HeadFlash

Security

Optimizely breach pushed malware onto CNN, Fox News and Salesforce sites

A supply chain attack on Optimizely injected malicious JavaScript into customer projects, hitting major sites and prompting a cleanup scramble.

Listen

This edition was produced with artificial intelligence. Text and voice are generated automatically.

Optimizely supply chain attack injected wallet-stealing script into major sites

Optimizely said an unauthorized party exploited a vulnerability in its Experimentation account signup flow to inject malicious JavaScript into customer projects, causing sites to load malware. Researcher Scott Helme identified 188 infected files belonging to 178 organizations; urlscan.io confirmed the script loaded or attempted to load on CNN, Fox News, BBC, Zoom, GitLab, Salesforce and others. The code pulled track.js from 1688360[.]cc, which swaps cryptocurrency wallet addresses, tries to exfiltrate card details, and loads iOS exploit kits. Optimizely blocked the vulnerable path, removed injected code, revoked tokens and reset affected accounts. Wallets tied to the campaign had received roughly $7,000 as of November 16, though some predated the attack.

Optimizely hack infected CNN, Fox News, BBC, Salesforce sites with malware →

Topics: Supply chain attacksMalware & infostealersCrypto security

Google: attackers deploy autonomous AI agents to weaponize patches at machine speed

Google Threat Intelligence Group reported that threat actors have moved past manual prompting to autonomous multi-agent offensive frameworks — ARTEX, CAIRN, CyberStrikeAI and PentAGI — to run multi-stage intrusions faster. GTIG said financially motivated actors and suspected China-aligned espionage actors are experimenting with these open-source tools, which automate discrete phases of live intrusions. Google also flagged LLM-driven pipelines that reverse-engineer patches via differential analysis, collapsing the gap between disclosure and exploitation, as seen with PaperCut NG/MF flaws in September. It cited LLMjacking, where stolen LLM credentials and cloud resources are resold through proxy services APIMart and JetNight, and a Chinese phishing platform, Jieyou, whose LLM chatbot lets low-skilled affiliates customize lures in natural language.

Google Finds Hackers Using AI Agents to Build and Run Exploits From Software Patches →

Topics: Autonomous AI agent cyberattacks in the wildState-backed hackers using AI

VulnCheck: Project Glasswing findings pile up as fixes stall

Patrick Garrity of VulnCheck examined vulnerabilities found by Anthropic’s Project Glasswing since its April 7 launch. Of 37,966 findings Anthropic reports, only 7,349 (19.4%) are public, and just 217 of those are marked fixed. Some 4.4% (1,654) are marked duplicates — 7.5 times more than fixes — and 419 were withdrawn. The ledger claims a 94.2% true positive rate, up from 90.8% at launch, despite more findings withdrawn or duplicated than fixed, which Garrity attributes to a likely feedback loop missing from the triage rates. He wrote that AI is valuable for finding vulnerabilities, but seven months in the evidence does not match the launch narrative, and the triage burden is shifting onto open source maintainers Glasswing promised to help.

Glasswing overwhelms maintainers with underwhelming findings (VulnCheck) [LWN.net] →

Topics: AI-discovered vulnerabilities and AI bug-report slop

Ethiopian Airlines warns ShebaMiles members after cyberattack

Ethiopian Airlines warned customers through its ShebaMiles frequent-flyer programme that some of their information may have been exposed in a recent cyberattack. The Addis Ababa-based carrier called the incident of limited scope and said there is currently no evidence customers’ financial information was accessed. It did not give the date of the intrusion, specify which data was involved, estimate how many people were affected, or say whether bookings and flight operations were impacted. ShebaMiles had more than 5.38 million members when the airline announced an Etihad Airways loyalty partnership in 2025. The disclosure follows Air Botswana’s statement that it detected unauthorized access on November 3 to systems holding employee and internal information, which it described as suspected ransomware.

Ethiopian Airlines reports cyberattack a week after Air Botswana →

Daily tech-news flash

The flash, every weekday.

Five minutes on AI, privacy and security — one short email per niche you pick, with a podcast to match.

Your niches