What problem does it solve? Query responses on the Internet Computer come from a single replica and are not verified by consensus, so a malicious or faulty replica could return fabricated data. This Skill shows how to certify canister data so query responses carry cryptographic proof of authenticity while staying fast. ## Core Features & Use Cases - Merkle-tree certification: Build an RbTree or CertTree over canister data, certify the 32-byte root hash during update calls, and generate per-key witnesses for query responses. - Rust, Motoko, and frontend coverage: Complete implementations using ic-certified-map, the ic-certification mops package, and client-side verification with @dfinity/certificate-verification. - HTTP certification: Certify custom HTTP canister responses with ic-http-certification so the HTTP gateway can verify them. - Use Case: A canister stores user records that a frontend displays. Use this Skill to return each record with a certificate and witness so the browser can verify the data against the IC root key and reject tampered responses. ## Quick Start Ask the AI to implement a certified key-value store on the Internet Computer where update calls set the certified root hash and query calls return the value with its certificate and Merkle witness.