AI is making cybersecurity research faster and cheaper than ever. Currently, these capabilities favor defenders, but that window of opportunity won't stay open for long.
We've signed @OpenAI's Call for Collective Action on Cyber Defense. The time to act is now.
An open letter for a global surge in cyber defense, signed by over 100 organizations including Anthropic, AWS, Google, Microsoft, OpenAI, and Oracle. https://t.co/uKXPS8LdAU
AI is making cybersecurity research faster and cheaper than ever. Currently, these capabilities favor defenders, but that window of opportunity won't stay open for long.
We've signed @OpenAI's Call for Collective Action on Cyber Defense. The time to act is now.
An open letter for a global surge in cyber defense, signed by over 100 organizations including Anthropic, AWS, Google, Microsoft, OpenAI, and Oracle. https://t.co/uKXPS8LdAU
@giovignone:
"We signed OpenAI’s Call for Collective Action letter because it clearly outlines the stakes of the world Octane operates in already.
Our approach to agentic security finds exploitable vulnerabilities in code that's been professionally reviewed for years. We see first-hand how vulnerable some of these systems are.
The defender's advantage is real, but it's temporary, and it goes to whoever moves fastest."
@gdb Octane is proud to be a day one signatory to this call for collective action on cyber defense.
It make the stakes clearer than ever, and we make the solution.
@kirkthebaird@v12sec@cantinasecurity@CertiK Octane is more than an automated scanner. Think of it as the context-aware agentic engine behind a wide range of security services.
A hands-on demo is always just a click away:
https://t.co/5IlTAc97Mm
@kirkthebaird@v12sec@cantinasecurity@CertiK Octane is more than an automated scanner. Think of it as the context-aware agentic engine behind a wide range of security services.
A hands-on demo is always just a click away:
https://t.co/5IlTAc97Mm
Automated tooling that drowns you in alerts is a SecOps nightmare.
Context-aware agentic security helps turn a noisy stream of findings into a tight set of evidence-backed attack paths.
Run it like a round-the-clock red team.
We released a new Cybersecurity Advisory, A Tale of Two SOCs: Insights From Two Red Team Assessments. Learn which actions make the difference for quickly containing a breach to protect your org from cyber threats 👉 https://t.co/NzbCFzJJyw
I think that “human in the loop” remains the best approach right now to agentic security for mission critical software.
Agentic only works fine when the stakes are low, human only is just limiting and capacity constrained.
@noegrets3@Oracle Any product that passes untrusted CUE/BIN images to an affected build could inherit the same host-side crash path.
https://t.co/a4jSuzC0UI
You use VirtualBox to isolate software you don’t trust.
But we crafted a CUE file that corrupts memory and crashes VirtualBox on the host before the sandbox can start.
@Oracle assigned CVE-2026-71125, our second VirtualBox CVE in 28 days.
https://t.co/qDKBYg6hMe
@atomiqlabs Wishing you luck. AI-assisted security is the best way to counter the growing number of AI-assisted attacks. Reach out whenever the team is ready, we'd love to help.
Every OmniVault gets its own isolated proxy contracts.
If one vault has issues, others remain unaffected. Your funds only touch audited Dexalot contracts.
Security through architecture, not promises.
Dual audits: @hackenclub + @octane_security
https://t.co/XYvSVlMBkS
@HHorsley Same goes for AI – eventually it all just becomes "software".
Finance and every other industry will use their own complex combinations of LLMs, blockchains, existing internet infrastructure, and tech that is yet to be invented.
We spent $609.59 on the compute that found three confirmed CVEs.
Chromium, VirtualBox, and QEMU are among the most reviewed codebases on earth, with thousands of contributors, continuous fuzzing, and fully public source.
Our researchers found bugs in all three.
Anyone can buy compute. At Octane, our bills are low because our researchers' aim is true.
Get the same agentic security analysis in your workflow today.
https://t.co/8rvlw9uYYp
A high-severity finding in Nethermind's Ethereum client paid out $50,000, while winning the Monad audit contest earned us $168,782.57.
https://t.co/VTQQ1LBAsz
Octane's AI just won one of the biggest contests in blockchain history (largest unconditional pot ever), placing first out of 1,600 SRs
We caught 3/4 high-severity findings while working with a novel Rust and C++ codebase
Octane secures crypto’s critical infra