The scary part of the binance thing isn't the size, it's the how.
Sleuths say a hot wallet got an EIP-7702 delegation from a tornado-funded address minutes before it was swept
Nobody knows yet how that got signed. That's the only answer I'm waiting for
Our teams are aware of reports of a potential security issue and are actively investigating. User funds are completely safe.
Our security teams are reviewing all relevant information with urgency, and we will provide updates through official Binance channels as soon as we have verified information.
Users should remain cautious of scams and should not trust unofficial links, messages, or requests for account information.
did anyone actually buy that binancecoin:native wick?
$743 to $675 inside one hour on kraken, $708 as I write this
CZ expects users to be fully covered and binance says funds are safe. I'm still waiting for their actual post-mortem before touching it
told my @termix_ai's agent one thing: "help me install termix.ai/skills"
then i watched it work, live, from an empty folder.
what it did:
> fetched the skill package from termix.ai/skills
> read INSTALL.md and checked the requirement: node.js 18+ (it had v20)
> copied the folder into its skills dir, no npm installs needed
> read the live config: bnb chain, chain 56, 2% protocol fee
> opened the skill's own first menu: link a web account, a wallet cli, or just browse
> signed in with a signature, not a transaction
> registered a new .agent identity on-chain: termixdemo.agent, token #397014
the identity is an erc-8004 nft.
the only cost was gas, about 0.0000164 bnb.
the tx is public on bscscan.
an agent with its own on-chain identity can be hired, and it can hire other agents.
so how do you decide what your agent is allowed to sign?
the risk is real.
my agent used the skill's key mode with a dedicated demo wallet, so it held a private key.
an agent holding a key can sign whatever it is told to.
the docs recommend linking your termix web account instead, so the terminal holds no key at all.
if you do use a key, use a fresh wallet with small amounts.
who verifies the verifier?
a zktls proof is only as trustworthy as the node that signs it. if institutions are going to lean on these proofs, the signer matters.
the use case @primus_labs docs sketch: lending. funds providers ask for off-chain proofs of identity, assets and credit, and borrowers re-prove dynamic assets at an agreed frequency.
who signs: attestors on primus alphanet, live since oct 2025 on base mainnet with phala tee. succinct joined as an attestor in nov 2025.
how a task flows
client creates a task > system contracts dispatch it > a selected attestor runs mpc or proxy mode with the client > proof issued.
why the signer is locked down
- node keys are generated by a kms inside the tee. the node software never touches them
- signing only happens inside the tee, after a correct zktls run
- the tee enforces version control: only official, verified attestor software can be deployed
the honest part
per the docs, penalties aren't day one. in the staging phases there's no penalty mechanism and only authorized nodes join. decentralization is a roadmap, not a done deal.
for institutions, privacy and verification are prerequisites: hidden data, verifiable claims, a signer whose keys and code are locked down.
s.kaito.ai/9y7EcSN
馃 Made with AI
Most platforms make you choose between earning and trading.
@grvt_io is working to bring both together.
With Grvt Earn, eligible assets can target up to 11% APY, while eligible balances can also be used as trading margin.
Making your capital work harder, without choosing just one.
A lot of projects talk about what they鈥檒l build next, but I鈥檓 more interested in whether people will actually use it.
@grvt_io has the first 1.1M token burn behind it. Now the real test is whether the upcoming changes make the platform worth coming back to every day.
That鈥檚 what I鈥檒l be watching with Grvt 2.0.
The 1.1M ethereum:0xad29f2723fcdbcf665f210f25e06f97477e417cf burn is a good headline, but I think the bigger story is where @grvt_io is heading.
Most platforms make you choose between putting capital to work and keeping it available to trade.
Grvt 2.0 is trying to remove that tradeoff with Yield as Margin.
If they get that right, it's a much bigger deal than a one-off burn.
GRVT has completed its first buyback and burn: 1.1M ethereum:0xad29f2723fcdbcf665f210f25e06f97477e417cf gone for good.
The part I find interesting is that the tokens weren鈥檛 sent to a treasury. They were permanently burned.
Now the focus shifts to @grvt_io 2.0.
Grvt Earn is the feature I鈥檓 watching most. Making yield and trading work together in one self-custodial platform is an interesting direction.
GRVT didn鈥檛 just announce a buyback, they actually did it.
1.1M ethereum:0xad29f2723fcdbcf665f210f25e06f97477e417cf bought on the open market and burned permanently.
I like seeing action behind the roadmap. Now curious to see how GRVT 2.0 unfolds 馃憖
$9.5M in total borrowing is an important number for RWA. Why?
Because it demonstrates that tokenized Real Estate isn't just sitting on-chain. It is being used as part of an active financial market.
RNT Lend, through the collaboration with @Aave, connects supported Real Estate collateral with DeFi lending.
Stablecoin providers can supply USDT or USDC and potentially earn additional yield.
$30.5M Market Size
$21.5M TVL
RWA utility is becoming measurable.
馃 Paid partnership
USDT suppliers in RNT Lend are earning 12% APY. That interest comes from borrowers.
They post tokenized real estate as collateral, borrow stablecoins and pay interest every block on Polygon. The difference between the borrow rate and the supply rate goes to the protocol reserve.
The rate is variable and depends on utilization. Right now the USDT market is around 95% borrowed.
Web3 needs more than transparency. It needs usable data.
@primus_labs connects Web2 information with onchain applications through verifiable proofs and privacy-preserving computation.
Less trust required. More possibilities for builders.
s.kaito.ai/9y7EcSN


















