Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Traditional RAG creates risk. In enterprise AI, accuracy and security aren’t optional. Most vector-only databases are built for speed, but they ignore enterprise realities like security and compliance. Without context, access controls, or accurate recall, they create compliance gaps that make AI unsafe for regulated industries. At Protecto, we built GPTGuard to change that — making enterprise AI safe by preventing data leaks, enforcing privacy, and keeping compliance intact.

One can’t discuss the modern state of endpoint security without mentioning a term that has quickly become ubiquitous with security solutions: artificial intelligence (AI). With a constantly evolving threat landscape and many security challenges plaguing organizations (e.g sprawling attack surfaces, monitoring and continuity gaps, alert overload, and limited resources), it’s clear that endpoint security must evolve as well, and the most-promising advancement is AI.

On August 20, 2025, Salesloft published an advisory describing a security issue potentially affecting the Salesloft Drift integration with Salesforce. On August 26, Google Threat Intelligence Group (GTIG) provided additional details about the campaign, in which a threat actor known as UNC6395 authenticated against Salesforce customer instances using compromised OAuth tokens tied to the Salesloft Drift integration with Salesforce.

Cybercriminals frequently target Small and Medium-sized Businesses (SMBs) due to their limited security resources. As cyber attacks become more sophisticated, SMBs must proactively defend their critical systems and sensitive data by investing in the right cybersecurity tools. Some cybersecurity tools that every SMB should consider adding to their security stack include a password manager, a Privileged Access Management (PAM) solution, a secure remote access solution and a secrets manager.

Last night, our automated Aikido Intel system alerted us that potentially malicious code was detected in some packages within the @nx scope, which include packages with as many as ~6 million weekly downloads. The scope and impact of this breach are significant, as the attacker chose to publish the stolen data directly on GitHub, rather than sending it to their own servers. This means that there’s a SIGNIFICANT amount of credentials that are publicly available on GitHub.