Threat Library

Agent-to-agent threats first — conventional coverage one click away.

Browse by hub: AI agent threats · Conventional watchlist · OWASP Agentic Top 10

Showing 20 of 1467 threats

outageavailabilityanthropicclaudeagent-relevant

Anthropic's Claude AI service experienced a major outage affecting multiple services, with users reporting login failures and degraded performance. This is an availability incident rather than a confirmed malicious cyberattack, though the root cause has not been publicly disclosed.

data-breachcryptocurrencyhardware-walletPII-exposuredark-web-sale

Cryptocurrency hardware wallet vendor SafePal disclosed a data breach affecting approximately 39,798 customers after an application flaw was exploited to exfiltrate customer order information. A threat actor is now advertising the stolen data for sale on underground forums, raising risk of targeted phishing and social engineering against affected customers.

ICSCISA-advisorySiemensfile-parsingmemory-corruptionout-of-bounds-readout-of-bounds-writeuse-after-freecritical-manufacturinglocal-code-execution

Siemens Solid Edge SE2025 and SE2026 contain seven high-severity memory corruption vulnerabilities (CVSS 7.8) triggered when parsing specially crafted PAR, PSM, or DFT files, which could allow an attacker to crash the application or achieve arbitrary code execution in the context of the current process. Exploitation requires a user to open a malicious file, making this a local-vector, user-interaction-required threat rather than a remotely exploitable one. Siemens has released patched versions (SE2025 V225.0.15+ and SE2026 V226.0.7+) and organizations should update promptly.

siyuanpdf-annotationrceelectronnode-jsstored-xssagent-relevantknowledge-managementrag-pipeline

SiYuan, an open-source Electron-based note-taking and knowledge management application, fails to sanitize annotation fields written via the setFileAnnotation endpoint prior to v3.7.4. This allows an attacker to embed malicious markup that executes as script with full Node.js privileges when a victim opens an annotated PDF, enabling complete host compromise.

wordpressplugin-vulnerabilityfile-deletionrceunauthenticatedcms-security

The Link Library plugin for WordPress (versions up to 7.9.4) contains an arbitrary file deletion vulnerability caused by insufficient path validation in the ll_delete_link_fields function. When the 'Delete local file on link deletion' option is enabled, unauthenticated attackers can submit malicious links that, once deleted by an administrator during routine moderation, trigger deletion of critical files such as wp-config.php, potentially leading to full remote code execution.

wordpressplugin-vulnerabilityprivilege-escalationauthorization-bypasscmsweb-application-security

The Pods – Custom Content Types and Fields WordPress plugin (versions up to 3.3.9) contains a critical authorization bypass flaw that allows unauthenticated attackers to escalate privileges to Administrator or reset any user's password, including the site owner's. This enables complete site takeover and has been assigned a CVSS score of 9.8.

wordpressplugin-vulnerabilityaccount-takeoverunauthenticatedprivilege-escalationweb-application

The TrueBooker WordPress plugin (versions up to 1.2.6) contains a critical account takeover vulnerability due to an insecure AJAX handler that allows unauthenticated attackers to change any user's email address, including administrators. Attackers can chain this with WordPress's native password reset flow to fully hijack accounts, including full site administrator access.

wordpressplugin-vulnerabilityauthentication-bypasstype-confusionprivilege-escalationcmsweb-application

The User Profile Builder plugin for WordPress (versions up to 3.16.4) contains a critical authentication bypass vulnerability caused by improper error handling during user registration. An unauthenticated attacker can exploit a type confusion flaw to obtain an autologin nonce bound to user ID 1, effectively logging in as the site's Administrator and achieving full site takeover.

benchmarkreasoning-effortperformancenot-a-vulnerabilitylocal-llmqwenSurface: ModelPropagation: None

This raw data is a blog post by Simon Willison reviewing the Qwen 3.8 27B model, noting that its default 'xhigh' reasoning effort setting causes excessive token usage and long generation times on consumer hardware. This is a usability/performance observation about model configuration defaults, not a security vulnerability, prompt injection, or agent-to-agent threat.

blog-posttoolingsvgmarkdownno-security-issueSurface: Human InterfacePropagation: None

This is a blog post by Simon Willison describing a personal web tool that renders SVG embedded in Markdown, adds export tabs (PNG/JPEG/MP4), and uses ffmpeg.wasm in the browser to compile animation frames into video. There is no mention of AI agents, prompt injection, tool poisoning, or any inter-agent communication in this content, and no security vulnerability is described.

microsoft-defenderpatch-bypassprivilege-escalationwindowsSYSTEM-accesszero-dayproof-of-conceptagent-relevant

A researcher known as Chaotic Eclipse released a public proof-of-concept called ShieldBreak that bypasses Microsoft's patch for CVE-2026-50656 (RoguePlanet), a Windows Defender vulnerability. The PoC reportedly grants SYSTEM-level access, meaning organizations that applied the original patch may still be exposed to full local privilege escalation.

SAPRCEactive-exploitationenterprise-softwarecommerce-platform

A maximum-severity remote code execution vulnerability in SAP Commerce Cloud, patched only three days prior, is already being actively exploited in the wild according to threat intelligence firm Defused. Organizations running unpatched instances face immediate risk of full system compromise, making rapid patching or mitigation critical.

google-workspaceoauthtoken-theftcloud-securityidentity-securitysaas-securityagent-relevant

This report highlights that attacks against Google Workspace increasingly bypass traditional phishing defenses by exploiting stolen OAuth tokens to gain access to Gmail, Drive, and connected third-party applications. Material Security emphasizes that organizations must defend the entire Workspace attack chain, not just the initial login, since attackers can pivot through connected integrations and persistent tokens. This represents a shift toward identity- and token-centric attack paths rather than credential phishing alone.

botnetmirai-variantlinux-malwarerouter-compromisesocks5-proxyiot-security

Evooo1Bot is a newly identified Mirai-based modular Linux botnet targeting internet-facing gateway devices and routers. Once compromised, infected devices are converted into SOCKS5 traffic relay nodes, likely to support proxy-for-hire services or to anonymize other malicious traffic.

ICSOTenergy-sectorhard-coded-credentialsmissing-authenticationrecoverable-passwordsVNC-exposureCISA-advisory

ANDRITZ HIPASE-250 and 250 SCALA industrial control system products (versions <=7.20) contain four vulnerabilities including recoverable password storage, missing authentication on data/config endpoints, an unauthenticated logging manipulation endpoint, and a hard-coded VNC credential used across engineering workstation deployments. Successful exploitation could allow an attacker to read sensitive process data, access engineering workstations, suppress audit logs, or recover stored credentials. These are primarily energy-sector ICS/OT vulnerabilities with no reported public exploitation to date.

ICSSCADASiemensPLCcryptographic-weaknesshardcoded-keypassword-hashingCWE-321CWE-759industrial-control-systems

Siemens LOGO! Soft Comfort versions prior to V9 contain two vulnerabilities affecting project-file encryption and password protection: a hardcoded AES master key and unsalted SHA-256 password hashes. A local attacker could exploit these flaws to decrypt project files, bypass or remove passwords, and perform efficient offline brute-force attacks, potentially gaining unauthorized access to sensitive PLC project logic and configurations.

ICSmedical-devicehardcoded-credentialsbluetoothCWE-798healthcareIoT

Flow Neuroscience FL-100 (and rebranded Halo Neuroscience FL-100) tDCS devices contain an undocumented hard-coded credential shared across all units, allowing any attacker within Bluetooth range to bypass authentication. Exploitation could let an attacker arbitrarily manipulate brain stimulation parameters and override built-in safety limits, posing direct physical harm risk to patients.

wordpressauthentication-bypassaccount-takeoverplugin-vulnerabilitycryptographic-failureunauthenticated-rce-adjacentcms

The User Session Synchronizer plugin for WordPress (versions up to 1.4.0) contains a critical authentication bypass vulnerability that allows unauthenticated attackers to impersonate any user, including administrators. The flaw stems from unvalidated request parameters and a cryptographic fallback that renders the encryption predictable when an unregistered session key is referenced. Full site takeover is possible with no prior knowledge of secrets, making this an urgent patch priority for any WordPress site running the plugin.

wordpressauthentication-bypassplugin-vulnerabilityprivilege-escalationunauthenticated-rce-adjacentagent-relevant

The 6Storage Rentals WordPress plugin (versions up to 2.27.0) contains a critical authentication bypass vulnerability allowing unauthenticated attackers to log in as any existing WordPress user, including administrators, simply by supplying that user's email address. This flaw stems from an insecure AJAX handler exposed to unauthenticated users that lacks nonce, capability, or ownership checks before establishing a full authenticated session.

wordpressplugin-vulnerabilityarbitrary-file-deletionunauthenticated-rcecontact-form-7web-application-security

The RapiSafe – Secure Multi File Upload plugin for Contact Form 7 (versions up to 1.0.4) contains an unauthenticated arbitrary file deletion vulnerability in its AJAX upload removal handler. Attackers can exploit exposed nonces to delete critical files such as wp-config.php, potentially triggering a reinstallation flow that leads to full remote code execution and site takeover. Given the plugin's popularity and the ease of exploitation (no authentication required), this poses a severe risk to any WordPress site running the affected component.