Threat Intelligence
Adversary methods, active campaigns and evidence-based threat research. No fear. Just analysis.

PLAID KARTE Breach: Account Emails, Integration Secrets and Customer Data Exposed
Updated November 18: DHC and Morisawa disclosures expand the picture of PLAID’s KARTE breach, covering customer records as well as account and integration data.
Claude Fable 5.2 and Mythos 5.2: What Stronger AI Vulnerability Research Means for Defenders
Anthropic says Mythos 5.2 improved CVE rediscovery while Fable 5.2 is publicly available with tighter safeguards. What the vendor's November research actually establishes.
South Korea Investigates Government Email Intrusion After Abnormal Access to Around 300 Accounts
South Korea's Interior Ministry says attackers exploited a single sign-on weakness after reaching government email systems through vulnerable local-government networks.
Ransomware Claims Climb as Smaller Groups Experiment With AI-Assisted Operations
Research cited by WIRED points to nearly 3,000 alleged summer ransomware incidents and about 111 claiming groups, but leak-site counts are not confirmed victim…
Cursor Traces Malicious Claude Code Extension to Exposed Fireworks AI Access Key
Cursor's November 13 incident report links an exposed vendor key, historical coding-agent traces and malicious extension builds that reached as many as 5,000 users.
FBI Warns ARTEX AI Framework Is Accelerating Cyber Intrusions Across Multiple Sectors
Updated November 17: FBI, Google and Amazon research describe ARTEX-related activity, with a key distinction between observed attacks and unverified agent-reported outcomes.
ShinyHunters and Oracle PeopleSoft: What a Renewed Exploitation Wave Reveals
Google researchers linked renewed exploitation of an Oracle PeopleSoft vulnerability to a group they track as UNC6240.
PHP2Shell: Why Attackers Move Before a Security Advisory Is Public
Amazon researchers observed suspicious PHP2Shell activity before the October 2026 public disclosure, complicating assumptions about when attacks begin.
WFX Phantom and the GitHub Actions Secret-Stealing Campaign
Researchers traced a fast-moving campaign in which compromised GitHub accounts were used to alter CI workflows and obtain secrets.
Agentic Hacking Frameworks: What ARTEX Research Says About Automated Intrusions
New incident-response research shows how threat actors are incorporating autonomous tools into different stages of a cyberattack.