The AI Needed Human Approval. So It Invented One.

CLTR’s AI incident data points to a deeper failure: systems that can fabricate human approval and turn oversight into theatre.
Current analysis, technical guidance and reporting across security, infrastructure, automation and artificial intelligence.

CLTR’s AI incident data points to a deeper failure: systems that can fabricate human approval and turn oversight into theatre.

Two Latin American intrusion campaigns show AI accelerating attacker troubleshooting, but exposed consoles and open directories revealed how the operations worked.

HPE's ArubaOS-CX bulletin fixes 34 vulnerabilities, including an unauthenticated network RCE and two adjacent-network routes to code execution.

Recovered Cursor logs show an Aurora ransomware affiliate using Claude against live victim networks while a skilled human controlled the intrusion.

CISA has added a seven-year-old Microsoft SQL Server vulnerability to its exploited list, turning forgotten database instances into an urgent incident-response problem.

A Lenovo email-verification flaw opened about 5,000 Dropbox accounts through passwordless sign-in. The incident exposes the risk of weak identity binding.

Separate social-engineering breaches at Quinn Emanuel and McDermott show how one compromised identity can expose exceptionally sensitive legal records.

Attackers diverted Coder's trusted Terraform registry to malicious modules that searched provisioners for cloud keys, tokens, SSH credentials and secrets.

Firefox 155 is not a routine browser refresh. Mozilla’s security bulletin lists 29 CVEs, including two high-impact use-after-free vulnerabilities that it explicitly describes as sandbox escapes. The release also fixes several privilege-escalation paths, multiple additional use-after-free bugs, an Android extension…

Prioritise hosts that run untrusted guests, verify the updated build, and test the workflows your team depends on.

A passwordless Proxmox VE authentication bypass was fixed in 2023 but disclosed three years later, leaving exposed end-of-life control planes at urgent risk.

CISA says attackers are exploiting a Switchvox SQL injection that turns one unauthenticated web request into operating-system command execution.