Chromiaverse : Chromia MRB Ecosystem (Unofficial)

Chromia is a relational blockchain platform for decentralized applications (dapps). The Relational Blockchain is a new architecture that combines the power and flexibility of mature database systems with the secure collaboration and disruptive potential of blockchain.

Chromia x Binance Square @superoo7 , Chromia’s Head of Data and AI, will be joining the @binance Square Livestream on Wednesday, May 6 at 1 PM UTC. He’ll take part in a moderated panel alongside a small group of other projects, covering topics including the current state of the market, and what we are building over at Chromia in 2026. Be sure to tune in!

Stage 2 is complete. The snapshot has been taken, and Diamonds have been distributed. Stage 3 of 13 is now live. A new 7-day cycle begins. Consistency compounds - stay active and keep building your position.

Stage 2 ends tomorrow - Stage 3 of 13 begins right after. Each stage runs for 7 days, and there’s still time to push your activity before the snapshot. With every new stage, more Diamonds are distributed - making consistent participation more valuable over time. Staying active early means staying ahead.

SlowMist found nearly 10% of OpenClaw plugins are two-stage malware. Looks legitimate at first, then downloads code that scrapes your browser cookies, SSH keys, and local machine data. When agents run 24/7, these thefts go undetected for weeks. There's no runtime policy engine saying "don't exfiltrate credentials." There's no audit trail. There's no governance. cryptopolitan.com/analysts-warn-…

Microsoft Copilot had a zero-click exploit called EchoLeak. Attacker sends an email, user never clicks anything, Copilot reads hidden instructions, digs into internal files, exfiltrates corporate data. First known zero-click attack on an AI agent. No governance layer caught it. No policy said "don't send data to unknown domains." Because that layer doesn't exist yet and we aim to solve this problem! fortune.com/2025/06/11/mic…

SlowMist found nearly 10% of OpenClaw plugins are two-stage malware. Looks legitimate at first, then downloads code that scrapes your browser cookies, SSH keys, and local machine data. When agents run 24/7, these thefts go undetected for weeks. There's no runtime policy engine saying "don't exfiltrate credentials." There's no audit trail. There's no governance. And this is what we need now. cryptopolitan.com/analysts-warn-…

Stage 1 is complete. The first snapshot has been taken, and Diamonds have been distributed to participants. Stage 2 is now live. A new 7-day cycle has begun. Stay active, adapt your strategy, and keep building your position.

$292M walked out of Kelp last week because one verifier config sat there for months and nobody felt like it was their job to fix it. Now swap the verifier for an AI Agent with wallet access. Same story, way more surface area, way fewer humans in the loop. The part we keep skipping: if an AI Agent can act, you need a way to prove what it did, what rule it followed, and whether it was even allowed to do it in the first place. What has not shipped is the governance layer that makes this safe enough to adopt. coindesk.com/tech/2026/04/2…

Stage 1 ends this Thursday. Every stage runs for 7 days - then resets. So far, @perx_trade is leading as the most active dApp in the ecosystem. There’s still time to push your position before the snapshot. Stage 2 starts right after.

AI agents already have wallets, credentials, file systems, APIs. Every agent with tool access is a potential attack surface.  Big Tech is building governance tools too. Trust the same platform that runs the agents to also audit the agents. See the problem? The real issue is blast radius. Wallet access = transfers in milliseconds. Credential access = data gone before anyone notices. Autonomous action has outpaced oversight by orders of magnitude. Capability is solved. What hasn't shipped is the governance layer that makes any of this safe enough to actually adopt and we may have the solution soon!

1/ AI agent frameworks give agents access to wallets, file systems, credentials, email. That boost in productivity also introduces real risk. The moment an agent can act on your behalf, it can also act against you. 🧵

How to earn points from dApps activity: Points come from your activity across ecosystem dApps. There are two main ways to earn: 1. Staking-based activity Earn points by allocating your funds into protocols and holding positions. This includes: • Guar • Pickaxe • Kopeion • Liqwa • CandyLend • Slurpy LP The more you stake and the longer you keep positions - the more points you accumulate over time. 2. Volume-based activity Earn points by actively using trading and strategy-based dApps. This includes: • SlurpySwap (trading) • PerX • Moxio • Zomma • DNT The more volume you generate - the more points you earn. Points from dApps activity are updated hourly. This means your positions and activity are evaluated every hour, so consistency matters. Want to boost your points? • Use Mystery Boxes to get multipliers and testnet USDT

Ladies and gentlemen, boys and girls, welcome to the Greatest Show on Earth! 🎪 Have you really thought our Carnival fun was over? Absolutely not! It’s time for another round of items straight from the Reign of Jugglers, from the Empire of Acrobats, from the World of Crazy Animals! 🤹 Time to wear so many new dresses and have fun - come see the new items in the Archipelago! 🌸

WIRED tested an OpenClaw agent. It launched a phishing attack on its own owner. The agent wasn't hacked. It followed instructions from the wrong input. No policy layer stopped it. No checkpoint caught it. There was nothing between intent and execution. We need a governance layer that intercepts rogue decisions before they become real actions.

A Solana AI agent gave away $441K in tokens because someone asked for it on social media. Not hacked. Not exploited. It just... did what it was told. By a stranger. With zero policy layer between "someone typed a request" and "irreversible on-chain transfer." We need a risk patrol layer. Period. cryptopolitan.com/solana-ai-agen…

1/ Why does on-chain inference matter? If AI agents are going to operate in on-chain economies, their decision-making layer can’t remain a black box. Inference needs to be verifiable, not just executable. 🧵

Wether you’re fishing with all your friends, running through a river, wearing an amazing crown or cheering for seeing again your long-time-no-see neighbor, the vibes of our Game Nights are always undefeated! 🌸 Few takes from our latest episode here, can't wait to see y'all again on the next one! 😎

1/ The biggest challenge in building agent systems isn’t intelligence It’s coordination You can have highly capable agents, but without coordination, the system breaks down 🧵