EUVD-2026-40723

Insufficient policy enforcement in Bluetooth in Google Chrome prior to 150.0.7871.47 allowed a remote attacker to perform privilege escalation via a crafted HTML page. Chromium security severity: Low...

EUVD-2026-40725

Insufficient validation of untrusted input in New Tab Page in Google Chrome prior to 150.0.7871.47 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. Chromium security severity: Low...

EUVD-2026-40708

Insufficient validation of untrusted input in WebXR in Google Chrome prior to 150.0.7871.47 allowed a remote attacker who had compromised the renderer process to perform UI spoofing via a crafted HTML page. Chromium security severity: Medium...

EUVD-2026-40639

Insufficient policy enforcement in USB in Google Chrome prior to 150.0.7871.47 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. Chromium security severity: Medium...

EUVD-2026-40652

Insufficient policy enforcement in WebView in Google Chrome on Android prior to 150.0.7871.47 allowed a remote attacker to bypass navigation restrictions via a crafted HTML page. Chromium security severity: Medium...

EUVD-2026-40676

Inappropriate implementation in Paint in Google Chrome prior to 150.0.7871.47 allowed a remote attacker to perform UI spoofing via a crafted HTML page. Chromium security severity: Medium...

EUVD-2026-40521

Inappropriate implementation in XML in Google Chrome prior to 150.0.7871.47 allowed a remote attacker to potentially exploit heap corruption via a crafted HTML page. Chromium security severity: High...

EUVD-2026-40554

Inappropriate implementation in Network in Google Chrome on Android prior to 150.0.7871.47 allowed a remote attacker who had compromised the renderer process to bypass site isolation via a crafted HTML page. Chromium security severity: Medium...

EUVD-2026-40514

Inappropriate implementation in Enterprise in Google Chrome prior to 150.0.7871.47 allowed a remote attacker to obtain potentially sensitive information from process memory via a crafted HTML page. Chromium security severity: High...

EUVD-2026-40511

Uninitialized Use in Dawn in Google Chrome prior to 150.0.7871.47 allowed a remote attacker to potentially exploit heap corruption via a crafted HTML page. Chromium security severity: High...

EUVD-2026-40472

Use after free in Ozone in Google Chrome prior to 150.0.7871.47 allowed a remote attacker to execute arbitrary code via a crafted HTML page. Chromium security severity: Critical...

CVE-2026-34100: Improper Neutralization of Special Elements used in an SQL Command

Guardian language-system passes the id GET parameter directly into an unsanitized SQL query in media.php line 17: SELECT id, filename, extension, type, duration, owner, private FROM files where id = '".$GET'id'."'. An authenticated attacker can perform error-based SQL injection to extract...

PT-2026-54736

Name of the Vulnerable Software and Affected Versions Guardian language-system affected versions not specified Description An authenticated attacker can perform error-based SQL injection to extract database contents. The issue occurs because the application passes the id GET parameter directly in...

CVE-2026-34109: Improper Neutralization of Special Elements used in an OS Command

Guardian language-system passes the id GET parameter directly into a PHP exec call in speech.php line 18 without sanitization: exec"php jobs/speechaudio.php ".$loginsession." ".$GET'id'." ...". No authentication is required. An unauthenticated remote attacker can append shell metacharacters...

PT-2026-54751

Name of the Vulnerable Software and Affected Versions Guardian language-system affected versions not specified Description An unauthenticated remote attacker can execute arbitrary operating system commands on the server. This occurs because the application passes the id parameter from a GET reque...

PT-2026-55082

This module enables you to test and run AI-driven workflows interactively through a chat interface. The module doesn't sufficiently enforce permissions on certain endpoints. Attackers may be able to trigger workflow execution incurring LLM spend and tool side effects or send messages into other...

📄 Flowise CSV Agent Prompt Injection Remote Code Execution

This vulnerability allows remote attackers to execute arbitrary code on affected installations of FlowiseAI Flowise. Authentication is not required to exploit this vulnerability. The specific flaw exists within the run method of the CSVAgents class. The issue results from the lack of proper...

CVE-2026-34117: Improper Neutralization of Special Elements used in an OS Command

Guardian language-system passes the id GET parameter directly into a PHP exec call in texttosubtitles.php line 19 without sanitization: exec"php jobs/texttosubtitles.php ".$loginsession." ".$GET'id'." ...". No authentication is required. An unauthenticated remote attacker can append shell...

PT-2026-54753

Name of the Vulnerable Software and Affected Versions Guardian language-system affected versions not specified Description An unauthenticated remote attacker can execute arbitrary OS commands on the server. The issue occurs because the application passes the id GET parameter directly into a PHP...

Cognitive Firewall: A Proactive, Zero-Trust, Multi-Gate Framework for LLM Safety

Large language models LLMs can be induced to produce harmful content through multi turn strategies in which no single user message appears clearly unsafe. Existing runtime safeguards commonly evaluate prompts or responses as isolated messages, which limits their ability to recover ac-cumulated...