Netcrook
25 Junio 2026
Salesforce Turns Support Bots Into Metered Agents, and Security Teams Should Pay Attention
Agentforce Help Agent is more than a chatbot launch: it ties autonomous customer service to outcome-based pricing, which raises the stakes around permissions, escalation, and abuse resistance.
OpenAI’s Daybreak Pushes AI Closer to Patch Work - and Closer to New Risk
The project is being expanded to explore whether large language models can help correct vulnerabilities at scale, a shift that could reshape remediation without removing the need for strict human control.
Duolingo’s Quiet Rollback Shows Why AI Usage Is a Dangerous KPI
When a company rewards employees for using AI, the metric can start measuring compliance instead of productivity, and that is where governance gets noisy.
The Quiet Weak Link in Water Security Is Now Getting a Federal Fix
NIST has issued guidance for water utilities that rely on remote access, spotlighting a control path that is convenient for operators but risky for critical infrastructure.
Chaos Claims a Roofing Victim, but the Evidence Trail Is Still Thin
An unverified extortion claim tied to Roof Depot shows how ransomware crews can weaponize names, directory entries, and identifiers long before defenders know whether a real intrusion happened.
Leak-Site Listing Turns a Roofing Firm Into a Ransomware Signal, Not Yet a Proven Breach
A public victim entry tied to Roof Depot shows how extortion crews can weaponize visibility long before anyone confirms whether data was stolen or systems were touched.
Interlock’s Claim Against Clearview Eye Centre Highlights a Familiar Ransomware Pressure Tactic
A public attack claim naming clearvieweyecentre.com is not proof of compromise, but it does show how quickly extortion messaging can force healthcare defenders back into triage mode.
Leak Page, Real Risk: Why a Named Clinic on a Ransom Note Matters
A victim listing tied to Interlock and a Calgary eye clinic is a reminder that the danger in ransomware is often the mix of stolen data pressure, business disruption, and uncertain facts.
Claude Tag Lands in Slack Beta for Enterprise and Team Users
Anthropic’s new Slack integration is a narrow product update on paper, but it adds another AI surface to the place where many organizations already handle their most sensitive day-to-day work.
When a Water Threat Stays Virtual: The OT Test Behind Cal Water’s Incident
A California utility faced a public disruption claim, but the key finding was narrower and more revealing: no evidence of OT activity, which keeps the case in the realm of verified cyber risk rather than confirmed physical-process interference.
GitLab Patch Wave Exposes the Real Risk in DevOps: One Upgrade Can Close Many Doors at Once
A security update for GitLab CE and EE resolves 13 flaws, including three rated high severity, and the practical lesson is simple: delayed patching can leave collaboration platforms sitting on multiple attack surfaces at the same time.
When the Intruder Logs In Looking Normal, Security Gets Harder
A webinar on account takeover points to a stubborn reality in cloud defense: once an attacker is using a valid identity, the attack can blend into ordinary business activity.
Google Brings Computer Control Into Gemini 3.5 Flash, and the Prompt Injection Problem Gets Realer
A built-in computer-use feature pushes Gemini into browser, mobile, and desktop workflows, but the security question is now how well an agent can be kept from acting on hostile instructions.
Akira’s Latest Name Drop Shows How Ransomware Pressure Starts Before Proof
A claim tied to JMS-Southeast illustrates the gap between extortion theater and verified compromise, where defenders must read the signal without mistaking it for certainty.
Leak-Site Claim Puts Industrial Sensor Maker in Akira’s Crosshairs
A named ransomware crew says it has data tied to JMS Southeast, a temperature-control supplier, turning a leak threat into a potential confidentiality problem for an industrial niche that runs on trust.
The Quiet Breach Path: How a Third-Party Token Spill Reached LastPass Data
A delegated-access compromise in a business SaaS layer shows how contact data can leak through an integration boundary even when a vendor’s core vault systems stay untouched.
Gaslight Shows Malware Can Try to Trick the Analyst’s AI
A newly identified macOS implant is notable not just for stealing data, but for embedding text meant to derail AI-assisted triage.
The Real Board Test Starts After the Slides End
A board meeting is not a finish line. For CIOs, the real security work is turning questions, concerns, and executive alignment into an ongoing risk conversation.
The Quiet Arms Race Inside Enterprise AI: Speed, Control, and the New CIO Burden
As companies push AI into everyday operations, security teams are being asked to do something difficult: open the gates fast, but keep the data, identities, and decisions inside the fence.
Old Malware, New Script Chain: Why LokiBot Still Fits Modern Windows Defense Gaps
A recent LokiBot campaign pairs obfuscated JScript with PowerShell, showing how native Windows scripting can still carry commodity credential theft past noisy perimeter controls.