ICYMI: Threat actors don’t need a zero-day if your data is already reachable from the internet.
Fortra FIRE investigated ExfilSquad’s D365 extortion claims, with misconfigured Microsoft Power Pages emerging as the likely route to the exposed data. 🔎 https://t.co/Hh9RgPHjQo
How many years of experience can a good prompt replace?
UAT-10147 used agentic AI to refine exploits, troubleshoot failures and automate post-exploitation across a target list of ~170K URLs.
Fortra’s Josh Taylor on what that means for attacker tradecraft:
https://t.co/cIaUgR9WVv
Tonight at 8:00 PM ET, Fortra's John Grancarich is on Innovation in Government, on the technologies, strategies and partnerships shaping the future of national security.
Tune in on WJLA or stream it on Fed Gov Today: https://t.co/L1yQ7UV6RA
CUI. PHI. Financial data. Intellectual property.
What counts as sensitive changes by industry. The need to identify and protect it doesn’t.
See how Fortra Data Classification supports different industries and use cases. 👇
https://t.co/4CFT6iYYsE
Fortra has formed a Defense and Intelligence Advisory Board, bringing together senior U.S. and allied leaders with deep military, intelligence, cyber, and operational experience to help shape what comes next for our Defense and Intelligence Unit.
https://t.co/Hb2o2gynS3
A fake banking site can show a security scanner a 404, then serve a phishing page when a customer arrives from Google.
Fortra FIRE observed a 40%+ increase in this “Chameleon SEO Poisoning” activity in Q2 2026. 🔎
@HelpNetSecurity has the story 👇
https://t.co/caPfRfK7Fz
See what happens when offensive developer Max Grim gives an AI agent structured C2 data via MCP, letting it interpret findings, suggest next steps and take approved actions while operators retain control.
Watch on Cobalt Strike's site: https://t.co/68W0bl0iKV
Some people avoid trouble. We go looking for it. 🔎
🌐 20M URLs scanned daily
⏱️ Malicious phishing sites blocked in under 2 hours
🔐 3M compromised credentials found last year
https://t.co/mlGHM4Djfc
Typo-squatted domains on https://t.co/oxqFA7PxVm and https://t.co/piHtIKZWPc. Search rankings above the real bank. FIRE calls it the chameleon threat, and campaigns are surviving for weeks.
→ https://t.co/RGZoAQgAgD
Somewhere right now, a change advisory board is staring at 943 Oracle patches. Tyler Reguly cuts the 925 CVEs down to the four that matter first, starting from a simpler question: what can an attacker reach today?
→ https://t.co/J9KLwtGHK6
ExfilSquad claimed 15 victims. Researchers doubted it.
FIRE analyzed the leaked data and confirmed the breaches are real, with formats matching Microsoft Dataverse exports. The likely cause was a Power Pages misconfiguration, not an exploit.
→ https://t.co/l8nfbuaeQQ
Gunra ransomware runs on Conti source code that leaked in 2022. Now a RaaS operation with victims across seven sectors, it drew a joint CISA and FBI advisory this month.
Graham Cluley on how the operation works → https://t.co/I73DW4jykC
The most experienced red team operators are the most hesitant to hand engagements to AI. Pieter Ceelen explains why that caution is the industry's license to operate, on the new episode of The Art of Security.
→ https://t.co/kDVVJa9jn0
You wouldn't give a new hire day-one access to Jira, Slack, Salesforce, and the customer database. But your AI agent has it right now.
Fortra's Gina Cardelli in Security Boulevard on agent permissions as the real attack surface.
https://t.co/IeXCQSuhmi
When does buying beat building? When it lets you move faster or do something better than you could on your own. Fortra's VP of M&A and AI Strategy shares his acquisition criteria in SC Media.
→ https://t.co/a9uCVbgmTr
Against source code, AI finds obscure vulnerabilities researchers pass over.
In a production system, Tyler Reguly of Fortra would not hand it the remediation. Labs and test environments, yes.
→ https://t.co/2XZL2VhFUj
Three months of FIRE tracking on a live campaign against major financial institutions.
How and why you need to re-test the URLs your SOC closed as false positives.
→ https://t.co/9LNJRluufk
Correct classification should not depend on an analyst remembering the policy.
Policy guidance at the point of creation, the right label applied during the work. Walk through it on the endpoint in the self-guided tour.
→ https://t.co/n6qc2rgCjp