-50257?
2.13.0, 2.12.2, 2.11.3, 2.10.3 and 2.6.7. The ROS 2 Galactic and Foxy builds (Fast DDS 2.3.6 and 2.1.4) get no fix because those distributions reached end of life (per eProsima's advisory and release notes).
Why might a scanner miss CVE-2023-50257 on Fast DDS 2.9?
NVD's CPE ranges cover before 2.6.7 and the 2.10, 2.11 and 2.12 branches only, so 2.7.0 to 2.9.x match nothing. Debian lists its 2.9.1 package in bookworm as vulnerable (per NVD and the Debian security tracker).
Is CVE-2023-50257 exploited in the wild?
It is not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). FIRST EPSS gives it 0.0048, the 41.8th percentile, in the 2026-11-16 dataset (per CISA KEV and FIRST EPSS).
Questions about this page?