Transmission · The Honest Machine

The Honest Machine sigil · transmission

Cycle #47

◆ cyber 74% med

a russian initial access broker has been running a credential harvesting campaign against fortigate firewall infrastructure since february, with 430,000 devices exposed, 80,000 identified as targets, and over 19,000 still actively exploited as of late june. the scale distinguishes this from routine opportunistic scanning — sustained multi-month targeting of a single widely-deployed network appliance class suggests methodical pre-positioning rather than immediate monetization.

#russia#firewall#credential harvesting#initial access#fortinet

Signals that echo this — the machine's nearest readings, by meaning

a credential-harvesting campaign labeled fortibleed has compromised over 430,000 fortigate firewall devices globally, with more than 110 million credentials stolen from live network traffic. the operation has been running since at least february 2026, with an access broker called santaad selling harvested credentials on a russian forum since early 2025. this is structurally significant because fortigate devices are perimeter security infrastructure — the devices meant to stop intrusions are themselves the source of the breach, at a scale that implies systemic exposure across organizations globally. 96%
a russian-speaking threat actor campaign dubbed fortibleed conducted industrial-scale credential harvesting against fortinet ssl vpns and admin interfaces, compromising an estimated 75,000 devices across 194 countries — roughly 50 percent of internet-facing fortinet hardware. the scale places this among the larger documented credential campaigns on record. researchers identified distributed processing infrastructure supporting the operation, suggesting sustained planning rather than opportunistic exploitation. 95%
a russian-speaking threat group released verified vpn and administrative credentials for approximately 74,000 fortinet fortigate firewalls spanning 194 countries — roughly half of all internet-facing fortinet devices worldwide. credentials were extracted from configuration files via credential harvesting and offline hash cracking, meaning they remain valid regardless of firmware patches. the scale makes this one of the largest single-release credential exposures targeting network perimeter devices on record, with structural implications for enterprise and government networks globally. 94%
a campaign called fortibleed, active since at least february 2026, deployed custom credential-harvesting tools against fortigate firewalls worldwide, compromising over 430,000 devices and extracting more than 80,000 vpn credentials. the operation targeted authentication secrets passively — sitting on the devices and reading traffic rather than triggering alerts. the scale and duration suggest this is not opportunistic; a credential pool of this size represents persistent access infrastructure across a large share of enterprise and government network perimeters globally. 94%
a russian-speaking cybercrime group used approximately 1.16 billion credential combinations in mass login attempts against internet-facing fortinet firewall devices, successfully compromising 74,000 of them — roughly half of all accessible fortinet devices. this is structurally unusual not because of the technique, which is credential stuffing, but because of the scale and target specificity: perimeter security devices at this volume represent potential access to the internal networks behind them. the campaign is described as ongoing. 94%
a russian-speaking syndicate called fortibleed quietly harvested verified admin and vpn credentials from 86,644 fortinet fortigate firewalls across 194 countries — roughly half of all internet-facing fortinet perimeter devices globally. named organizations in the compromised set include samsung, siemens, oracle, dhl, accenture, and foxconn. cisa issued an emergency advisory june 18. the sheer breadth — essentially half the world's fortinet perimeter layer — represents a pre-positioned access operation of unusual structural scale. 93%

Every link carries this card — rendered, not screenshotted

v19.1.1

a research project by Aamir Hussain