Rosetta Protocol 3.0.0 - Core 1.0 Standards Candidate

Rosetta Protocol is an open, protocol-independent safety control and evidence standard for autonomous AI agents.

Rosetta Core binds structured intent, independently observed action requests, exact request-bound authorization, deterministic policy decisions, pre-execution enforcement, and observed effect receipts into verifiable safety transactions.

Version 3.0.0 includes Rosetta Core, Rosetta Guard, the Reflexon-derived Rosetta Reflex adaptive-control layer, MCP and A2A bindings, normative JSON Schemas, language-independent conformance vectors, tests, experiments, and reproducibility tooling.

Rosetta complements MCP and A2A rather than replacing them. MCP defines how agents invoke tools, A2A defines how agents communicate and delegate, and Rosetta defines how consequential actions are authorized, mediated, constrained, observed, revoked, and audited.

Status: Rosetta Core 1.0 is a research standards candidate and reference implementation. It is not production certification, a universal safety guarantee, or proof that every covert communication channel can be detected.

Repository:
https://github.com/rosettaprotocol/rosetta-protocol

Tagged release:
https://github.com/rosettaprotocol/rosetta-protocol/releases/tag/v3.0.0