The Language of Security: How Prompt Syntax Shapes Secure Code Generation in Open LLMs

Large Language Models LLMs are increasingly used for source code generation despite their outputs often exhibiting security vulnerabilities. Prior work shows that prompt engineering can mitigate such risks, yet 1 they focused on high-level prompting strategies, neglecting recent evidence that...

CVE-2026-32915 OpenClaw < 2026.3.11 - Sandbox Boundary Bypass via Subagent Control Surface

OpenClaw before 2026.3.11 contains a sandbox boundary bypass vulnerability allowing leaf subagents to access the subagents control surface and resolve against parent requester scope instead of their own session tree. A low-privilege sandboxed leaf worker can steer or kill sibling runs and cause...

CVE-2026-32915 OpenClaw < 2026.3.11 - Sandbox Boundary Bypass via Subagent Control Surface

OpenClaw before 2026.3.11 contains a sandbox boundary bypass vulnerability allowing leaf subagents to access the subagents control surface and resolve against parent requester scope instead of their own session tree. A low-privilege sandboxed leaf worker can steer or kill sibling runs and cause...

CVE-2026-32915 OpenClaw < 2026.3.11 - Sandbox Boundary Bypass via Subagent Control Surface

OpenClaw before 2026.3.11 contains a sandbox boundary bypass vulnerability allowing leaf subagents to access the subagents control surface and resolve against parent requester scope instead of their own session tree. A low-privilege sandboxed leaf worker can steer or kill sibling runs and cause...

CVE-2026-32915

OpenClaw before 2026.3.11 contains a sandbox boundary bypass vulnerability allowing leaf subagents to access the subagents control surface and resolve against parent requester scope instead of their own session tree. A low-privilege sandboxed leaf worker can steer or kill sibling runs and cause...

CVE-2026-32915

OpenClaw contains a sandbox boundary bypass vulnerability affecting versions before 2026.3.11. A low‑privilege, sandboxed leaf subagent can access the subagent control surface and resolve against the parent requester scope instead of its own session tree, enabling steering or killing of sibling r...

CVE-2026-32915: Incorrect Authorization

OpenClaw before 2026.3.11 contains a sandbox boundary bypass vulnerability allowing leaf subagents to access the subagents control surface and resolve against parent requester scope instead of their own session tree. A low-privilege sandboxed leaf worker can steer or kill sibling runs and cause...

GHSA-4W7M-58CG-CMFF OpenClaw: Leaf subagents could steer sibling sessions across sandbox boundaries

Summary In affected versions of openclaw, sandboxed leaf subagents could still access the subagents control surface and resolve against the parent requester scope instead of remaining confined to their own session tree. Impact A low-privilege sandboxed leaf worker could steer or kill a sibling ru...