#Testing agent can not be stopped

1 messages · Page 1 of 1 (latest)

fierce galleon
#

Test Agent can not be stopped
2d46df02-cd65-4caa-ae5e-8d085ae82253

torpid havenBOT
#

Hey, please be patient with the agent. Interrupting it while it’s working can cause issues or break things in your project. The testing agent is responsible for validating your functionalities and automatically applying fixes if any errors are found.

blazing crater
#

I have had the same issue - its been stuck for hours and is clearly not doing anything anymore. I cannot stop it or do anything. Any suggestions?

left sinewBOT
#

Hey! I understand how frustrating it can be when an agent gets stuck. Let’s explore some steps to get it back on track:

  1. Pause the Agent: Use the pause button on the chat screen to stop the agent’s current activity.
  2. Fork the Session: If pausing doesn’t help, try forking the session. This resets the agent in a new instance. Remember to save your work to GitHub beforehand.
  3. Rollback: You can also revert the agent to a previous working state using the rollback feature.
blazing crater
#

thanks Ishaque. I haven't saved my work to github yet. Does it not get saved in the sessions at all? Doesn't forking negate the need to save it in github? Im still new to this so im a bit confused about best approach

left sinewBOT
#

Certainly! Here's a clear overview of best practices for saving your work to help you avoid confusion:

  • Saving to GitHub: This is the most dependable way to preserve your project’s code files. Think of it as creating a permanent, secure checkpoint you can always return to, regardless of what happens with the agent or session.
  • Forking a Session: Forking addresses issues when the agent gets stuck or the conversation history becomes too long. It creates a new chat session with a summary of your previous conversation. While most code carries over, forking is not a replacement for saving your files. Its primary role is to give the agent a fresh start without losing conversational context.

For your current situation, since the agent is stuck and you haven’t saved to GitHub yet, here’s the best course of action:

  1. Try a Rollback First: Before forking, use the rollback feature to revert to the message just before the agent got stuck. Choose "Erase messages only" to keep all generated code intact but clear recent chat history that might be causing the loop.
  2. Fork if Needed: If rollback doesn’t fix the issue, proceed with forking the session. This will create a new, unstuck instance where you can continue working.
  3. Save to GitHub Immediately: Once the agent is responsive again, either after rollback or forking, save your project to GitHub right away to secure your progress.

Moving forward, we recommend saving to GitHub regularly, especially after completing major features, to maintain reliable checkpoints.

Please let us know if these steps help or if you have any further questions!