Apache Roller's Second CVE Widens the Affected Range to 4.0.0, and CISA Lists It With Three Others Due Nov. 23
Security / news
Apache Roller's Second CVE Widens the Affected Range to 4.0.0, and CISA Lists It With Three Others Due Nov. 23
A pre-authentication deserialization flaw first listed only 6.1.5. A vCenter bypass in the same batch was used by a suspected PRC-nexus actor, per Google.
An unauthenticated remote attacker can run code on Apache Roller 4.0.0 through 6.1.5 by sending a crafted request to the XML-RPC endpoint, with no configuration change needed. CISA added the flaw to its Known Exploited Vulnerabilities catalog on Nov. 20, 2026, with a federal due date of Nov. 23. Upgrading to Roller 6.1.6 or later fixes it.
The score is 9.8 (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H) and it is not inflated: the request is parsed before authentication.
The second CVE changed the version range
The Apache Software Foundation announced CVE-2026-82384 on its announce list on Sept. 25, 2026. That record listed one affected version, Roller 6.1.5, and credited a finder who goes by n0mi1k. It said the servlet is mapped unconditionally, so parsing happens even when the global XML-RPC feature is disabled.
NVD's record for CVE-2026-129448, published Nov. 17, calls it "the same vulnerability" and extends the range to 4.0.0 through 6.1.4. It also says the issue "is being actively exploited." One caution: NVD's structured field reads "lessThan 6.1.5", which excludes 6.1.5 itself, while the description and the first CVE cover it. A scanner reading only one record sees part of the range.
The workaround is narrower than it looks. Apache says to set "enabledForExtensions" to "false" in WEB-INF/web.xml, or remove the "/roller-services/xmlrpc" servlet mapping, and restart. Switching XML-RPC off in the administration UI is not enough, because older versions deserialize the request before checking that setting.
Four entries, one due date
CISA's catalog shows four additions on Nov. 20, each due Nov. 23, three days later.
| CVE | Product | NVD base score |
|---|---|---|
| CVE-2026-129448 | Apache Roller | 9.8 |
| CVE-2026-123890 | Broadcom VMware vCenter | 9.8 |
| CVE-2026-76466 | Cisco Expressway Series | 9.8 |
| CVE-2025-42884 | SAP NetWeaver Enterprise Portal | 6.5 |
The SAP flaw scores only 6.5 because it discloses or modifies server information with no availability impact, yet CISA's SSVC data lists exploitation as active and automatable, so the score understates the urgency. The Cisco flaw (published Oct. 21) allows root command execution for an attacker holding persistent cookies, because client certificates are insufficiently validated when TLS Verify Mode is Enabled for clustering. CISA's required action cites its BOD 26-04 directive and forensic triage requirements. The catalog also added a Synology DSM flaw, CVE-2024-10441, on Nov. 18; Synology's separate advisory trouble is in our report on its third critical DSM advisory in 62 days.
Who is exploiting vCenter
Of the four, only vCenter has a public account of an attacker. Google Threat Intelligence Group and Mandiant published on Nov. 19 that ARMORY CASTLE, formerly UNC5221 and described as a suspected PRC-nexus espionage actor, used CVE-2026-123890 in the vmdir RPC interface on port 2012. The flaw gives an authenticated session as administrator@vsphere.local.
Google dates the malware BONESAW to early September 2026 on a vCenter already patched against an earlier flaw, CVE-2026-59309. Its build timestamp of July 13 predates that CVE's July 29 disclosure by 16 days, though Google warns build metadata can be altered. Fixed vCenter builds are 9.1.1.0200, 9.0.2.0200 and 8.0 U3m. Google advises restricting ports 2012, 443, 22 and 5480 to allowlisted addresses.
Operators running Roller should patch to 6.1.6 today and check the web.xml setting if they cannot. For anything on the table above, the date to meet is Nov. 23. For a different 9.8 that behaves less cleanly, see our report on OpenSSL's CVE-2026-121481, where a reviewer called the reproducer synthetic.
Sources
More in Security
- 01Signal Now Blocks Siri AI on iOS 27 and Auto-Enables Screen Security on Android, but Cannot Stop macOS ScreenshotsMeredith Whittaker's Nov. 20 post lists what Signal changed on each platform and where its protection runs out, then asks Apple, Google and Microsoft for an opt-out.
- 02OpenAI's Cyber Tier Answers 95.0% of Advanced Requests; Anthropic's Lowest Tier Blocked 46 of 50 TrialsBoth labs now gate their least-restricted cyber models behind vetting. The two published figures measure different things, and both come from the vendors.
- 03Claude-BugHunter's README Advertises 681 Attack Patterns and Cites 433, and Counts 83 Skills Where the Repo Says 82The 14.8k-star Claude Code bundle for bug hunting is MIT-licensed and carries scope warnings, but its own figures disagree and Anthropic's access tiers decide who can run it.
- 04Windows Server 2028 Preview Hardens Server Core by Default in Build 29706, but NTLM Is Still OpenHundreds of settings are on from first boot, Microsoft warns Server 2025 workloads may break, and the 33 percent of servers still on 2019 are untouched.