Breaking Update
Israeli Startup Tackles AI Agent Security Risks With Hardened OpenClaw
Israeli cybersecurity firm Minimus has released a secure version of OpenClaw, reducing over 2,000 vulnerabilities by 99% for AI agent deployments.
Jerusalem, 24 February, 2026 (TPS-IL) — Israeli Cybersecurity startup Minimus has released a hardened version of OpenClaw, a fast-growing open-source AI agent tool, after identifying more than 2,000 security vulnerabilities in existing deployments. The company said its minimal container image cuts the number of flaws by 99 percent while preserving full functionality, and is being offered free to the developer community.
OpenClaw allows organizations to run AI Agents locally, integrating with messaging platforms and internal systems, but its rapid, LLM-driven development introduced serious Security Risks. Some companies, including Meta, have barred its use, even as OpenAI recently hired the tool’s creator. Minimus was founded by former Palo Alto Networks executives.
