IronClaw, an open-source AI agent solution, was launched to enhance security for users relying on personal AI assistants by utilizing NEAR AI’s cryptographically secure infrastructure. Unlike its predecessor OpenClaw, which exposed users to risks like prompt injections that could compromise sensitive information, IronClaw operates within encrypted enclaves and employs a defense-in-depth security model. It stores credentials in a secure vault that never interacts with the AI directly and features real-time leak detection, ensuring that user data remains protected against unauthorized access. Built entirely in Rust, IronClaw also enforces compile-time memory safety, minimizing the likelihood of common exploits.
IronClaw secures AI agents with NEAR AI’s hardware-enforced security
