← Back to feed
AI SecurityEmerging1 sourceAug 31, 2026 · 11:22via CyberPress

OpenClaw 2.0 Released With Enhanced AI Agent Security and Credential Protection

Brief

OpenClaw has shipped version 2026.

  • 1, branded OpenClaw 2. 0, the largest update in the open-source AI agent platform’s history, built by 933 contributors across more than 16,000 pull requests.

The release introduces a substantial security overhaul focused on credential handling and agent trust boundaries, as well as a rebuilt browser Control UI and shared multiplayer sessions.

The headline security feature allows an agent to request a credential via a masked prompt, so the secret’s value never enters the chat transcript or the model’s context window.

OpenClaw 2.0 Released With Enhanced AI Agent Security

Users can enable an opt-in proxy that restricts substitution of these protected secrets to explicitly approved destination hosts.

Read more on CyberPress