Issue 22
Microsoft patches exploited SharePoint zero-day and 160 other vulnerabilities
Microsoft's patch package of 14 April fixes an actively exploited SharePoint zero-day flaw and 160 other vulnerabilities.
Who is affected and how
- security leader
- The large number of patches requires advance planning of the testing and deployment window.
- security operations
- Patching SharePoint servers takes priority because the zero-day is being exploited.
CISA adds seven actively exploited vulnerabilities to its KEV catalog
CISA added seven vulnerabilities to its Known Exploited Vulnerabilities (KEV) catalog on 13 April, based on evidence of active exploitation. The list includes a 2012 CVE affecting Microsoft products…
security leaderThe additions signal active exploitation, so exposure and patching records should be aligned with the list.security operationsThe catalog is a guide for patch prioritisation: the listed products should be checked and patched first.incident · European UnionSweden says pro-Russian hackers attempted to breach thermal power plant
Sweden said pro-Russian hackers attempted to break into the systems of a thermal power plant; the news was published on 15–16 April. Hungarian outlets reported it as a Russian cyberattack.
security operationsAt energy and district-heating operators, external access and the segmentation of industrial networks move to the forefront.public-sector decision-makerThe case concerns an attempt against critical infrastructure, which justifies a review of protection and reporting frameworks.vulnerability · globalAdobe issues emergency fix for Acrobat and Reader zero-day flaw
Adobe on 13 April patched an actively exploited zero-day flaw in Acrobat and Reader. According to Dark Reading, an attacker had used malicious PDFs to exploit the flaw for at least four months.
security operationsThe patch should be deployed urgently, because the flaw has been exploited for at least four months.citizenPDF readers need to be updated, and opening suspicious PDFs should be avoided.regulation · United StatesNIST moves National Vulnerability Database to risk-based CVE triage amid record submissions
NIST said on April 16 that, because of record numbers of CVE submissions, the NVD is switching to a risk-based system. CVEs on the CISA KEV list and those affecting critical software get priority…
security leaderTools and processes should be reviewed for their dependence on the NVD.security operationsVulnerability management that relies on the NVD needs another source for CVEs that have not been enriched.journalist and researcher"Not Scheduled" CVEs do not disappear, but NIST will not enrich them automatically.Hungary
More key stories of the week
Claude Code, Gemini CLI and GitHub Copilot agents vulnerable to prompt injection via comments
According to an April 16 report, the Claude Code, Gemini CLI and GitHub Copilot agents are vulnerable to prompt injection through comments.
security leaderThe permissions and access of developer agents should be narrowed.developerCode comments and external content should be treated as untrusted input when an agent works with them.vulnerability · globalCritical Nginx UI authentication bypass flaw now actively exploited
According to a 15 April BleepingComputer article, the critical Nginx UI authentication bypass vulnerability is now being actively exploited.
security operationsOrganizations running Nginx UI must update immediately or restrict access to it.developerAuthentication on management interfaces must not be bypassable.law enforcement · globalOperation PowerOFF identifies 75,000 DDoS users, takes down 53 domains
With Europol's support, 21 countries held a coordinated action week on 13 April against the DDoS ecosystem and its users. According to the announcements, more than 75,000 users were identified and 53…
security operationsTaking down DDoS services may temporarily reduce the load, but the need for protection and load balancing does not change.journalist and researcherThe operation also targeted the user side of DDoS-for-hire services, not only the operators.research · globalAnalyses say Anthropic's Mythos model could pose serious risk to banks
An article of 13 April says that attacks supported by Anthropic's Mythos model could have severe consequences for banks. An analysis by CETaS on 14 April examined what the model means for the future…
security leaderAs AI-assisted attacks accelerate in the financial sector, risk assessments need to be reviewed.journalist and researcherThe analyses draw conclusions from the model's capabilities; they are not measured incidents.law enforcement · United StatesTwo US nationals sentenced for aiding North Korean remote IT worker fraud scheme
The US Department of Justice said on April 15 that two US nationals have been sentenced for facilitating a fraudulent remote IT worker scheme for North Korea. According to the department, the scheme…
security leaderWeaknesses in internal identity verification for remote employees should be examined separately.business leaderWhen hiring remote workers, identity verification and background checks are critical, because the scheme operated through company payrolls.incident · globalHackers exploit Marimo flaw to deploy NKAbuse malware from Hugging Face
According to an April 16 report, hackers are exploiting a vulnerability in Marimo to deploy NKAbuse malware from Hugging Face.
security operationsTraffic to AI development environments and external model sources should be monitored.developerMarimo users need to update, and components downloaded from Hugging Face should be checked.law enforcement · globalAuthorities dismantle W3LL phishing network
According to April 13 reports, authorities have dismantled the W3LL phishing network. The reports say AitM techniques allowed login sessions to be intercepted in real time.
security operationsAitM phishing captures both the password and the session, so phishing-resistant authentication is needed.business leaderTaking down the network does not remove the threat from AitM techniques.tool · globalOpenAI expands Trusted Access for Cyber and offers GPT-5.4-Cyber to vetted defenders
OpenAI said on 14–15 April that it is expanding its Trusted Access for Cyber program and making the GPT-5.4-Cyber model available to vetted defenders.
security leaderAs AI capabilities grow, defenders may also need access programs.security operationsThe model, built for defense, is available through vetted access.incident · United StatesMcGraw Hill says Salesforce misconfiguration led to data leak
Education company McGraw Hill said on April 15 that a Salesforce misconfiguration led to a data leak.
security operationsAccess and sharing settings on SaaS platforms need to be checked regularly.business leaderA cloud provider configuration error can also cause a data breach.incident · globalBooking.com warns customers that a hack exposed their data
The Guardian reported on 13 April, and a Hungarian Booking.com article on 14 April, that a hack exposed customer data. Booking.com is notifying the affected customers.
citizenCustomers who booked recently should watch for notifications from the provider and be wary of messages that misuse their data.research · globalData exfiltrated despite Microsoft patch for Copilot Studio prompt injection
According to a VentureBeat article of April 15, Microsoft patched a Copilot Studio prompt injection, but data was exfiltrated anyway.
security leaderFor AI agents, having a patch in place does not guarantee that data exfiltration has stopped.developerAgents' data outputs and integrations should be verified independently.report · globalMicrosoft warns that AI is now powering cyberattacks
Microsoft warned on April 13 that AI is now powering cyberattacks.
security leaderAccording to the vendor's warning, AI-assisted attacks require dedicated preparation.journalist and researcherThe summary gives no detailed data; the claim is Microsoft's warning.In brief
Weekly email
One email a week with the main stories from the Agent Age editions. You confirm by email and can unsubscribe in one click.
Almost done: we sent you an email. Open it and press the confirm link. If it does not arrive within a few minutes, check your spam folder.
Please enter a valid email address and tick the consent box.
Something went wrong. Please try again later.