Cisco Systems Inc. has unveiled new features aimed at enhancing the security of artificial intelligence agents, debuting at the RSAC 2026 Conference. Key enhancements include the open-source tool DefenseClaw, designed to identify vulnerabilities in AI agents. Customers will benefit from improved regulatory capabilities through Duo IAM, which enables the registration of AI agents, allowing administrators to define access rules for various tools. This includes specific interactions with sensitive data and time-limited access, thereby minimizing hacking opportunities. Cisco also introduced cybersecurity developments for its Splunk platform, including sophisticated detection tools and automated task management. Additionally, DefenseClaw scans for malicious code in AI software outputs and tracks cybersecurity risks over time. Alongside DefenseClaw, Cisco released two other complimentary cybersecurity tools: the LLM Security Leaderboard and AI Defense: Explorer Edition, aimed at assessing large language models’ resilience against attacks. These initiatives strengthen Cisco’s commitment to AI security and cybersecurity innovation.
Source link
