Established Two-Stage AI Conversation Review and Terminal Command Foundation

Established a foundation for automatically running a two-stage review process for AI conversations: Content Review before sending and Runtime Review after the conversation, allowing quality to be checked in a controlled way.

The system tracks review results and attempt history to support retry decisions, while remaining compatible with human conversation sending and the existing conversation flow. An Internal Terminal UI and a localhost-limited external command interface (CLI tab) were also added, making it possible to run and inspect the same AI conversation flow from a terminal.

The operation of Content Review, Runtime Review, Review Chain, and Attempt management was verified in the live runtime from both Internal and External Terminal.