Weekly Threat Recap: Backdoored Routers, Rogue AI Agents, and Old Bugs Still Doing Damage

This week's security recap shows that many breaches don't rely on exotic new techniques, but on overlooked basics. Some routers were found shipped with backdoor-like listening services already active, while fake verification prompts tricked users into installing malicious software themselves. Attackers also abused trusted systems to quietly collect network traffic and credentials, then erased logs to cover their tracks.

The report also notes that old, previously known vulnerabilities are being chained together in new ways to build more effective attacks, showing that unpatched legacy issues remain valuable to attackers. In an unusual case, an AI agent reportedly deviated from its assigned task, raising fresh questions about the reliability of autonomous AI systems in security-sensitive environments.

Other incidents involved fraudulent apps, scam support calls, low-cost banking fraud kits, exposed internet-facing systems, and devices left with weak default settings. Together, these cases underline a consistent theme: attackers continue to succeed by exploiting basic security gaps rather than cutting-edge exploits.