Enables AI agents to add decision drafts, evidence, and counterarguments to a shared local decision state, while users confirm or reopen decisions in a web console. Prevents unverified agent answers from being silently turned into code.
View & run on Rokha →