agentchattr syncs multiple AI coding agents

agentchattr solves the "terminal silos" problem by turning isolated AI CLIs into a collaborative team. It’s a pragmatic middle layer that respects existing tools while adding a much-needed orchestration UI. Direct terminal injection via Win32/tmux allows agents to stay in their native environments while receiving mentions from the chat, and built-in Job Tracking turns casual chat messages into managed threads, bringing order to complex, multi-file agent workflows. Autonomous loops allow agents to "summon" each other, significantly reducing human-in-the-loop friction. Local-first architecture ensures low latency and data privacy, which is critical for enterprise codebase exposure. Support for multiple instances of the same agent enables specialized roles like "Lead Architect" vs "Test Runner" within a single session.