All articles

A new report examines how internal coding agents could escape review in OpenAI's earlier monitoring system.

New research describes attackers delegating decisions and commands to AI agents.

Microsoft's 19 October update focuses on controlling what company agents can do as their tasks and risk change.

Lumen researchers describe attacks on vulnerable AI infrastructure. The target is the server running the service, not the model's answers.

A malicious npm release could run during installation. The case exposes a security boundary outside an AI agent's sandbox.

Open-source testing software and AI models appear in a real intrusion campaign.

Defense, Red Team and Specialized Access cover different security tasks. Verification and data-handling requirements remain part of the program.

A safety decision about one question does not necessarily account for the workflow its answer becomes part of.

Hacking capability and permission to reach other systems are different questions. A new research update explains how containment tests examine the second.

ISO 27001 is a voluntary standard you can be certified against. NIS2 is EU law with mandatory duties, reporting clocks and fines. They overlap in content, not in consequence.

Delinea's new survey highlights the difference between an assistant's approved task and its actual access to company data.

New research examines what agents did when ordinary data retrieval failed.

A lab test shows why an email assistant's model and its surrounding automation must be understood together. The results differed between Gmail's sidebar and a custom workflow.

Zimperium describes Android malware that asks an AI assistant where to tap and scroll. The model helps navigate a phone the malware has already compromised.

Plugin4Shell exposed a gap between approving an AI coding plugin and checking the code delivered. A published Codex fix explains the missing verification.

Hacktron describes Claude-assisted research linking a forum image-upload flaw to OpenAI account access. Here is how the chain worked.

CrowdStrike's new PhantomRaven report connects likely AI-written malware, npm packages and bug bounty activity. Here is what the findings establish.

Falcon Data Security aims to recognise confidential information in ordinary text. Here is how classification on the PC works.

Agentic IdP gives AI agents verifiable identities and connects their access to the person or system behind them. Here is how it relates to Guardian and Continuous Identity.

SafeMind combines offensive and defensive AI models with software that runs security tasks. Here is how Red Tempest, Blue Solano and the testing loop fit together.

All 14 capabilities in CrowdStrike's AIDR-to-Guardian comparison explained, from shadow AI discovery to agent controls, investigations and cost analytics.

Visual prompt-injection research exposes a practical boundary: an agent reading customer images should not be able to change its own operating instructions.

CrowdStrike's Fal.Con 2026 announcements connect AI activity, agent identities and AI-powered defence. Here is what Guardian, SafeMind and Agentic IdP each add.

CrowdStrike is extending AI security into agent activity. Here is what Guardian adds and what we would check before a rollout.

RubyGems removed more than 500 malicious packages in May. New research attributes the campaign to AI agents and highlights risks in automated documentation builds.

The September report describes AI-assisted intrusions with varying human involvement. The response should focus on access and observable behaviour.

An agent-assisted campaign moved from planning to credential theft in hours. Check how quickly your cloud response can contain unauthorised activity.

Check Point demonstrated a cross-account channel combined with prompt injection. The channel was closed, but connector permissions still deserve review.

Researchers reconstructed about 18,000 agent posts on DSEwiki. The findings raise practical questions about writable websites and agent network controls.

ESET describes malware comments intended to derail AI-assisted analysis. Your pipeline needs an explicit path for incomplete or refused results.

Researchers describe an Aurora operator using Cursor during attacks. Focus on the actions and access, rather than treating the tool's name as a verdict.

The industry calls for faster cyber defense. For business leaders, the next step is to turn that commitment into funded, accountable work.

OpenAI's investigation describes unauthorised agent communication and uneven safeguards. Shared storage and test environments need explicit security boundaries.

Aikido's lab recreation shows an agent bypassing a booking restriction. The result is a reminder to enforce permissions on the server.

Only 12 of 405 samples appeared in Unit 42's endpoint dataset. That is useful evidence, but not a measure of every AI-assisted attack.

An NSA-led advisory describes reconnaissance against Siemens PLCs. Review exposure and remote access with the people responsible for the process.

Adversa demonstrated a Grok attack using encrypted webpage content. Transforming external data must not turn it into trusted instructions.

OpenAI reported a training pause and stricter controls after signs of critical cyber capability. The announcement also makes the cost of agent oversight visible.

A preprint shows instructions propagating between agents. Its warning-prompt result is promising, but should not replace access controls.

OpenAI's cyber model responds to more advanced security requests. Its 95 percent completion figure measures refusals, not a universal exploit success rate.

Dream's research describes a four-day intrusion using AI agents. The findings show why weak authentication and exposed APIs still matter.

Researchers recovered sensitive data from published reasoning blocks. Treat agent traces as potentially confidential before sharing them.

JFrog describes malicious hooks and editor tasks in the August campaign. Review executable project settings as well as package dependencies.

Tenet's research shows how attacker-controlled log content can steer an agent. Reading an event should not give that event authority to change systems.

The ISO 27001 timeline depends on your starting point, the controls you need to implement and the certification body's audit schedule.

AI can help draft policies and organise evidence. People still need to make risk decisions, implement controls and verify that the records describe reality.

Build shared security processes, but track legal duties and certification separately. Applicable legal deadlines take priority over commercial preferences.

Compare ISO 27001 proposals on implementation, internal time, tools, audit fees and ongoing maintenance. Scope matters more than an unsupported headline price.

Clarify the certificate scope, deadline and acceptable evidence before promising a date. Then assess the gaps and agree a credible plan.

Secured by FM CyberSecurity combines security operations, advice and ISO 27001 preparation. Here is what to clarify when buying that work as a subscription.

AI agents breached Hugging Face during a security evaluation. What does it mean for a model to find weaknesses and act on its own?

Palo Alto Networks introduced Idira in May 2026. Here is how to navigate the new name, existing CyberArk documentation and licence questions.

Start with account ownership and high-impact access. Plan credential rotation, emergency access and ongoing operation before expanding the rollout.

The next-gen label does not define one architecture or price model. Evaluate source coverage, detection, search, retention and the work behind the service.

DORA distinguishes general resilience testing from designated TLPT. Choose methods against risk, document coverage and verify remediation.

The EPM control panel gives Windows users access to administrative tasks enabled by IT policy. It is separate from both Windows Control Panel and the admin console.

Tenable's vulnerability service is available separately and within the wider platform. Compare the extra coverage and operating work before choosing a package.

Software can organise registers, evidence and technical findings. Choose it around the decisions, data and responsibilities in your DORA programme.

A response plan should identify who can act, how evidence is preserved and which notifications apply. The first decisions depend on the incident, not a fixed script.

Aikido includes cloud posture checks. Assess its configuration coverage, permissions and reporting against the cloud risks you need to manage.

The ICT agreement register, planned-contract notices, incident reports and internal audit have different purposes. Assign owners and use the current submission rules.

Three incidents reached real systems from an evaluation setup with unintended internet access. A prompt describing a sandbox cannot enforce its boundaries.

Ransomware can affect availability, confidentiality and recovery. Norwegian businesses need to assess operational harm and personal-data obligations together.

A useful scan needs authorised scope, reliable collection and follow-up. Understand the difference between missing findings and missing coverage.

Continuity plans should follow the services the firm must sustain. Map dependencies, set recovery objectives and exercise the decisions as well as the technology.

EPM lets users work without permanent local administrator rights. Successful deployment needs precise rules, an approval process and support for legitimate tasks.

Agents left the intended boundaries of an OpenAI evaluation and compromised Hugging Face. The incident puts shared infrastructure and agent permissions in focus.

CRA manufacturer reporting began on 11 September 2026. The main product requirements follow in December 2027. Check product scope and your role now.

CISSP is a broad security credential with experience requirements. Verify it, then assess the person's fit for your specific systems and assignment.

Implementation and audit are different assignments. Check the credential issuer, level and current status, then assess the consultant against the work you need.

PCI DSS, DORA, NIS2, ISO 27001 and GDPR have different testing requirements. Match the method, scope and tester qualifications to the actual obligation.

A pentest report needs context: scope, release, findings, remediation and retesting. Connect those records to the risks and controls in your ISMS.

Use a Tenable trial to answer a defined security question. Confirm eligibility, modules, access and assessment work before the evaluation begins.

An identity assessment can help prioritise account risks. Agree trial availability, access, coverage and the written findings before connecting your directory.

A PAM project needs decisions about access, dependencies and recovery. Here is what to prepare for a useful architecture conversation.

A useful first assessment establishes coverage, validates findings and assigns remediation. Agree the scope and scan conditions before collecting results.

Scanning needs a routine for coverage, triage, remediation and verification. Here is a practical cadence to adapt to your systems and responsibilities.

Vulnerability management addresses weaknesses throughout their lifecycle. Exposure management adds a broader view of access paths, identities and business impact.

Johan Vorgaard, Kenny Le, Maximilian Sharoyan and Fredrik Standahl represented FM CyberSecurity at Arrow ECS Norway's Summer Cloud Festival in Oslo.

Choose by assessment scope, collaboration and the work you can operate. The scanner, vulnerability service and broader exposure platform serve different needs.

Tenable visited FM CyberSecurity's Oslo office in May. Guy March also took a turn in the racing simulator, recording 1:36.052 at Silverstone.

Nessus assesses vulnerabilities and configuration on defined targets. The wider Tenable products add management and coverage for other security areas.

Fredrik Standahl's Digi.no op-ed argues for approved AI tools and clear rules for handling business information.

In E24, Fredrik Standahl discusses the gap between employees' AI use and the controls their employers have put in place.

Start with the AI tools people use, agree what data they may handle, and test the controls. A practical approach using CrowdStrike.

Unapproved AI use can expose business data and give agents excessive access. Start by understanding the work people are trying to do.

A security operations centre combines detection, investigation and response. In-house, managed and shared models should be compared against the same responsibilities.

A SIEM collects and analyses security events across systems. Start with the detection and investigation gaps, then decide how the service will be operated.

Modern antivirus does more than match known files. EDR adds investigation and response capabilities. Check the coverage and who acts on the alerts.

A static website can keep publishing separate from visitor traffic. Git, a build pipeline and a controlled publishing service each have a distinct role.

Falcon brings endpoint, identity and other security capabilities into a shared platform. Modules, integrations and the managed service have different roles.

Clear security records help answer customer and investor questions. Build evidence around the controls you operate, with owners and an honest view of gaps.

Let actual customer requirements define the SOC 2 scope, report type and timing. Reuse existing security evidence, but assess the gaps before budgeting.

SOC 2 Type 2 reports on controls over a defined period. Read the system scope, opinion, testing results and customer responsibilities, not just the report title.

The Digital Security Act has applied since October 2025. Check the service definitions, exceptions and reporting rules for your category of provider.

ISO 27001 certification gives buyers independent evidence about a defined information security management system. Its scope and the tender wording matter.

Define scope, assess risk, implement controls and gather evidence. A practical checklist for preparing a Norwegian SMB for ISO 27001 assessment.

Confirm the applicable regime, map critical functions and connect the evidence. Updated for Norway's September 2026 extension of DORA-based rules.

A NIS2 preparation checklist for management: scope, accountable owners, risk measures, supplier oversight and a reporting process matched to applicable law.

NIS2 scope depends on service, size, jurisdiction and exceptions. Distinguish direct legal duties from customer requirements and Norway's existing NIS1-based law.

FM's pentest offering uses Aikido, with FM reporting. The aim is repeatable testing tied to application changes, clear scope and follow-up on findings.

Tenable provides assessment and exposure analysis. The service around it must connect findings to asset owners, business priorities and verified remediation.

Endpoint and identity signals need an agreed response process. Falcon provides the platform; the service agreement defines who monitors, contains and escalates.

Turn NIS2 preparation into a practical work plan: confirm applicable rules, assess controls, exercise reporting and track corrective actions.

AI can help investigate security alerts. The useful test is whether it improves decisions and response in your own operation.

Fredrik Standahl's Shifter commentary asks startup teams to make security checks part of releasing AI-generated code.

VG Dine Penger interviewed Fredrik Standahl about starting a cybersecurity company and recruiting in the sector.

Norwegian Cybersecurity Cluster interviewed Fredrik Standahl and Maximilian Stiegler Sharoyan about founding FM CyberSecurity.