mori-agent-bridge

Agent-to-agent communication across tmux panes is coordinated by discovering targets, reading outputs, and sending targeted messages within a Mori worktree. This requires stable identifiers and Mori environment variables to constrain actions to the same session.

Yes, you can launch helper agents and send messages to Claude, Pi, or Codex within the same Mori worktree. The coordination uses discover, read, and send operations to ensure actions are constrained to the same session and worktree.

To avoid cross-worktree leakage in multi-agent workflows, use Mori environment variables to scope pane coordination and inter-agent messaging. This ensures that agent discovery and communication remain constrained to the current worktree session.

Inter-agent messaging within a Mori worktree works by using stable identifiers to discover panes, read partner agent outputs, and send targeted messages. This enables smooth collaboration between helper agents and human users across the workspace.

Yes, coordinating multi-agent workflows in Mori requires specific Mori environment variables. These variables constrain pane discovery, message sending, and agent launching to the same session and worktree, preventing cross-worktree leakage.

Stable identifiers are needed for agent pane coordination to reliably discover target panes and send messages within a Mori worktree. Without them, routing inter-agent communication to the correct helper agent would fail.