A Blind Trust, the Bloody Thrust: When Attacker-Controlled Hook Updates Steer AI Agent Harnesses Towards Malicious Behaviors

Modern AI agent harnesses expose lifecycle hooks that bind shell commands to runtime events such as session start, tool calls, and file edits. These commands run with host privileges yet ship as lifecycle-hook configuration and may fire at times the LLM never observes. We identify the...

container-tools:ol8 security, bug fix, and enhancement update

aardvark-dns 2:1.10.1-2 - build off the RHEL maintenance branch - Resolves: RHEL-59129 buildah 2:1.33.14-3 - rebuild for CVE-2025-68121 - Resolves: RHEL-149262 cockpit-podman 84.1-1 - update to https://github.com/cockpit-project/cockpit-podman/releases/tag/84.1 - Related: Jira:RHEL-25557 conmon...