AK Alissa Knight On CyberWire Daily

“Our developers are doing a great job at authentication, but a really bad job at authorization. I'm authenticated, I'm allowed to be there, but I'm not authorized to take remote control of the entire fleet of satellites. Maybe I'm authorized to have read-only access to one satellite. But I demonstrated that, because of authorization vulnerabilities, I could take control of every satellite in the constellation.”

CyberWire Daily · Software Engineering & Infra · November 2026

“Our developers are doing a great job at authentication, but a really bad job at authorization. I'm authenticated, I'm allowed to be there, but I'm not authorized to take remote control of the entire fleet of satellites. Maybe I'm authorized to have read-only access to one satellite. But I demonstrated that, because of authorization vulnerabilities, I could take control of every satellite in the constellation.” — Alissa Knight, CyberWire Daily

Context

Asked what an attacker gains from the APIs that task and monitor satellites, Alissa Knight says most of the weaknesses she finds in ground control systems are authorization flaws, not login flaws.

Transcript

CyberWire Daily Around 17:11 into the episode
Speaker 2

…attack surface. It is the new battle space, so to speak, on in satellites that are orbiting our planet. There's a lot of military. Tactical satellites in what's called Leo or low Earth orbit that are orbiting around the planet that our warfighters use for imagery and other tactical capabilities.

Maria Varmazis

Yeah, okay. And during this demo, going back to sort of the API topic as well, the tasking and telemetry APIs, can you help sort of define the ground truth here on what those do relative to the ground system and what would an attacker actually get out of going after those two APIs?

Alissa Knight

Sure. So there's a number of APIs in this particular ground segment. There's 10. And there's the majority of the systemic weaknesses that I've discovered in my experience and vulnerabilities around ground control is around authorization vulnerabilities. So our developer is effectively doing a great job at authentication, but a really bad job at authorization. So I'm authenticated. I'm allowed to be there, but I'm not authorized to take remote control of the entire fleet of satellites. Maybe I'm authorized just to be able to have read-only access to just one satellite, but demonstrated the capability because of authorization vulnerabilities, being able to actually take control of every satellite in the constellation. And what's unique about offensive AI platforms like ours is you can actually redirect the entire, we call it a strike group, but the entire strike group of agents at a nearby satellite. So actually in the middle of the attack, redirected the entire strike group to a nearby satellite, which I told the AI was a Russian satellite. Sorry if there's any Russians listening right now, but we have that ability because it's not deterministic to be able to redirect those non-kinetic effects…

Speaker 2

command will ask to borrow the satellite of another command to use it because their satellite is no longer within line of sight of the antenna. So you have all these idiosyncratic nuances that are different about space vehicles and space systems like satellites that you don't run into as challenges…

Maria Varmazis

We'll take a quick break now. Be back shortly.

Speaker 2

Quick question. Who or what has access to your critical systems right now? I'll give you a hint. It's not just employees anymore. It's vendors, machines, applications, workloads, and so many AI agents. Zage Security helps organizations control every identity with zero trust policies that govern…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

More from CyberWire Daily