Being more reliant on agents has made me feel like my developer experience was going downhill overall. It was clear to me that my workflow and tooling needed to evolve as well.
As I’m working on a platform with a microservice architecture and shared libraries, my day-to-day work requires me to work on multiple repositories and I ended up with either multiple tabs on my terminal, or a single tab but multiple panes to run my agents. This started to get really difficult to keep track of when the agents needed my input, needed my review, or were ready for new tasks. Keeping up with the changes made by the agents was also a pain.
Then I discovered herdr, which provides a TUI dashboard view of the agents’ statuses. I’ve been toying with this new tool for a couple of weeks now and I really liked it. With herdr integrated into my tooling now, the feedback loop between myself and my running agents has gotten a lot quicker! There’s a sound notification to notify me that my attention is needed, and the herdr TUI dashboard visually tells me exactly which agent. I typically use lazygit to browse through the changes made by the agents and it was easy to weave it into my herdr workflow too. The big plus for me is that it’s agent-agnostic, it works well with both claude and opencode which are my primary agentic tools.
There are also lots of community plugins for herdr which I have not really explored yet but overall, this tool seems really promising to me in terms of configuring your own workflow to manage multiple agents.