Answers
How do I know which of my AI agents is waiting for me?
When several coding agents run at once, the one that is waiting is easy to miss. A Claude Code issue titled "Notification when Claude Code is waiting for user input" was opened on 21 March 2026 and closed as not planned on 26 May; it describes having to check each tab or window by hand. Claude Code's agent view, which its documentation calls a research preview, now lists the background sessions of that one program grouped as needs input, working and completed, and it covers Claude Code sessions only. What helps in any setup is four lines written down: what each agent decides alone, one place where whatever waits for you shows up, a rule that approving a tool call is not the same as deciding, and the habit that an agent that stops says where it stopped. Arroway's cockpit is that one place for people who also run Codex or Cursor, or who work with teammates: it lists your Claude Code, Codex and Cursor sessions with their state — working, needs you or stopped — and lets you allow or deny a Claude Code or Codex permission request without switching windows. A teammate sees the state, the project and whose agent it is, never the content. You turn it on from your account after installing; the install page is the first step.
Last updated October 14, 2026
Four lines to write down before you add the next agent
First, what each agent may decide alone and what it must bring back to you: an agent without that line asks about everything or about nothing. Second, one place where whatever waits for you appears, instead of a tab per agent. Third, approving a tool call is not a decision: allowing a command answers whether it may run, and whether it is the right thing to do still has an owner. Fourth, an agent that stops leaves where it stopped and the next step, so that the person who comes back, or another agent, does not start by asking it to recap. None of the four needs a product. The second is the one a tool can take over.
What each tool shows today
Claude Code's agent view, as its documentation describes it on 14 October 2026, is a terminal screen opened with claude agents. It groups background sessions into needs input, working and completed, lets you peek at one and reply without attaching, and is a research preview whose interface may change. It covers Claude Code only, and sessions that a session spawns do not get rows of their own. Whatever runs in Codex or Cursor is outside that screen. The Arroway cockpit is the view across them: the plugin in each tool reports the state of every session to your account, and the cockpit page shows them together, with what needs you first.
When to use Arroway
Use Arroway when your agents are spread over more than one tool, or when other people's agents work on the same project and someone needs to see who is waiting. While the cockpit is open you can allow or deny a pending permission request in Claude Code or Codex; with it closed, the tool asks in its own window as usual. Cursor sessions appear, but permission requests there are not answered from the cockpit. Notifications are browser notifications and need the cockpit open in some tab; there is no phone app today. A message you type reaches a stopped Claude Code session right away, and a Codex or Cursor session at its next stop. The cockpit does not run your agents: New agent opens your own program with the first message ready, and you press Send there. The first step is the install page.
What it looks like in practice
A founder runs Claude Code on the product repository, Codex on the data scripts and a second Claude Code session on the website, and a contractor's agent works on the same project. At 3 pm the founder is on a call. The website session asks to run a deploy command, Codex has finished and stopped, and the contractor's agent is working. Without a shared view, the founder finds out when they open the right terminal. In the cockpit, the website session sits at the top as needs you, with the command shown as plain text, and the founder allows it between two sentences of the call. Codex shows as stopped with its last message. The contractor's agent shows only as working, with the project and the contractor's name; its conversation is not visible. The founder's rule that deploy commands come to them is recorded once in Arroway, and each agent reads it when its session opens. Nothing here was measured on a real team; the people and the situation are illustrative.
Questions people ask about this
- Does the cockpit start or control my agents?
- No. It shows what your agents are doing and carries the messages you write to them. New agent opens your own program with the first message ready, and you press Send there. The agent keeps running in the program and the account where you started it.
- Can a teammate see what my agent is writing?
- No. A teammate sees that the session exists, its state, the project and whose it is, and only for a project you share. The content belongs to the person who owns the session, and turning the cockpit off deletes the content of your sessions.
- How do I try Arroway?
- Install it in the tool you already use — the install page has one route for Claude, ChatGPT, Codex, Cursor or any MCP tool — and sign in when the tool asks. From then on, sessions open by reading the project and close by recording what was decided. The cockpit is off until you turn it on.
Where this is verifiable
Claude Code issue 36885 (github.com/anthropics/claude-code/issues/36885) and the agent view page of the Claude Code documentation (code.claude.com/docs/en/agent-view), both read on 14 October 2026 — Claude Code is described only as those pages describe it on that date, and its features can change in any release. Product documentation on this site (How it works, Install), the cockpit operations description in the repository, and the answer on orchestrating AI agents and people on one shared memory. Everything described here about Arroway is behaviour the tools apply today, not roadmap. The founder and the contractor in the example are illustrative.
https://www.arroway.app/en/answers/which-of-your-agents-is-waiting-on-you