BlinkOps announced BlinkOS on November 19, a platform it calls a secure operating system for agentic security operations. The company says it combines one agent layer, one shared context layer and one action layer across a security stack, with more than 30,000 pre-built integrations. It is available now and is listed on the Microsoft Azure Marketplace.

The design choice a SOC buyer should look at is the split between reasoning and action. BlinkOps says agents decide what needs to happen, while execution runs through predefined, governed workflows. Its release lists agents with narrow responsibilities and no direct access to credentials, workflows that run on Blink-managed runners or on customer infrastructure, and least-privilege access, tenant isolation, human approvals and audit trails. In the release, co-founder and CEO Gil Barak says: “Agents provide the intelligence, deterministic workflows provide the control, and security teams can finally scale their operations without scaling their headcount.”

The company also says the platform can bring in outside models, and gives the example of a customer with access to Anthropic’s Cyber Verification Program. The release does not say how many customers use BlinkOS or give performance figures, so those claims are not yet testable.

In our reading, the part worth testing in a pilot is the execution layer. An agent that can reach many tools is only as safe as the permissions on the path between its decision and the action. Ask any vendor in this category to show which actions run without approval, where the credentials sit, and what an analyst sees in the audit trail for a single automated containment. The same questions apply to the SOAR products that hold response credentials today. Splunk’s November advisories rate its SOAR product Critical, and Rapid7 reported exploitation of central tools in its Apache Axis report on Confluence and N-central.

Source: BlinkOps, BlinkOS announcement on PR Newswire