Northstar Security Context Graph: AI Agents Fix Toxic Risks

Security teams rarely lose because of one dramatic flaw. More often, three ordinary issues line up: an unpatched laptop, an account with too much access and a path to a critical system. Northstar Security, which emerged from stealth on October 28, 2026 with $7 million in seed funding, is building AI agents designed to spot those lineups and help fix them before anyone else notices.

• Funding: $7 million seed round led by Wing Venture Capital, with MVP, Outliers and individual security veterans participating.
• Core product: a real-time Security Context Graph linking identities, assets, applications, vulnerabilities and business systems.
• Integrations: more than 70 security and IT tools, including Okta, GitHub, Workday, Microsoft Defender and Entra, AWS and Google Cloud.
• Open access: external AI agents can query the graph through the Model Context Protocol (MCP).

What Is a Security Context Graph?

Most companies already own plenty of security tools. The problem is that each one sees a slice of the picture. Northstar ingests data from those existing tools and connects it into a single graph that updates in real time, showing how people, devices, apps and business systems relate to each other.

That connection is what exposes toxic combinations. SecurityWeek and SiliconANGLE both describe Northstar's example: a laptop missing endpoint protection might look like a minor ticket. In the graph, it turns out to belong to a privileged user who can reach the payroll system. SiliconANGLE notes that confirming that risk manually could mean checking six separate tools. The graph answers it in one step.

How Do Northstar's AI Agents Fix Security Risks?

Analysts can ask questions in plain language, such as which unmanaged devices can reach finance systems. Once a risky combination is confirmed, Northstar's agents can start remediation: installing endpoint protection, notifying the user or limiting their access. Importantly, people can supervise those actions where appropriate, which keeps humans in control of changes that affect real systems. That balance between speed and oversight is exactly what defenders need as AI agents take on more operational work.

Why Does MCP Support Matter for Security Teams?

Because Northstar exposes its graph over the Model Context Protocol, customers are not locked into one assistant. They can point their own AI agents, built on Claude or other models, at the same context and design custom workflows. It is a smart architectural choice. Context is the scarce resource in agentic security, and making it available through an open standard lets teams build on it rather than duplicate it. It fits the broader push to give security teams better visibility into agent behavior, from Cisco's Project Quarry detection rules to Shufti's Know Your Agent verification.

Who Is Behind Northstar Security?

Co-founder and CEO Amer Deeba summed up the problem as security teams "drowning in data but starving for context," according to SiliconANGLE. The round also drew individual investment from cybersecurity veterans Gerhard Eschelbeck, Renaud Deraison and Justin Somaini, a vote of confidence from people who have built and run security programs at scale.

The takeaway for defenders is simple: connecting the data you already have can reveal risks that no single tool flags, and AI agents can then close them faster. For more defensive tools and practical security news, follow our AI security coverage.

Sources: SecurityWeek — November 1, 2026; SiliconANGLE — October 28, 2026.