zktls

It’s getting easier to place orders with AI Agents, but if one says a trade has gone through, would you just take its word for it?

One angle worth watching with Primus Labs is its effort to make Agents’ execution results verifiable.

For example, you ask an Agent to buy an asset. In the end, it gives you a text description or a screenshot. That may show what it claims to have done, but it may not be enough to verify what actually happened. And when execution takes place in off-chain systems such as exchanges, on-chain records alone aren’t enough either.

Primus’s zkTLS approach generates verifiable credentials for data from websites or APIs. This lets applications check where the data came from and whether it meets agreed conditions, while minimizing unnecessary disclosure. The team has listed verifiable AI Agent execution as an application scenario.

More concretely, Primus has publicly released tools such as Binance Verifiable Skills, which bring some Binance API calls into a zkTLS verification workflow. The repository also includes integration guides for credential generation, local verification, and on-chain verification.

This points to a more practical way to use Agents: after carrying out a task, they return the result along with a credential that can be independently checked. Customers don’t have to rely solely on the Agent’s own explanation; other programs can verify it too.

Of course, the limits should be clear: proving what an API returned doesn’t prove that a trading strategy was sound, nor does it guarantee a profit. You still need to check which account the credential refers to, when it was issued, which fields it covers, and whether it covers the entire task.

To see whether these tools can be adopted in real-world business, watch how costly verification is, how easily developers can integrate them, and whether errors and failures are recorded.

As Agents begin handling money and business tasks on people’s behalf, the ability to verify what they’ve done will directly affect how much authority people are willing to give them.

This article is for analysis only and does not constitute investment advice. DYOR

#Primus #zkTLS #AIAgent

Primus Labs has been getting a lot of buzz on X over the past couple of days, but after looking around, I’d say about 80% of it is probably being pumped by Kaito farming accounts.

"Zktls primus good"
"primus shaping the future of zktls"
"fhe and zktls is huge for primus"

Replies like these get single-digit views, some only 3 to 6. Vietnamese long-form posts even end with "Bắt đầu tại @KaitoAI," and people conveniently add the Primus Points rewards link too. That’s not community consensus; that’s the rewards quest zone going live.

To be fair, the technology itself isn’t nonsense. zkTLS is used to prove off-chain data, like a balance on a certain website, without exposing the details. In the summaries on X, they say they’ve already verified more than $59 million in reserves for projects like Unitas, and there’s also a confidential vault running on BNB Chain.

But there are two things not to get misled about:

First, zkTLS proves that "a certain webpage showed this number at a certain time." A reserve proof only verifies the asset side; who verifies the liabilities? FTX was never short on screenshots of assets.

Second, when it’s tied to Kaito as a "reputation tool," it’s basically just adding another layer of verification to a yap leaderboard. You’re verifying how many posts I made, not whether what I said is right.

Infrastructure can be watched over time, but all those good, huge, and we’re early posts on the timeline are just check-ins for an airdrop. They have nothing to do with whether the project is actually valuable. Anyone treating those replies as a signal will most likely end up being the one left holding the bag for the points crowd.

#zkTLS #PrimusLabs

🔐 A new approach to on-chain privacy: hide the amounts, not the truth? Primus Labs raises $6.5 million

An old problem with public blockchains: a fund’s positions and moves are all out in the open. A few transfers can be enough for others to reverse-engineer its strategy. But hide everything, and counterparties can’t verify that anything is real.

Primus Labs is aiming for a middle ground: “confidentiality,” not “anonymity.”

🔹 Funding: $6.5 million combined across Pre-Seed and Seed rounds, led by Dispersion Capital, Symbolic Capital, and VanEck, with participation from Samsung Next, Alchemy, and others
🔹 zkTLS: Proves that data from a website or API is genuine (such as an exchange account balance) without revealing the exact figures
🔹 Confidential Vault: Now live on BNB Chain. Using FHE (fully homomorphic encryption), users can convert U into encrypted eU to make confidential transfers or earn yield, while balances remain encrypted during use. The launch announcement cited an APY of around 9.69%, plus a $10,000 joint incentive with Unitas Labs
🔹 Important distinction: The FHE feature is live; zkFHE is still at the research-paper stage
🔹 Also built ZKredit with BNB Chain and Brevis, aiming to make DeFi lending depend on more than just collateral

My take: One major reason institutions hesitate to deploy large amounts of capital is that “full transparency = running around naked.” This is the right direction; both proof of reserves and private transactions can benefit from it. But keep two things in mind: First, the community has also pointed out that proof speed and integration difficulty remain hurdles. Second, many recent posts about it on X are campaign-task posts; hype doesn’t equal usage. The APY figure reflects the rate at launch, so check the current yield and smart contract risks before depositing.

Would you swap your stablecoins for encrypted versions to hide your balance? Or do you think on-chain transparency is what makes you feel secure? Let’s discuss in the comments 👇

#隐私计算 #BNBChain #zkTLS #PrimusLabs