Agents

Mooring is built around Claude Code, and can also start any command you set up as an agent.

An agent tab runs one of two things:

Claude Code, which Mooring is built around.
A custom command, any program and arguments you set up. Mooring starts it in a tab the same way, without the Claude features below.

Agent setups

Each agent setup has its own button in the project strip. Mooring starts you with three Claude setups:

Point at a button to see what it starts.

To add or change setups, open Settings ▸ AI Agents. Each setup has:

A Name and an optional Description.
A Type: Claude Code or Custom command.
For Claude Code: a Model, a Skip permission prompts switch, and Extra flags.
For a custom command: the Command and its Arguments.
A Mooring MCP tools switch, which lets that agent use Mooring's tools.

Click Set Default to choose the setup that ⇧⌘T opens.

Mooring's tools. Mooring has a built-in MCP server, so Claude in a tab can list and manage sessions, read git status, manage worktrees, and read and write cards in Up Next. See Claude Code agents.
Knowing when Claude finishes. Mooring passes Claude Code hooks to every Claude tab it starts, so it knows when Claude finishes a turn or is waiting for you. Nothing is written into your project. The bell in the project strip decides whether that project's tabs light up. See Status badges.
Restart in place. Right-click a Claude tab and choose Restart Session. Claude stops and resumes the same conversation in the same tab.
The Conversation window. Read the selected tab's conversation, formatted, beside the terminal. See Conversation window.

A custom command tab is a plain terminal. It gets Mooring's tools only if the command accepts Claude's --mcp-config option.

You can run as many agent tabs as you like, in one project or across its worktrees. Mooring tracks each one on its own.