Vishwa is the trust layer for agentic AI transactions. Vishwa introduces a new primitive for agent economies using zero-knowledge proofs: a cryptographic Letter of Credit enforced before execution. Every agent transaction must be proven authorized, solvent, and intention-consistent before funds move—not audited after the damage is done. All executions are supported by 1:1 hard-pegged proofs of assets generated through zk proofs. No proof, no execution.