Blog
Meta’s AI Safety System Manipulated by Space Bar Characters to Enable Prompt Injection
A bug hunter discovered a bypass in Meta's Prompt-Guard-86M model by inserting character-wise spaces between English alphabet characters, rendering the classifier ineffective in detecting harmful content.
Meta’s AI Safety System Manipulated by Space Bar Characters to Enable Prompt Injection
A bug hunter discovered a bypass in Meta's Prompt-Guard-86M model by inserting character-wise spaces between English alphabet characters, rendering the classifier ineffective in detecting harmful content.
Meta’s AI Safety System Manipulated by Space Bar Characters to Enable Prompt Injection
A bug hunter discovered a bypass in Meta's Prompt-Guard-86M model by inserting character-wise spaces between English alphabet characters, rendering the classifier ineffective in detecting harmful content.
US State Department Says UN Cybercrime Treaty Must Include Human Rights Protections
The U.S. State Department emphasized the importance of including human rights protections in the upcoming United Nations cybercrime treaty. The final round of negotiations for the treaty, which began on Monday and will conclude on August 9, 2024.
New PowerShell Backdoor Linked to Zloader Malware
The newly discovered backdoor has limited samples available on VirusTotal, making detection more difficult. It operates by collecting system information and sending it to a command and control server, awaiting further instructions.
New Mandrake Spyware Found in Google Play Store Apps After Two Years
A new iteration of a sophisticated Android spyware called Mandrake has been discovered in five applications that were available for download from the Google Play Store and remained undetected for two years. The applications attracted a total of more than 32,000...
Massive Phishing Campaign Exploiting Proofpoint’s Email Protection to Dispatch Millions of Perfectly Spoofed Emails
The campaign began in January 2024 and peaked at 14 million emails in June. The emails were designed to steal sensitive information and included authentic-looking signatures to bypass security measures.
The Power and Peril of RMM Tools
As more people work remotely, IT departments must manage devices distributed over different cities and countries relying on VPNs and remote monitoring and management (RMM) tools for system administration. However, like any new technology, RMM tools can also be used...
Cyber Threat Intelligence: Illuminating the Deep, Dark Cybercriminal Underground
Learn about critical threats that can impact your organization and the bad actors behind them from Cybersixgill’s threat experts. Each story shines a light on underground activities, the threat actors involved, and why you should care, along with what you can do to...
Cybercriminals Target Polish Businesses with Agent Tesla and Formbook Malware
Cybersecurity researchers have detailed widespread phishing campaigns targeting small and medium-sized businesses (SMBs) in Poland during May 2024 that led to the deployment of several malware families like Agent Tesla, Formbook, and Remcos RAT. Some of the other...
FREE GUIDE