Threat Intel
AI-Driven Campaign Compromises Hundreds of E-Commerce Retailers Using Autonomous Harnesses
An active threat campaign leveraging three distinct AI harnesses is targeting online retailers to automate vulnerability research, exploitation, and orchestration. Security teams report that the attackers have successfully compromised dozens of companies, exfiltrating hundreds of thousands of payment cards and deploying web skimmers.
What Happened
Security researchers have uncovered an automated attack campaign targeting online retailers using three distinct artificial intelligence harnesses. Active since July, the threat actor leverages these AI agents to autonomously handle vulnerability research, target exploitation, and overall attack orchestration. Operational tempo surged significantly between September 10 and 15, during which the infrastructure launched 105 distinct attack projects and successfully compromised at least 27 organizations.
Impact
The automated intrusions have yielded severe operational and data security breaches. In just two of the compromised environments, attackers exfiltrated data from over 600,000 unexpired payment cards. Additionally, malicious skimmer scripts were injected into five e-commerce platforms. The campaign has also breached a Fortune 500 hospitality provider, a US airline, an industrial supplies distributor, and a prominent online fashion retailer. Analysts note that where access was achieved, the automated framework completed the intrusion lifecycle in less than a day, and frequently within a few hours.
What to Do
Defenders and SecOps teams managing e-commerce environments must immediately audit internet-facing web assets and review access logs for rapid, anomalous exploitation patterns typical of automated tooling. Key defensive actions include:
- Deploy robust File Integrity Monitoring (FIM) and Content Security Policy (CSP) headers to detect unauthorized JavaScript injection and web skimmers.
- Review API endpoints, authentication boundaries, and edge infrastructure for rapid reconnaissance signatures.
- Implement strict egress filtering and monitor for unusual outbound data transfers indicative of cardholder data exfiltration.
- Prioritize patching for known web application vulnerabilities to reduce the attack surface available to automated scanners and AI agents.
Related vulnerabilities
ThreatsDay Bulletin: AI‑Powered Zero‑Day Chain, 543K Live Secrets, Model Inspection RCE
ThreatsDay highlighted an AI‑driven zero‑day exploitation chain, a list of 543,000 exposed secrets, and a remote code execution flaw in model inspection tools. Defenders using ML inference services and secret repositories should audit, patch, and harden their systems immediately.
Read briefing
CyberXero AI‑Powered Campaign Compromises 628,000 Ukrainian Records via WordPress Exploits
CyberXero, a Russian‑speaking threat actor, used Anthropic’s Claude Code to orchestrate an AI‑augmented attack that leveraged WordPress vulnerabilities (CVE‑2026‑63030, CVE‑2026‑60137, CVE‑2026‑4815) and Cobalt Strike to compromise 429 admin panels and exfiltrate 628,000 records from Ukrainian organizations. The campaign also targeted critical infrastructure in Ukraine and other countries. Defenders should patch the CVEs, monitor for unauthorized admin activity, and inspect exposed directories for AI logs and stolen data.
Read briefing
August 24th Threat Intel: Latvia Breach, AI‑Driven Attacks, and Critical Vulnerabilities
The latest Check Point research bulletin highlights a sweeping data breach at Latvia’s Road Traffic Safety Directorate, affecting over 1.2 million citizens and 200,000 organisations. Sakura Internet exposed up to 1.36 million customer accounts through a compromised rental‑server environment. A Canadian children’s hospital suffered a third‑party data leak, while Berlin’s urban ministries were isolated after a security incident. AI‑driven threats are on the rise: an autonomous GitHub Actions agent stole Snowflake tokens, and U.S. authorities warn of AI‑assisted attacks on Siemens S7 industrial c
Read briefing
CrowdStrike Launches Agentic Identity Provider to Secure AI Agents
CrowdStrike has unveiled its Agentic Identity Provider, a next‑generation identity framework that assigns a unique, verifiable identity to every AI agent deployed across an organization. By tying access rights to real‑time contextual signals, the solution enforces zero‑trust principles and tightens privileged‑access controls for autonomous agents. The move addresses growing concerns around the security of AI‑driven operations and aims to prevent misuse of AI tools while maintaining operational agility.
Read briefing