OpenClaw founder discusses privacy and security challenges

OpenClaw founder Peter Steinberger recently highlighted the platform's commitment to privacy and security during an interview, emphasizing that all user data is stored locally without any cloud uploads. This approach allows users to maintain full control over their access permissions and memory data, which is crucial in an era where AI automation and agent workflows are increasingly prevalent. Despite these strengths, Steinberger acknowledged ongoing security challenges, particularly concerning AI agent vulnerabilities and the risks posed by prompt injection attacks.

As of February 2026, OpenClaw identified 341 malicious plugins in its skills marketplace, reflecting an 11.3% contamination rate that raises significant supply chain concerns. To address these threats, OpenClaw has engaged contract security experts to enhance its defenses and assist users in navigating emerging risks. The founder underscored that while OpenClaw offers robust privacy solutions, the security of AI agents is heavily reliant on the underlying model's capabilities, necessitating continuous model updates and vigilant monitoring of integrations and developer tooling.