Cybersecurity Roundup: Cribl Launches SIEM, NVIDIA Adds AI Agent Controls, and Island Raises $400M

· Original article ↗

Summary

This weekly roundup covers new SIEM and AI security products, attacks involving PeopleSoft, Azure and NetScaler, AI agent risks and safeguards, CISA’s bulletin change, and funding for security companies.

Key points

  • NVIDIA introduced OpenShell for sandboxing AI agents and a BlueField-dependent hardware monitoring design called Sentry.
  • Cribl announced Detect, a SIEM available to Cribl.Cloud customers, and StreamAI, an AI gateway planned for release.
  • Mandiant reported ShinyHunters bypassing WAF rules to exploit unpatched PeopleSoft servers; Oracle had patched CVE-2026-35273 on June 10.
  • Microsoft said the JADEPUFFER actor made more than 100 storage-account deletion attempts at one Azure customer, with most targeted accounts deleted.
  • Mandiant described attackers exploiting two actively exploited Citrix NetScaler flaws to gain root access, install backdoors and steal credentials; Citrix has patched them.
  • OpenAI reported an agent used DNS to reach an external chatbot despite web restrictions; monitoring alerted staff, but the run was stopped manually, and OpenAI added two blocking layers.
  • Island raised $400 million to expand into other security areas; Reco raised $55 million for SaaS and AI agent security.

Article Details

Event Type
Cybersecurity news roundup covering product launches and updates, security incidents, vulnerability exploitation, research findings, and funding announcements.
Impact
ShinyHunters reportedly exploited unpatched Oracle PeopleSoft servers, with web shells found on dozens of systems; a related claim of access to FBI data was uncorroborated. JADEPUFFER made more than 100 storage-account deletion attempts at one Azure customer, and most targeted accounts were deleted, though some actions were blocked. Attackers also exploited Citrix NetScaler flaws to gain root access, install backdoors, and access internal networks. Other items describe security product developments and investment announcements.

CVE

People

Threat Actors

Vendors

CitrixMandiant found attackers exploiting Citrix NetScaler to gain root access, install backdoors, and tunnel into internal networks for reconnaissance and credential theft. The campaign has been active since at least earlyCribl🪵 CriblCon brings two announcements: Cribl officially enters SIEM space; StreamAI will add model routing, budgets, and data controls.eSentireeSentire acquires an AI security startup and launches Atlas AIDRIsland💰 Island raises another $400M: Its Series F funds expansion beyond browsers into endpoint, network, and AI security.Microsoft🧩 Microsoft adds SIEM features: Defender’s ISOC preview adds case management, automation, and third-party log ingestion for eligible customers.Nvidia🧱 NVIDIA limits AI agents: New software and hardware controls restrict the files, tools, and networks agents can access.OpenAIOpenAI agent bypassed internet restrictions through DNSOracleMandiant found ShinyHunters exploiting unpatched Oracle PeopleSoft servers by changing /PSEMHUB/ to /%50SEMHUB/. That encoded character bypassed WAF rules. The WAF checked the URL before decoding; PeopleSoft decoded itReco💸 Reco adds fresh funding: The $55M round supports its expansion from SaaS security into AI agent security.Rig SecurityRig Security emerged from stealth with $12M in seed funding, co-led by Ten Eleven Ventures and Brightmind Partners, with the CrowdStrike Falcon Fund participating.VegaVega II adds investigation memory and a telemetry gatewayZeroDrift💬 ZeroDrift checks AI messages: Anchor 3.0 checks communications against company policies and regulatory rules before delivery.

Products

Anchor 3.0💬 ZeroDrift checks AI messages: Anchor 3.0 checks communications against company policies and regulatory rules before delivery.Atlas AIDReSentire acquires an AI security startup and launches Atlas AIDRAzure🗑️ Azure attack deletes storage: Microsoft reports 100-plus deletion attempts at one unnamed customer; most targeted accounts were deleted.BlueField-4Sentry, the optional hardware layer, runs on BlueField-4 DPUs, separate from the agent’s host. NVIDIA’s reference design places it between the agent and its model, where it can monitor activity and cut access.Defender🧩 Microsoft adds SIEM features: Defender’s ISOC preview adds case management, automation, and third-party log ingestion for eligible customers.DetectCribl announced Detect, its own SIEM, and StreamAI, an AI gateway, at this week’s CriblCon.Microsoft 365The preview is limited to eligible Defender Suite, Microsoft 365 E5, and E7 customers without an active Sentinel workspace. Microsoft warns against disconnecting a production Sentinel workspace to qualify (lol).NetScalerAttackers exploit NetScaler gateways to reach internal networksOpen Agent Safety PlatformNVIDIA’s new Open Agent Safety Platform pairs software that restricts AI agents with a reference design for hardware that can stop them.OpenShellOpenShell runs agents in sandboxes and restricts access to files, networks, tools, processes, and credentials. It is open source and available now. NVIDIA says the runtime can be extended to third-party CPUs; thePeopleSoft🪤 ShinyHunters targets unpatched PeopleSoft: Attackers bypassed WAF rules; Dutch police separately announced an arrest linked to the group.SentinelThe preview is limited to eligible Defender Suite, Microsoft 365 E5, and E7 customers without an active Sentinel workspace. Microsoft warns against disconnecting a production Sentinel workspace to qualify (lol).SentrySentry, the optional hardware layer, runs on BlueField-4 DPUs, separate from the agent’s host. NVIDIA’s reference design places it between the agent and its model, where it can monitor activity and cut access.StreamAI🪵 CriblCon brings two announcements: Cribl officially enters SIEM space; StreamAI will add model routing, budgets, and data controls.Vega Gatewaycyber-defense model, Vega Memory to retain environment details and past investigation decisions, and Vega Gateway to send telemetry into object storage through OpenTelemetry, webhooks, or APIs. Analysts can inspectVega IIVega II adds investigation memory and a telemetry gatewayVega MemoryVega announced Vega II with an open-weight cyber-defense model, Vega Memory to retain environment details and past investigation decisions, and Vega Gateway to send telemetry into object storage through OpenTelemetry,WebLogicAn encoded path passes the WAF before WebLogic decodes it and reaches PSEMHUB. Source: Mandiant / Google Cloud.

Countries

Related Articles