Bitdefender Releases Free AI Guardian Beta for macOS

Bitdefender has released a public beta of AI Guardian, a free macOS security tool that checks actions by autonomous software agents before they take effect. The standalone app initially supports Claude Code 2.1.121 and OpenClaw 2026.6.6.

AI Guardian compares agents’ requested operations with rules set by users, then allows, flags or blocks each request. It records decisions for later review. Prompt processing happens locally, though selected services, such as website reputation checks, may use Bitdefender’s cloud infrastructure.

The tool is designed to detect hidden instructions that try to manipulate agents and to stop suspicious or altered Model Context Protocol tools before an agent can use them. It also checks supported agent skills before execution. When API credentials are exposed, AI Guardian can restrict access to protected resources, including SSH keys and system credentials.

Bitdefender cited research that tested 20 AI agents against more than 1,300 tool-poisoning attempts. The attacks succeeded 36.5% of the time; one model was manipulated in 72.8% of attempts. A separate study reported more than 1.2 million exposed AI service secrets during 2025, an 81% annual increase, and identified more than 24,000 credentials exposed through public MCP configurations that year.

The company says AI Guardian is free during the public beta, which is limited to English and macOS. Bitdefender plans to bring it to other operating systems. Its agent-security lineup already includes two other products.