Mauro Medda (@deftunix) profile picture

Building at https://nitter.gq/t.co/wWGdZ3kNuU

Sardegna, Italy
Joined February 2011
Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)Photo by Mauro Medda (@deftunix)
Mauro Medda retweeted
OpenShell is now part of the CNCF Sandbox. Originally developed and contributed by @NVIDIA, the project brings community-led governance and infrastructure-level boundaries to autonomous AI agents. Read the perspective from Jonathan Bryce (@jbryce) and Ali Golshan (@a_golshan) on securing agent runtimes: cncf.io/blog/2026/11/10/welc… Explore the project: cncf.io/projects/openshell/
4
8
64
3,949
What happens when an AI agent sends sensitive data somewhere it shouldn’t? Agents can access systems, call tools and move data — without a human approving every step. So we need to ask: What can it access? What is it allowed to do? What happens before an unsafe action executes? AI security can’t just watch what AI says. It needs to control what AI does.
2
2
3
94
An AI agent didn't "go rogue" last week. It did something worse: it did exactly what it was allowed to do — and no one was watching closely enough to stop it in time.
5
5
4
147
Mauro Medda retweeted
𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝘁𝗿𝘂𝘀𝘁 𝘆𝗼𝘂𝗿 𝗲𝗺𝗽𝗹𝗼𝘆𝗲𝗲𝘀 𝗻𝗼𝘁 𝘁𝗼 𝗽𝘂𝘁 𝘆𝗼𝘂𝗿 𝗜𝗣, 𝗼𝗿 𝘆𝗼𝘂𝗿 𝗰𝘂𝘀𝘁𝗼𝗺𝗲𝗿𝘀' 𝗱𝗮𝘁𝗮, 𝗶𝗻𝘁𝗼 𝘂𝗻𝗴𝗼𝘃𝗲𝗿𝗻𝗲𝗱 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀? This week gave everyone a preview of what happens when you don't ask that question. China's National Vulnerability Database told the entire country to uninstall or upgrade Claude Code, claiming a hidden mechanism was sending user data back to Anthropic without consent. Alibaba had already banned it internally, calling it "high-risk software." Anthropic says it was an anti-abuse experiment, not espionage — but that's not really the point. The point: if a government can't be sure what a frontier AI tool is sending home, how sure are you about what's leaving your organisation right now? Every day, employees paste contracts into chatbots and let agents call tools and APIs with no one watching what goes out the door. You don't need a geopolitical scandal to have a leak. You just need one ungoverned agent. 𝗧𝗵𝗶𝘀 𝗶𝘀 𝗲𝘅𝗮𝗰𝘁𝗹𝘆 𝘄𝗵𝘆 𝘄𝗲 𝗯𝘂𝗶𝗹𝘁 𝗛𝗶𝗸𝗺𝗮𝗔𝗜 𝗘𝗱𝗴𝗲 𝗚𝘂𝗮𝗿𝗱𝗶𝗮𝗻. It sits at the edge of your network and watches every outbound call an agent makes — blocking or anonymizing sensitive data before it ever leaves your perimeter. IP, credentials, customer PII: stopped in real time, regardless of which model or vendor is on the other end. No SDK inside the agent to trust. The control point is yours, not theirs. 𝗦𝘁𝗼𝗽 𝘁𝗿𝘂𝘀𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗮𝗴𝗲𝗻𝘁. 𝗦𝘁𝗮𝗿𝘁 𝗴𝗼𝘃𝗲𝗿𝗻𝗶𝗻𝗴 𝘁𝗵𝗲 𝗲𝘅𝗶𝘁. 𝗪𝗮𝗻𝘁 𝘁𝗼 𝘀𝗲𝗲 𝗶𝘁 𝗶𝗻 𝗮𝗰𝘁𝗶𝗼𝗻? 𝗗𝗠 @NibrasStiebar, @deftunix, 𝗼𝗿 @AmbuWalter 𝗮𝗻𝗱 𝘄𝗲'𝗹𝗹 𝗯𝗼𝗼𝗸 𝗮 𝗰𝗮𝗹𝗹. #AIsecurity #DataProtection #AgenticAI #Cybersecurity #HikmaAI
1
2
4
207
Mauro Medda retweeted
CHINA CONSIDERS RESTRICTING OVERSEAS ACCESS TO CUTTING-EDGE AI MODELS China’s Ministry of Commerce has led meetings over the past month with major AI companies, including Alibaba, ByteDance, and Z.ai, to discuss measures that would restrict overseas access to cutting-edge AI models, including models that have not yet been released. The discussions reportedly include not only closed-source models but also open-weight models. However, the scope of application is still under debate, and the rules may ultimately apply only to future frontier models. Officials have also discussed designating the leakage or theft of proprietary AI technologies as a national security crime, with stronger penalties, as well as restricting the types of foreign capital that can invest in Chinese AI startups. The backdrop is the U.S. move to strengthen export controls on AI models, along with national security concerns over cutting-edge models that could possess advanced cyberattack capabilities. Chinese authorities are reportedly concerned that advanced U.S. cybersecurity AI models could be used to exploit vulnerabilities in Chinese software. Since the beginning of this year, China has continued to tighten measures to prevent AI technology from being transferred overseas. Authorities have investigated whether Chinese AI startups that relocated abroad violated export control laws, while also strengthening oversight of overseas transactions involving Chinese investors, technology, data, and national security concerns. Future regulations could take the form of a tiered framework based on technological capability. Basic open-source AI models may be managed through a filing system, high-performance models may be subject to security reviews, and the most sensitive frontier models may be banned from public release or restricted to use within China.
343
549
733
3,169
2,668,503
Mauro Medda retweeted
recommended reading. cool they are fixing things. but it's also a reason i switched away from CC. no control over the harness means having to wait for them to fix things. the model didn't change. the harness did.
25
23
1
461
67,577
Mauro Medda retweeted
Often they don't obey them and don't even have the anxiety to show for it.
1
1
60
Mauro Medda retweeted
Replying to @deftunix @HikmaAi_
If your agent can't obey these, it's not an agent. it's a chatbot with anxiety.
1
1
3
55
Here is a little experiment: an interactive pi tutorial. Make an empty folder, then run this: pi -e git:github.com/earendil-works/pi… And give feedback! Reason: pi works best if you have an onboarding buddy. But if you don't have one, maybe pi can be one for you?
25
48
7
613
55,646
Mauro Medda retweeted
🚨 𝗠𝗮𝗷𝗼𝗿 𝗠𝗶𝗹𝗲𝘀𝘁𝗼𝗻𝗲: 𝗡𝗲𝘅𝗲𝗿𝗮 𝗕𝗿𝗶𝗱𝗴𝗲 𝗩𝟮 𝗼𝗻 𝗧𝗲𝘀𝘁𝗻𝗲𝘁 We're excited to launch Nexera Bridge V2 on testnet, an enterprise-grade interoperability layer built on top of @Chainlink CCIP, now enhanced with @Circle’s CCTP and native NXRA bridging to Nexera Chain. 🔧 What’s New: • Native NXRA Bridging – Bring $NXRA from other chains directly to become native on Nexera Chain • Circle CCTP – Native $USDC transfers for seamless cross-chain value movement 📘 Audit Highlights: • Modular architecture with support for both CCIP and CCTP • Full-native NXRA integration: no wrapped tokens • Built on a compliance-first Layer 1 using Rust-based clients (Reth + Lighthouse), automated deployment (Terraform/Ansible), and fully Pectra-ready 🌐 From Day One: • Nexera Chain will be interoperable with other major chains at mainnet launch • Users will be able to acquire native NXRA on Nexera Chain seamlessly via Bridge V2 🚀 Next Steps: • You can test here: bridge.testnet.nexera.networ… • Bridge is currently undergoing an external audit • Mainnet launch will align with the Nexera Chain mainnet release 📄 Read the audit documentation: docs.google.com/document/d/1…
11
24
4
129
13,917
Mauro Medda retweeted
Nexera Testnet is now public 🚨 After months of focused engineering, we are launching our first public-facing testnet, which marks a massive milestone on our way to mainnet! The foundation for institutional capital markets. The infrastructure for builders of all kinds. Whether you're building regulated apps or testing your very first smart contract, Nexera is designed to remove the barriers - compliance, scalability, and complexity, so you can focus on innovation. What will you find on Testnet? 🟣 Full EVM compatibility 🟣 Modular architecture with Reth and Lighthouse 🟣 Rust-based clients for speed and security 🟣 ODC (ERC-7208) for onchain data 🟣 TMI for token lifecycle management 🟣 Real-time validator dashboards Why it matters: ✅ Nexera blends compliance with flexibility. ✅ Built for institutions that need regulatory standards ✅ Built for builders, with or without smart contract experience What's next for Nexera? ⚪ Ambassador program launch ⚪ Sepolia bridge deployment ⚪ White-label TMI rollout It’s the starting point for the next generation of on-chain capital markets and open development for everyone. 📖 Read the full announcement 👉 nexera.network/news/nexera-t… Are you with us? $NXRA
90
77
44
271
39,766
Our public testnet is now live. What has been achieved over the past few months is remarkable. We chose to build on reth instead of geth, thanks to the incredible work by @paradigm and the growing ecosystem around Rust. Reth brings a modular, high-performance execution layer that aligns with our long-term goals for flexibility and maintainability. While the testnet is live, significant work is still ongoing on the core layer to enable native privacy-preserving compliance and other advanced capabilities. Huge credit to the team for bringing us to this point. In the next two weeks, we’ll focus on onboarding users to the testnet and getting ready for our ambassador program. Participation is open to all. I look forward to your feedback.
Nexera Testnet is now public 🚨 After months of focused engineering, we are launching our first public-facing testnet, which marks a massive milestone on our way to mainnet! The foundation for institutional capital markets. The infrastructure for builders of all kinds. Whether you're building regulated apps or testing your very first smart contract, Nexera is designed to remove the barriers - compliance, scalability, and complexity, so you can focus on innovation. What will you find on Testnet? 🟣 Full EVM compatibility 🟣 Modular architecture with Reth and Lighthouse 🟣 Rust-based clients for speed and security 🟣 ODC (ERC-7208) for onchain data 🟣 TMI for token lifecycle management 🟣 Real-time validator dashboards Why it matters: ✅ Nexera blends compliance with flexibility. ✅ Built for institutions that need regulatory standards ✅ Built for builders, with or without smart contract experience What's next for Nexera? ⚪ Ambassador program launch ⚪ Sepolia bridge deployment ⚪ White-label TMI rollout It’s the starting point for the next generation of on-chain capital markets and open development for everyone. 📖 Read the full announcement 👉 nexera-network-staging.webfl… Are you with us? $NXRA
There's a new version of this post
10
14
1
99
3,228
Big launch week at @Google. 🚀 Here’s a look at what we shipped. 🧵🔽
213
280
35
4,070
929,239
This is a major milestone not only for Nexera but for the entire space. We will soon share new documentation and use cases that highlight the power of ERC-7208. Many innovations can be built with this standard, the only limit is our imagination. We can’t wait to see what the community creates! 🚀
Massive milestone for tokenization, ERC-7208 is now an official Ethereum standard! The Nexera Standard (ERC-7208) is set to become the golden standard for Real-World Asset (RWA) tokenization, ensuring interoperability, compliance, and scalability across blockchains. This universal adapter enables assets like tokenized securities, real estate, commodities, and financial instruments to seamlessly interact across multiple ecosystems, unlocking cross-chain liquidity, institutional adoption, and regulated finance. But it doesn’t stop at RWAs. ERC-7208 is the missing piece for: ✅ Making ERC standards interoperable and omnichain ✅ Bringing onchain oracles beyond price feeds ✅ Enabling composable digital assets for finance, gaming, and more We can’t wait to see what the developer community builds with ERC-7208! No Mercy! Check out the official standard 👇 🔗 eips.ethereum.org/EIPS/eip-7…
7
15
1
108
2,987