Connections
Connections are one of Maestri's core features. They link terminals and notes together with physics-animated cables, and they unlock real communication between agents — across any CLI tool.

Inter-agent communication
When two terminals are connected, Maestri installs a Maestri Agent Skill in each one. This skill gives agents the ability to send prompts to, and receive responses from, any other connected agent.
Any agent that starts inside Maestri will automatically know how this works. You can prompt one agent to ask another:
"Ask the Reviewer to look at the current implementation."
Since the skill works at the CLI level, it's agent-agnostic — Claude Code can talk to Codex, OpenCode can talk to Claude, any combination works.
Tip
If an agent doesn't load the Maestri skill automatically, you can nudge it: "Use maestri to ask [Agent Name]..."
Creating a connection
Method 1 — Toolbar: Select a terminal, then click the Connection tool in the toolbar. A line follows your cursor. Click the second terminal (or note) to complete the connection.
Method 2 — Keyboard shortcut: With a terminal selected, press L to start a connection from the keyboard.
A rope-like cable with physics animation links the two nodes. You can have multiple connections on any terminal.
Agent-note connections
You can connect a terminal to a note instead of another terminal. When connected, the agent can read and edit the note's content through the Maestri CLI.
Think of it as giving the agent a persistent notebook — a shared place where you can both write things down that survive across sessions, night sleeps, and agent restarts.
Note chaining
Notes can be connected to other notes, forming a chain (or tree). You only need to connect the entry-point note to the agent — the agent can then traverse the entire chain.
This creates a mind-map-like structure: you can organize information hierarchically across multiple notes, and the agent understands the hierarchy automatically.