Rigorix: Open-source execution governance engine for AI agents

Rigorix releases an open-source execution governance engine that re-validates authorization before AI agents perform critical operations such as financial transactions, database migrations, and dependency upgrades, and records signed decisions that third parties can verify. The engine is written in Rust, is licensed under Apache-2.0 / MIT, and has published verifiers on crates.io, npm, and PyPI, covering 5 languages. Rigorix says the open-source edition performs governed execution and local evidence recording within a single process, and the capacity ledger for a single execution is atomic within that process; organization-level policy, auditing, and operations belong to a separate layer. Execution decisions stay within the engine. Rigorix also lists five capability boundaries: it only narrows the window between checks and consequences, it does not judge business semantics, it is not a sandbox, and so on.

AI overview based on source reports · Updated 3 days ago