Pinned Tweet
Introducing Antares: @Cisco's family of small language models for locating known vulnerabilities in code.
Antares-350M and Antares-1B are live on Hugging Face now. They can outperform many larger closed- and open-weight models at a fraction of the cost.
Small enough to run locally. No shipping sensitive codebases to the cloud.
Why it matters: vulnerability triage is expensive and slow. Antares helps democratize AI-assisted security for all.
Explore the models + read the new Vulnerability Localization Benchmark: blogs.cisco.com/ai/introduci…
AI agents are moving beyond screens and servers into the physical world.
The new DefenseClaw Edge Connector brings AI agent security to robots, drones, and IoT devices — enforcing security policies locally, in microseconds.
Physical AI is moving intelligence to the edge. DefenseClaw moves security with it.
Read more: blogs.cisco.com/ai/agents-at…
Cisco AI retweeted
Is your team still stuck in the cycle of manual network management?
In this episode of Inside Product, Jeetu Patel and Jeff Schultz discuss why automation is the backbone of Cisco’s current portfolio strategy. They aren't just talking theory—they’re looking at how we can deliver real results for your business by automating the complexities of the AI era.
What’s the most time-consuming task you’ve successfully automated in your network this year? Let us know below! 👇
Check out the full episode here: cs.co/6013BHExQx
How is @Cisco securing agentic AI at enterprise scale?
Join experts from Cisco’s CIRCUIT and AI Defense teams on October 21 to explore how Cisco AI Defense and DefenseClaw protect internal AI workflows through runtime guardrails, scanning, and policy enforcement.
Register now: cloudsecurity.cisco.com/webi…
How can AI agents continuously improve from real-world performance?
Cisco @fdtn_ai introduces FAFO (Fully Automated Flow Optimization), turning production traces and sparse user feedback into an automated optimization loop.
The result: recursive self-improvement grounded in real-world data.
Read more: cisco-foundation-ai.github.i…
September at Cisco Foundation AI: new benchmarks for understanding where cyber-capable AI should be strong — and where it should fail.
🔍 VLoc Bench: 500 vulnerabilities, 290 repos, 147 CWEs
📊 27 models evaluated
🤖 Antares-3B ranks second overall on VLoc Bench
🛡️ Safety-VLoc-Bench explores attacker/defender asymmetry
See what @fdtn_ai shipped in September: blogs.cisco.com/ai/foundatio…
How well can a zero-shot decision model handle content safety?
@Cisco put Jev and Laya head-to-head with a policy-trained classifier across 24 harm categories and 6 datasets.
The takeaway: policy-specific training still wins, and Jev came surprisingly close to a 31B-parameter LLM judge on zero-short safety classification tasks.
See what we learned: blogs.cisco.com/ai/jev-decis…
Cisco AI retweeted
Cisco Talos observed multiple AI agents coordinating against a single target. Adversarial automation is already here.
Nick Biasini opens our new series, "Cisco Security in the AI Era."
▶️ Watch: cs.co/6011BGLkBH
Cisco AI retweeted
In this episode of "Inside Product at Cisco", Jeff Schultz joins Jeetu Patel to delve into the evolution of the Cisco Portfolio and how this focus on critical business outcomes for the AI era is delivering real results for the ecosystem.
Watch the full conversation on YouTube: cs.co/6010BGmwBM
Cisco AI retweeted
The AI era will be built on innovation and trust. Proud to be part of the ecosystem doing both.
Today, with over 100 industry partners, we introduced the NVIDIA Open Agent Safety Platform, bringing together OpenShell and Sentry.
Artificial intelligence is extraordinary technology that will advance discovery, productivity, security, health, and prosperity for generations to come.
But its full promise can only be realized when people have confidence that AI is being built to be safe and deployed with wisdom and responsibility.
This is bigger than a single product. It's the beginning of an open ecosystem to build the trust layer for safe agent systems.
Together, we are building the foundation of the AI economy.
Trust and innovation are not in conflict. Safety is how trust is earned. We must build not only the most capable AI, but the most trusted AI, so that this extraordinary technology can realize its enormous promise for the world. nvda.ws/4hcoq7m
As AI models increasingly power agents, understanding their susceptibility to jailbreaks and prompt injections is critical.
A new update to the @Cisco LLM Security Leaderboard expands model security evaluations across text, image, and audio—with 102 new evaluations since June.
Explore how 136 models hold up against adversarial attacks across modalities: blogs.cisco.com/ai/llm-secur…
As AI agents take on more critical security tasks, how do you make sure they can be trusted?
@jpatel41, @djsampath, and Matt Caulfield discuss why identity is key to securing the agentic era.
Watch the conversation to learn more: youtube.com/watch?v=rGL4RCkR…
Bigger AI models aren’t always better for security.
Join Cisco @fdtn_ai experts @aminkarbasi and Didier Chapoteau to learn how small language models trained for specific security tasks can improve security workflows.
Using Cisco’s open-weight Antares models, they’ll explore where specialization can outperform scale.
Register now: cloudsecurity.cisco.com/webi…
Most security benchmarks assume the vulnerable code is already known. In practice, defenders first have to find it.
Cisco @fdtn_ai introduces VLoc Bench, a benchmark for repository-scale vulnerability localization: given a CWE and read-only access to a real codebase, can an agent identify the files associated with the weakness? (1/5)
Explore VLoc Bench and read the research:
cisco-foundation-ai.github.i… (5/5)
Read the blog to learn more: cisco-foundation-ai.github.i…
Cisco AI retweeted
Networking isn't just about software. The real breakthroughs happen at the physical layer. Learn why Cisco is co-designing custom #silicon for better efficiency ➡️ cs.co/6013BGyGVt
#innovation #infrastructure @jpatel41
As AI agents gain more autonomy, the risks are evolving too.
We’ve updated our Integrated AI Safety and Security Framework with a new objective: Agentic Autonomy Failures.
Explore the new taxonomy for Excessive Agency, Goal Drift, Reward Hacking, and more: blogs.cisco.com/ai/security-…
