After the command. Is sent. It will then pause and I'll have to press the little button control C button down in the corner for it to start thinking again, but I'm unsure did it really execute that command because I had to press that little control C button down in the corner?
#When warp agent does a command,
hidden hemlock
abstract hedgeBOT
Hi @hidden hemlock. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
Based on the available knowledge sources, when you run a command in Warp Agent Mode, the agent will either execute the command directly or, if it needs your approval, prompt you to confirm before running it. After the command is executed, if the agent is waiting (for example, for a long-running command to finish or for further input), you may see a control panel with a "Ctrl+C" button in the corner. Pressing this button sends an interrupt signal (like pressing Ctrl+C on your keyboard) to stop the running command or process.
If you have to press the "Ctrl+C" button for the agent to "start thinking again," it usually means the command was running and needed to be interrupted—perhaps because it was stuck, waiting for input, or running longer than expected. The command does execute, but if it hangs or enters a state where the agent can't proceed (such as being stuck in a pager or a long-running process), you need to manually interrupt it so the agent can regain control and continue processing.
This behavior is confirmed in several sources, which mention that the agent can get stuck on long-running or interactive commands, and that the cancel (Ctrl+C) button is provided to let you interrupt these processes. After interruption, the agent should resume its workflow or allow you to continue interacting with it. If you had to press Ctrl+C, it means the command was indeed running, but needed to be stopped for the agent to proceed. This is not an indication that the command was skipped; rather, it was executed and then interrupted by your action so the agent could continue its logic or provide further output Managing Agents, git branch -a command gets stuck.
If you notice that the agent does not proceed after pressing Ctrl+C, or if the cancel button does not work as expected, this may be a bug or a limitation with certain types of commands, as noted in user reports. In such cases, the command was still executed, but the agent may not have been able to handle the output or state transition properly.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
hidden hemlock
@ That's great, but are they working on a solution for this? A great solution would be. A maybe a time out somehow? Or a set amount of time before agent. Can see that it got hung up or stuck and then. Proceed to get it unstuck himself.
@abstract hedge That's great, but are they working on a solution for this? A great solution would be. A maybe a time out somehow? Or a set amount of time before agent. Can see that it got hung up or stuck and then. Proceed to get it unstuck himself.
abstract hedgeBOT
Let me check the knowledge base for you again. One moment! :books:
Yes, the Warp team is aware of this issue and is actively working on improving how the agent handles long-running or stuck commands. There are several user reports and official responses indicating that the current behavior—where the agent can get stuck on commands that trigger pagers or run for a long time, requiring manual intervention with the cancel (Ctrl+C) button—is not ideal.
Specifically, Warp team members have acknowledged that:
- The agent sometimes runs interactive or long-running commands that can get stuck, and the cancel button does not always successfully end these commands.
- They are working on making it possible for the agent to directly interact with these types of "long running" commands so that it doesn't get stuck in these cases.
- Improving the UI and agent logic to better handle and recover from these situations is an active area of development. For example, there is mention of auto-expanding blocks for long-running commands and improving the information given to the agent when commands are killed with SIGINT (Ctrl+C) so it can better understand and recover from these states.
While there is no explicit mention of a timeout feature or automatic detection and recovery from being "hung up," the team is focused on making the agent more robust and responsive in these scenarios. They are also considering enhancements to ensure that the cancel button reliably cancels stuck commands and that the agent can proceed without requiring manual intervention as often git branch -a command gets stuck, Chat freezing and NO more chat!.
If you have specific suggestions like a timeout or automatic un-sticking, the Warp team encourages users to submit feature requests or upvote existing ones to help prioritize improvements.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
hidden hemlock
@abstract hedge Is there a way I can get them to focus on this as their top priority because that is huge. When it comes to interactions with getting me to stay to want to pay for their service. That's so crucial it makes me want to switch to something else.
abstract hedgeBOT
Let me check the knowledge base for you again. One moment! :books:
The knowledge sources show that several users have already expressed strong frustration about this issue, with some even stating they canceled their subscription because of it, and others emphasizing how critical it is for their workflow. Warp team members have acknowledged the problem and indicated that improvements are in progress, but there is no official process mentioned for escalating a feature request to top priority beyond upvoting issues and providing feedback in GitHub threads.
If you want to increase the chances of this being prioritized, the best actions—according to the knowledge sources—are:
- Upvote and comment on relevant GitHub issues to show how important this is to you. The Warp team has stated that upvotes help them gauge user interest and prioritize features.
- Clearly describe the impact (e.g., that it’s a blocker for you and a reason you might stop using or paying for Warp) in your comments. Other users have done this, and it’s visible to the team Add setting to automatically prevent pager output in AI agent commands.
- Subscribe to the issues so you’re notified of updates and can continue to provide feedback.
There is no mention in the sources of a direct escalation or support channel that guarantees a feature will become top priority, but user feedback and upvotes are the main mechanisms the team uses to prioritize work.