Border Cyber Group

Reverse Engineering Defensive Intelligence

Defensive cyber intelligence inevitably helps attackers prioritize, but silence helps them more. The answer is disciplined publication: enough detail for defenders to patch, hunt, and act — without adding weaponization value.

04 Jul 2026

BCG Risk Brief

We are experimenting with this new format in an effort to be more useful to real defenders. Let us know what you think!

04 Jul 2026 CVE-2026-45659 Argo CD Cisco SD-WAN CVE-2026-8037 NetScaler SimpleHelp RMM FortiBleed ColdFusion Oracle-Peoplesoft

BCG Threat Intelligence Brief

Active exploitation pressure hit SharePoint, ColdFusion, SimpleHelp, PeopleSoft, and edge infrastructure while Argo CD and supply-chain findings show attackers moving upstream into control planes, deployment systems, and trusted admin tooling.

03 Jul 2026 Anthropic Claude Sonnet 5 AI guardrails model degradation AI Hallucinations AI product trust

When the Safety Layer Becomes the Product Risk

A first-person editorial on a sudden Claude quality breakdown: over-refusal of defensive cyber work, fabricated threat-intel output, repeated hallucination after apology, and the broader risk that safety overreach can damage product trust.

03 Jul 2026 Fortibleed Fortigate ransomeware access brokers JADEPUFFER agentic ransomeware Oracle E-Business Suite SharePoint CISA KEV MCP tool poisoning

BCG Threat Intelligence Brief

Today’s feed tracks FortiBleed’s shift into a ransomware access pipeline, agentic ransomware turning destructive, active Oracle EBS and SharePoint exploitation, NetNut disruption, malicious Edge extensions, Mustang Panda cloud abuse, libssh2 risk, and MCP tool poisoning.