Weekly Cybersecurity Roundup: Ransomware Sentencing, AI Agent Attacks and Critical Flaws

Summary
This roundup covers a ransomware developer’s sentencing, reported AI-agent and npm malware threats, cloud token defense guidance, and critical vulnerabilities affecting SAP, a WordPress plugin, TP-Link cameras and coding-agent plugins.
Key points
- A Zurich court sentenced a Ukrainian IT specialist to nearly 13 years for developing ransomware used in extortion attacks; the verdict is subject to appeal.
- Mandiant’s report describes AI-agent risks, including a coding assistant linked to a worm across roughly 100 repositories and an accounting agent that incurred about $50,000 in cloud costs.
- CrowdStrike attributes the PhantomRaven npm information stealer to a financially motivated actor; it targets CI/CD environment variables, and researchers found no evidence the stolen data was sold.
- NIST and CISA published guidance for protecting cloud identity tokens, covering validation, secrets management and detection.
- SAP urged emergency patching of internet-facing systems for CVE-2026-44756, an unauthenticated memory-corruption flaw for which researchers demonstrated remote code execution in lab testing.
- Attackers exploited a file-upload flaw in the WooCommerce Wholesale Lead Capture plugin to upload PHP webshells; site owners are urged to update to version 2.0.3.2 and check for suspicious files.
- Plugin4Shell enables malicious plugin updates in several AI coding agents; Anthropic and OpenAI have issued fixes, while Microsoft had not patched Copilot and Google said its deprecated Gemini CLI would not be fixed.
Article Details
- Event Type
- Cybersecurity news roundup covering vulnerability exploitation and disclosure, malware, ransomware, cybercrime prosecutions, and AI-agent security.
- Impact
- Reported impacts include unauthenticated remote code execution risk in SAP products, PHP webshell uploads to vulnerable WordPress sites, potential takeover of AI coding-agent plugins, theft of CI/CD environment data, ransomware extortion, and fraud against US victims. The article reports fixes for some vulnerabilities and products, while Microsoft had not patched Copilot and Google said the deprecated Gemini CLI would not be fixed.
MITRE ATT&CK
CVE
CVE-2026-15315TP-Link fixed both issues, tracked as CVE-2026-15315 and CVE-2026-15316, in firmware V5_1.4.6 released in August.CVE-2026-15316TP-Link fixed both issues, tracked as CVE-2026-15315 and CVE-2026-15316, in firmware V5_1.4.6 released in August.CVE-2026-44756Organizations using SAP have been warned about CVE-2026-44756, a maximum-severity flaw in its Extended Passport processing code that lets unauthenticated attackers trigger memory corruption before any login check occurs.
Threat Actors
Malware
LockergogaThe court identified him as the lead developer behind the Lockergoga, MegaCortex, and Nefilim ransomware families, though it described his role as closer to a technical consultant than the operation’s mastermind.MegaCortexThe court identified him as the lead developer behind the Lockergoga, MegaCortex, and Nefilim ransomware families, though it described his role as closer to a technical consultant than the operation’s mastermind.NefilimThe court identified him as the lead developer behind the Lockergoga, MegaCortex, and Nefilim ransomware families, though it described his role as closer to a technical consultant than the operation’s mastermind.PhantomRavenCrowdStrike has tied an npm-based information stealer called PhantomRaven to a financially motivated actor who moonlights as a bug bounty hunter.
Vendors
AnthropicAnthropic and OpenAI have shipped fixes for Claude Code and Codex, Microsoft has not yet patched Copilot, and Google says the deprecated Gemini CLI will not be fixed at all.GoogleAnthropic and OpenAI have shipped fixes for Claude Code and Codex, Microsoft has not yet patched Copilot, and Google says the deprecated Gemini CLI will not be fixed at all.MicrosoftAnthropic and OpenAI have shipped fixes for Claude Code and Codex, Microsoft has not yet patched Copilot, and Google says the deprecated Gemini CLI will not be fixed at all.OpenAIResearchers at Air’s security lab disclosed Plugin4Shell, a zero-click flaw affecting Claude Code, OpenAI Codex, GitHub Copilot, and Gemini CLI that lets an attacker controlling a plugin’s repository swap a pinned,SAPOrganizations warned of critical SAP vulnerabilityTP-LinkTP-Link patches Tapo camera flaw that skips password checks
Products
Business SuiteThe bug touches a wide range of SAP products, including S/4HANA, NetWeaver, and Business Suite.Claude CodeResearchers at Air’s security lab disclosed Plugin4Shell, a zero-click flaw affecting Claude Code, OpenAI Codex, GitHub Copilot, and Gemini CLI that lets an attacker controlling a plugin’s repository swap a pinned,Gemini CLIlab disclosed Plugin4Shell, a zero-click flaw affecting Claude Code, OpenAI Codex, GitHub Copilot, and Gemini CLI that lets an attacker controlling a plugin’s repository swap a pinned, reviewed commit for maliciousGitHub Copilotat Air’s security lab disclosed Plugin4Shell, a zero-click flaw affecting Claude Code, OpenAI Codex, GitHub Copilot, and Gemini CLI that lets an attacker controlling a plugin’s repository swap a pinned, reviewedOpenAI CodexResearchers at Air’s security lab disclosed Plugin4Shell, a zero-click flaw affecting Claude Code, OpenAI Codex, GitHub Copilot, and Gemini CLI that lets an attacker controlling a plugin’s repository swap a pinned,S/4HANAThe bug touches a wide range of SAP products, including S/4HANA, NetWeaver, and Business Suite.SAP NetWeaverThe bug touches a wide range of SAP products, including S/4HANA, NetWeaver, and Business Suite.Tapo C200OPSWAT researchers found two flaws in TP-Link’s Tapo C200 security camera, including an authentication bypass that lets an attacker on the network replay a value from the camera’s own challenge-response process to gainWooCommerce Wholesale Lead Capture for WordPressDefiant says attackers have exploited a critical file-upload flaw in the WooCommerce Wholesale Lead Capture plugin, blocking more than 100,000 exploit attempts since the bug was disclosed in February.