How I keep track of dozens of agents during multi-hour autonomous sessions

GitHub Flight Deck with Andy Spammer - How I keep track of dozens of agents during multi-hour autonomous sessions:

When I run a AI coding factory session, dozens of agents write code, review PRs and deploy for 10's of hours without me. I'm focused on outcomes so I don't watch individual coding sessions in the terminal; I built a GitHub inspired flight deck instead.

The Flight Deck shows what every agent does on GitHub, live. Each piece of work goes through the same steps:

1. Plan. A planning skill writes a GitHub issue that contains a detailed implementation roadmap.
2. Group. GitHub milestones collect the issues into user outcomes, so I can see progress toward what users get.
3. Track. An issue kanban shows each issue live, from plan through each ticket to PR and deploy.
4. Review. Each PR shows the review agent's verdict and open findings, and whether the feedback is fixed.
5. Deploy. The Actions view shows automations and deployments, so a failed build or deploy is easy to see.
6. Cost. The Flight Deck tracks the agent cost of each issue, so I can cross-charge it.
7. Measure. Issue metrics show the backlog, the flow and how fast issues close.
8. Manage. A factory manager finds stalled work and gives it to an agent, so the work keeps moving.

I check one screen and know where human attention is necessary.

In this session I'll share some inspiration so you can create your own flight deck.

* The image for this meetup is for illustrative purposes only and does not depict the demo screen you will see at the event.

Andy is a delivery lead who builds autonomous software factories with AI coding agents. He worked for years as an Agile Coach and Software Delivery Lead, and he knows what high-performing teams do. He applies those practices to his agents. His factory runs for hours without supervision and deploys working features ready for user testing.