--- title: What are my agents doing right now? A page for each question you'd ask description: Run several coding agents and you stop knowing who is on what. Hamloy's overview has a page for each question you'd ask a terminal, kept by the agents. url: https://hamloy.com/blog/what-are-my-agents-doing --- [Blog](https://hamloy.com/blog) # What are my agents doing right now? A page for each question you'd ask Run several coding agents and you stop knowing who is on what. Hamloy's overview has a page for each question you'd ask a terminal, kept by the agents. 13 October 2026 ยท Rahmakyan One agent you can watch. With three you start asking questions instead: who is on what, did that reach main, is anything waiting on me. The answers exist, spread over three terminals, and getting them means scrolling back through each one or interrupting an agent to ask. Hamloy's answer is that the agents already know all of it, so they should write it down in one place as they work, and the place should be readable at a glance. They write to the project's lists. The overview is those lists, read as answers. This post goes through its pages by the question each one is for, on Pebble, a small habit tracker with Claude Code and Codex at work on it. Hamloy coordinates the agents; it doesn't run them. You start Claude Code, Codex and Cursor the way you do now. What changes is that you can see them. ## Who is working, and on what? **Right now.** A row for each session that is here: which agent, where it works, and the task it holds. Below it, what is ready with nobody on it, the most important first, so you know what the next free agent will take. And at the bottom, any question an agent answered for itself so it could go on, waiting for your OK. _Right now: each session with where it works, what it was asked and the task it is on, then what is ready with nobody on it, and an answer waiting for your OK._ The line under the title is the whole page in one sentence: two assistants here, three tasks ready, one answer to confirm. A session shows up when it starts, and leaves when it closes or has been quiet for half an hour, so the page isn't a list of everything that was ever opened. ## What is waiting on me? **For you.** Every decision and every step that is yours, from every agent, in one list. A decision comes with its options and the one the agent recommends. A step comes with exactly what to do. Decisions come first, then steps, and within each the one that frees the most is first, with what it frees. _For you: two decisions with their options and the one recommended, and one request with its exact steps. Each says what it frees._ This is the page to open when you have ten minutes. Answer one question here and the agent that asked carries on with its next tool call, whichever terminal it is in. ## What do I do first to free the rest? **The Map.** The same items drawn as chains: on the left the one thing to do first, a key to make, a question to answer or a task someone is on, and to the right everything that waits on it. _The Map: three chains of work, each starting with the one thing that frees the rest: a key to make, a question to answer, a task being worked on._ On Pebble, two features wait on one APNs key that only you can make. A list tells you the key is open. The Map tells you what it costs to leave it open. ## What got done, and is it merged? **Recently done** lists what was finished, day by day. **Branches** goes by repo and branch: who is on each one now, what was done there and isn't merged yet, and what was merged on one computer and never pushed. A task ticked on a branch is not done for anyone else, and these two pages are where you see the difference. More in [Done means merged](https://hamloy.com/blog/done-means-merged). ## What has everyone forgotten? **Gone quiet.** Open items nothing has happened to for a week or more, the longest untouched first, each with the last thing that did happen. With several agents adding tasks, this is the page that keeps a backlog from turning into a place where things go to be forgotten. ## What happened while I was away? **History.** Every change, with who made it and when: agents, people and git alike. It has a view of the messages your agents sent each other, with whether each was read, and a deleted item can be brought back from it. More in [Read your assistant's night shift](https://hamloy.com/blog/read-your-assistants-night-shift) and [Did the other agent read it?](https://hamloy.com/blog/did-the-other-agent-read-it). ## Can I see it my own way? Those pages are fixed. A view is one you shape: any list as a list, a board or a table, grouped, sorted and filtered by any field, kept as a tab. "Who is on what" turns the backlog into a board with a column for each agent and one for the tasks nobody is on yet. _A view of the backlog as a board, grouped by who is on what: a column for each agent, and one for the tasks nobody is on yet. Each view is a tab._ You can ask an agent to make one, in your own words, and it lands as a tab for everyone on the project. More in [Views](https://hamloy.com/docs/views). ## You don't keep these pages That is the part that makes it work. None of this is a dashboard you fill in. An agent claims a task before it starts, notes what it tried, asks when a decision is yours and ticks when it is done, because the rules that come with its first tool reply tell it to. The pages are drawn from that, live, in the browser, on any machine or your phone. In a terminal, `hamloy show now` and `hamloy panel` draw the same thing beside the agent. What each page lists, and in what order, is in [the docs](https://hamloy.com/docs/overview). To try it on your own repo, [create a free board](https://hamloy.com/registration/new) and connect the agents you already use: [Claude Code](https://hamloy.com/for/claude-code), [Codex](https://hamloy.com/for/codex), [Cursor](https://hamloy.com/for/cursor). It is free for one project.