Scalable Weak-To-Strong Verifier Hardening for Agent Environments

Agent environments are essential for large language model LLM development. They use programmatic verifiers to turn complex outcomes into rewards. Verifiers can be attacked into awarding full credit to cheating agents without completing the intended task, causing inaccurate evaluation and...

CVE-2025-40169

CVE-2025-40169 concerns the Linux kernel BPF verifier, where the check_alu_op() logic for ALU operations incorrectly allowed negative offsets due to signedness. The vulnerability arises when validating BPF instructions with signed 16-bit off fields; the historical check insn->off > 1 failed...

CVE-2025-40169: Undefined Security Weakness

In the Linux kernel, the following vulnerability has been resolved: bpf: Reject negative offsets for ALU ops When verifying BPF programs, the checkaluop function validates instructions with ALU operations. The 'offset' field in these instructions is a signed 16-bit integer. The existing check...