Northstar launches with $7M to help AI agents repair security risks
Northstar joins data from more than 70 security and IT tools in one map, and its agents start the repair.
Northstar Security AI, a cybersecurity startup, launched with $7 million in seed funding, which Wing Venture Capital led. Its platform joins data from the security and IT tools of a company in one live map. The map is a Security Context Graph. AI agents use the map to find risks and start repairs. Customers can use the platform at this time.
How it works
A known weak point, a device with no control or an account with too much access can look safe when seen one by one. Together, they can cause a risk to a primary system. In one example, a laptop with no endpoint protection belongs to a user who can use the payroll system. Its agents can start a workflow to install protection, and persons can supervise the steps.
The numbers
Northstar works with more than 70 systems, with those from Okta, GitHub and Workday. The company gave no test results. Randy Barr of Cequence Security says that its team uses it to repair problems.
What is not known
The source does not tell how frequently agents work without a person. External agents, such as Claude from Anthropic, can use the graph through MCP. MCP is a connection that lets AI tools use other software.
Insight: A map of weak points is open to external agents
Northstar opens its map of company weak points to external agents through MCP. BeyondTrust data from 21 October 2026 show that 85% of the coding agent user that it saw had an MCP server. This reads as a move to make the map a step that agents must use. The company that holds the map gets the control. Give each MCP connection to such a map the control of an admin account.
Sources
Posted
