The rise of AI-enabled hacking has led to a significant shift in cybersecurity, with the focus extending beyond large language models to the development of bespoke tools that harness their power. Enterprises are now creating their own technology platforms to customize these models, mitigating risks and controlling their behavior. This "harness" is crucial in limiting the potential damage of AI-enabled attacks, which can exploit vulnerabilities such as CVE-2022-30190. By building these tailored tools, organizations can better defend against autonomous hacking threats. The harness is not just a theoretical concept, but a practical solution that industry professionals are leveraging to stay ahead of emerging threats1. This matters to cybersecurity practitioners because it highlights the need to prioritize the development of customized tools that can effectively harness the power of AI, rather than relying solely on general-purpose models.