Primus Labs has added 6 new tasks, bringing the total to 25. Remember to complete them to boost your future airdrop weight. Task: https://s.kaito.ai/xKLA1Sx I saw a scenario the other day that suddenly made me realize the problem Primus is trying to solve is quite specific. Someone wants to use off-chain income or assets to borrow money on-chain, but doesn't want to reveal their bank statements or specific balances. Or an agent needs to prove they checked prices and made a certain call according to the rules, but can't disclose all the sensitive data in the process. Previously, in both cases, you either completely trusted a third party or sacrificed privacy. @primus_labs' approach separates proving authenticity from not exposing the original content. zkTLS can prove that the data returned by a webpage or API is authentic without exposing the original content; FHE performs calculations directly on encrypted data, and sensitive information such as balances and amounts doesn't need to be decrypted throughout the process. Ultimately, you can verify a claim's validity, but you can't see the intermediate details. Trusted Credit Infrastructure is one application of this approach. You can transform off-chain asset, income, or behavioral data into privacy-preserving on-chain proofs of credit. Others can verify that you meet certain criteria, without needing to know your specific wealth or its origin. For institutional risk control and compliant lending, this granular control is more practical than complete concealment or complete transparency. The AI Agent is similar. When executing tasks, the agent can use zkTLS to prove that it indeed called a certain API and performed a certain step according to the policy, without disclosing sensitive data during the process. It verifies that the behavior conforms to the rules, rather than laying out all intermediate results. What I'm particularly interested in is that it doesn't pursue complete anonymity, but rather strikes a balance between confidentiality when necessary and the validity of claims. You need to prove you have money, qualifications, and reserves so others can verify the validity of the claim; but they don't need to know your specific amount, its origin, or how you plan to use it. Of course, it's still in its early stages, and the actual implementation and performance remain to be seen. But in terms of direction, it is indeed answering an increasingly real question: when financial activities require both privacy and verifiability, how can we simultaneously achieve both? If you're interested, you can check out their Trusted Credit or Agent verification content to get a feel for this confidential yet verifiable approach.