Last released Jul 2, 2026
Execution governance for AI agents: preview -> canonical hash -> operator accept -> server-side re-derivation -> apply. Fail-closed, idempotent, audited.
Supported by
