Netcrook
07 August 2026
When Hiring AI Starts Listening to Applicants, Security Gets a New Front Door
Candidate-facing AI turns resumes and interview answers into machine-readable input, which means hiring platforms can become decision systems that need the same defensive scrutiny as any other exposed enterprise app.
DHS Remark Draws a Line Around the AI Safety Panic
A blunt comment about a "Terminator factory" captures a quieter policy message: AI companies are being told that recent incidents matter, but that technical discipline may matter more than a fresh wave of lawmaking.
Cheap First-Year Security Can Hide an Expensive Lesson for Small Businesses
A Norton Small Business promotion puts endpoint protection, backup, password management, and dark web alerts into one bundle, but the real story is how SMB security is now sold as a mixed stack of defense, recovery, and convenience.
Inside the Airline Cutover That Turned a Merger Into a Systems Test
Alaska Airlines’ move to a shared passenger service system with Hawaiian Airlines shows how a merger can become a live stress test for booking continuity, customer experience, and operational discipline.
The New Battleground in Circular Economy Is the Record, Not the Trash
As circular economy models mature, product sustainability is increasingly judged by the quality of the data that follows each item through the supply chain.
When a PR Becomes a Trust Boundary: The Claude Code MCP Risk Developers Should Not Ignore
A reported path from repository changes to local code execution shows how AI coding tools can turn configuration into an attack surface if teams treat project files as harmless text.
AI-Assisted HTTP Terminator Spots New Desync Paths and an Apache Traffic Server Zero-Day
PortSwigger says its AI-assisted research system explored 30,000 candidate desync vectors, produced new HTTP desynchronization techniques, and helped surface a separate zero-day in Apache Traffic Server.
The B2B Sales Graph Is Getting Sharper - and So Is the Attack Surface
As prospecting moves from fairs and cold calls to platform-driven outreach, the same visibility that helps sales teams find decision-makers can also help attackers map targets and build convincing pretexts.
When a Trusted Repo Becomes a Command Channel
A newly disclosed Claude Code issue spotlights a dangerous edge case in AI-assisted development: repository-scoped trust may be able to shape what runs on a developer workstation.
Mailbox Theft for Profit: How AiTM Phishing Is Hunting Microsoft 365 Finance Teams
A phishing wave aimed at Microsoft 365 is using adversary-in-the-middle tricks and residential proxy camouflage to reach the mailboxes that sit closest to payroll and finance workflows.
Chrome’s Latest Patch Wave Hides a Bigger Story: 41 Bugs, 6 Critical Risks
Google’s newest Stable build for Windows, macOS, and Linux is a reminder that browsers are high-value attack surfaces, not just everyday apps.
When Banking Apps Start Thinking: The Quiet Expansion of GenAI Into the Control Layer
Embedded AI is moving beyond chat interfaces and into banking workflows, but the real challenge is not speed - it is preserving traceability, approvals, and audit-grade control.
When a Task App Becomes a Mirror for Your Day
Hackaday published a post about FlipBuddy with a teaser centered on time, tasks, and the appeal of tracking every task, which makes the product feel less like a gimmick and more like a lens on how much routine software can learn.
NatJack Turns the Router’s Memory Into an Attack Surface
A newly disclosed NAT-focused attack class puts connection state, not just packets, at the center of a quiet but serious network integrity problem.
UNC6671’s Name Changes Point to a Moving Target in Vishing Extortion
A group tracked as UNC6671 is described as shifting from BlackFile to Redact, Pink, Helix, and Falcon, a pattern that may complicate defender tracking.
Old Kernel, Sharp Edge: How a Quiet SCTP Bug Became a Host Risk
A long-lived Linux memory bug in SCTP shows how a narrow networking path can matter far beyond the container it touches.
When AI Speed Comes Chained to the Chip
AMD’s move toward model-specific inference silicon could sharpen the economics of enterprise AI, but it also turns flexibility, patching, and lifecycle management into hardware problems.
When Isolation Becomes a Lifeline: The New Logic of OT Defense
BlackBerry's analysis puts a hard truth at the center of critical infrastructure security: disconnecting industrial systems is not enough unless operators can still communicate, coordinate, and recover.
Windows Hello’s Quiet Shortcut: How a Live Session Can Bleed Into Cloud Identity
A newly disclosed Windows Hello for Business technique suggests that a user session, not a PIN, may be the real prize for attackers hunting Entra ID access.
When AI Starts Acting, Governance Becomes the Real Security Control
AI agents are moving from conversation to execution, and that shift turns autonomy, supervision, traceability, and accountability into pre-deployment security requirements, not afterthoughts.