@@ -81,6 +81,26 @@ _Exploiting these vulnerabilities requires access to the system and the ability
8181to execute code there, thus the categorization as Local Privilege Escalation (LPE),
8282not Remote Code Execution (RCE)._
8383
84+ ## Update 2026-05-13: Fragnesia
85+
86+ Yet another network fragment handling issue which fails to properly prevent
87+ in-place en/decryption which can be made to hit sensitive page cache contents
88+ that was put there with splice has been discovered with
89+ [ Fragnesia] ( https://www.openwall.com/lists/oss-security/2026/05/13/3 ) .
90+ It's a simple logic error where the sharing property is forgotten in buffer
91+ coalescing. It has gotten CVE-2026 -46300.
92+
93+ ## Update 2026-05-15: ssh-keysaign-pwn
94+
95+ This LPE is unrelated to the others; under certain circumstances, the kernel
96+ fails to prevent the dumping of process memory under ptraces, which can be used
97+ to read out sensitive data. As it can be used to read e.g. ssh keys, it has been called
98+ [ ssh-keysign-pwn] ( https://github.com/0xdeadbeefnetwork/ssh-keysign-pwn ) . It has
99+ gotten CVE-2026 -46333.
100+
101+ This vulerability allows local attackers to read sensitive data (Information
102+ Disclosure - ID) which may be useful to escalate privileges.
103+
84104## Impact
85105
86106Any system where normal (non-root) users can log in to execute code under their
@@ -137,7 +157,7 @@ A fully effective workaround is again to prevent loading the affected modules
137157by placing another file ` dirtyfrag.conf ` in ` /etc/modprobe.d/ ` :
138158
139159``` shell
140- # Temporary workaround for Dirty Frag CVE-2026-43284, CVE-2026-43500
160+ # Temporary workaround for Dirty Frag CVE-2026-43284, CVE-2026-43500, CVE-2026-46300
141161# This breaks IPsec
142162install esp4 /bin/false
143163install esp6 /bin/false
@@ -146,6 +166,16 @@ install rxrpc /bin/false
146166
147167Note that these workarounds prevent IPsec from working.
148168
169+ Update 2026-05-15: This mitigation helps against Fragnesia as well; AppArmor preventing
170+ unprivileged user namespaces (ubuntu) also helps. While a patch
171+ for Fragnesia was proposed, a fix has not been merged into the upstream kernel yet.
172+ Alma Linux has gone ahead and published a fixed kernel anyhow (for testing).
173+
174+ Update 2026-05-15: The stable kernels 7.0.8, 6.18.31, 6.12.89, 6.6.139, 5.15.207, and
175+ 5.10.256 contain a fix for the ptrace vulnerability (CVE-2026 -46333 aka ssh-keysign-pwn).
176+ The latter is hard to mitigate, as there's no Security Module hook. Setting up a seccomp
177+ that disallows ptrace would help.
178+
149179If a system is suspected to already have been exploited, the system owner can
150180dispose of the page cache by doing ` echo 3 > /proc/sys/vm/drop_caches ` as root
151181and unload the affected modules to prevent re-exploitation.
@@ -237,10 +267,40 @@ in the meantime, as advised above.
237267The SCS community infrastructure was secured on May 8 by disabling the
238268relevant modules.
239269
270+ ## Outlook
271+
272+ The density with which severe issues are currently found is notable. While the
273+ research in the referenced issues originated from human beings being suspicious
274+ that certain things were not handled correctly everywhere, they were assisted
275+ by AI tools in the search. This pattern is likely to turn up a relevant number
276+ of such issues which are reachable now using such help until we may enjoy a
277+ more quiet time again. Until a new class of weaknesses comes into the reach of
278+ yet more powerful AI tools.
279+
280+ The other notable thing is that the quiet closing of issues within the Linux
281+ Kernel no longer works which has broken the responsible disclosure process
282+ that gave defenders some headway against attackers. One reason is that
283+ once a pattern emerges, several teams compete for speed to find related issues
284+ and decide in favor of disclosing before someone else finds it also. Another
285+ reason is that there are people that now use AI tools to analyze merged bugfixes
286+ to the kernel using AI tools - the fact that they are not flagged as security
287+ issues no longer is a good cover. So software distributors need to expect more
288+ of these occurences where they are struggling to patch software, validate the
289+ fixes and ship to customers before large-scale attacks happen. Operators will
290+ need to be fast as well.
291+
292+ So yes, the times are changing in software security and open source software is
293+ hit first being the most accessible to security researchers. On the proprietary
294+ side, we can suspect similar research is happening, but whether the research
295+ teams also go for fame and visibility or maybe have other prevalent incentives
296+ is an open question.
297+
240298## Thanks
241299
242300The authors would like to thank Taeyang Lee at Xint (who initiated the
243301research on copy.fail) and Hyunwoo Kim (@v4bel , who discovered Dirty Frag).
302+ Update 2026-05-15: They would also like to acknowledge William Bowling (V12) for Fragnesia
303+ and Jann Horn and Qualys for ssh-keysign-pwn.
244304They would also like to thank the upstream Linux kernel maintainers and
245305Linux distributors for their reliable work no handling the issues and
246306getting fixes out.
@@ -257,3 +317,5 @@ SCS security contact is [security@scs.community](mailto:security@scs.community),
257317- Mention succssful patching of community infra, v0.3, 2026-05-09, 13:30 CEST.
258318- Correct facts on the failure of the responsible disclosure. Release as v1.0, 2026-05-09, 20:00 CEST.
259319- Update on final rxrpc fix in stable kernels. v1.1, 2026-05-12, 08:45 CEST.
320+ - Add Fragnesia and ssh-keysign-pwn LPEs. v1.2, 2026-05-15, 17:15 CEST.
321+ - Add note on AI changing the game (outlook). v1.3, 2026-05-15, 17:30 CEST.
0 commit comments