The PKCS#11 feature in ssh-agent in OpenSSH before 9.3p2 has an insufficiently trustworthy search path, leading to remote code execution if an agent is forwarded to an attacker-controlled system. (Code in /usr/lib is not necessarily safe for loading into ssh-agent.) NOTE: this issue exists because of an incomplete fix for CVE-2016-10009.
https://www.vicarius.io/vsociety/posts/exploring-opensshs-agent-forwarding-rce-cve-2023-38408
https://www.openssh.com/txt/release-9.3p2
https://www.openssh.com/security.html
https://support.apple.com/kb/HT213940
https://security.netapp.com/advisory/ntap-20230803-0010/
https://security.gentoo.org/glsa/202307-01
https://news.ycombinator.com/item?id=36790196
https://lists.debian.org/debian-lts-announce/2023/08/msg00021.html
https://github.com/openbsd/src/commit/f8f5a6b003981bb824329dc987d101977beda7ca
https://github.com/openbsd/src/commit/f03a4faa55c4ce0818324701dadbf91988d7351d
https://github.com/openbsd/src/commit/7bc29a9d5cd697290aa056e94ecee6253d3425f8
http://www.openwall.com/lists/oss-security/2023/09/22/9
http://www.openwall.com/lists/oss-security/2023/09/22/11
http://www.openwall.com/lists/oss-security/2023/07/20/2
http://www.openwall.com/lists/oss-security/2023/07/20/1
http://packetstormsecurity.com/files/173661/OpenSSH-Forwarded-SSH-Agent-Remote-Code-Execution.html