Inherited Access, AI Permissions, Supply Chain Attacks & Edge Exposure

This week’s conversation focuses on a shift that’s getting harder to ignore: initial access is no longer a clean moment in time. It’s showing up as a side effect of systems that already belong in the environment.

Across the headlines, attackers aren’t forcing entry so much as inheriting it. AI platforms are exposing access through default permissions. Trusted open-source packages are delivering backdoors as part of normal installs. Edge devices are getting exploited before patch cycles even start, or aging into a state where they can’t be fixed at all. At the same time, session hijacking techniques are bypassing MFA entirely by operating inside legitimate logins, and malware is moving onto infrastructure that can’t be taken down.

The common thread here is control. Access is being granted, reused, and persisted through systems defenders rely on every day, which makes it harder to define where initial access actually begins.

Key Takeaways:

The FCC Just Banned All New Foreign-Made Routers. Everything You Need to Know to Keep Your Network Safe, CNET

https://www.cnet.com/home/internet/fcc-just-banned-all-new-foreign-made-routers-everything-you-need-to-know-to-keep-your-network-safe/

• What Matters: The ban targets new devices, but the real risk is already sitting on networks today. Existing routers will age into unpatchable, internet-facing systems that continue handling traffic without a fix path. That creates a durable attack surface attackers can rely on being there. From the team’s perspective, this reinforces a model many already assume: treat edge infrastructure as untrusted. Over time, this becomes less about vulnerability management and more about managing permanent exposure.
• What’s Overhyped: The policy itself doesn’t reduce current risk. In the short term, it may increase it through rushed replacements and misconfigurations. The bigger issue is how little visibility and control most teams have over-edge devices once deployed.

Vertex AI Vulnerability Exposes Google Cloud Data and Private Artifacts, The Hacker News

https://thehackernews.com/2026/03/vertex-ai-vulnerability-exposes-google.html

• What Matters: AI agents with default permissions are exposing cloud environments through access that already exists. Attackers can extract service account credentials and move laterally without exploiting anything. This is access being inherited through trusted systems, not forced entry. Small permission gaps can quickly turn into broad environment access. The scale comes from how widely these agents are deployed with similar defaults
• What’s Overhyped: The AI framing draws attention, but the root issue is familiar. This is a permissions and identity problem, not a new class of attack.

Google Attributes Axios NPM Supply Chain Attack to North Korean Group UNC1069, The Hacker News

https://thehackernews.com/2026/04/google-attributes-axios-npm-supply.html

• What Matters: A trusted NPM package delivered a backdoor as part of normal install behavior. No exploit required, just routine dependency use. This is supply chain access embedded directly into developer workflows. The attacker invested time to blend in, making detection difficult. Once executed, the code runs with the same access as the developer environment.
• What’s Overhyped: This isn’t limited to NPM or a specific ecosystem. The real risk is how much trust is placed in dependencies that aren’t fully validated.

Citrix NetScaler products confirmed to be under exploitation, Cybersecurity Dive

https://www.cybersecuritydive.com/news/citrix-netscaler-exploitation-vulnerabilities/816097/

• What Matters: Exploitation began almost immediately after the vulnerability became public. Systems at the edge were targeted before organizations could patch. Attackers move fast, scanning and harvesting access within hours. High-value systems that prioritize uptime are especially exposed. In many cases, compromise happens before defenders even start remediation.
• What’s Overhyped: This isn’t just a problem for large or regulated industries. If a system is visible, it’s a target. Patch speed alone doesn’t close the gap.

AitM Phishing Targets TikTok Business Accounts Using Cloudflare Turnstile Evasion, The Hacker News

https://thehackernews.com/2026/03/aitm-phishing-targets-tiktok-business.html

• What Matters: Attackers are capturing session tokens in real time, bypassing MFA by operating inside legitimate logins. Once the session is valid, they don’t need to authenticate again. Initial access and persistence happen almost at the same moment. From there, accounts can be reused for malware delivery or further access.
• What’s Overhyped: MFA still has value, but it’s not a complete control. Defensive visibility drops off after login, where activity is assumed to be legitimate.

Malware Is Sleeping on the Blockchain, and It's Already Infected Dozens of Global Targets, PCMag

https://www.pcmag.com/news/malware-is-sleeping-on-the-blockchain-and-its-already-infected-dozens-of

• What Matters: Malware hosted on blockchain infrastructure can’t be easily removed. The payload is effectively permanent once deployed. This removes the option of takedown and forces a shift to containment. Defenders have to focus on blocking interaction, not removing the source. Detection depends on identifying behavior that doesn’t match normal usage.
• What’s Overhyped: The blockchain angle is new, but the tactic is not. Behavioral detection still applies, even when the infrastructure can’t be taken down.