Describe the bug
The agent launched two background agents early in the investigation. Both took ~140 seconds to complete. By the time they returned results, the agent had already found most of the same information through direct grep/view calls. This consumed resources without meaningfully accelerating the investigation in this case
Affected version
GitHub Copilot CLI 1.0.36
Steps to reproduce the behavior
- Ask the agent to investigate a large codebase
- Agent launches multiple background agents
- Agent also starts direct investigation with grep/view
- Observe explore agents complete after the direct investigation has progressed past the agents' scope
Expected behavior
The agent should either wait for agents before duplicating work, or cancel agents once the main thread has found the same info, or better calibrate when to use agents vs direct investigation based on complexity
Additional context
No response
Describe the bug
The agent launched two background agents early in the investigation. Both took ~140 seconds to complete. By the time they returned results, the agent had already found most of the same information through direct grep/view calls. This consumed resources without meaningfully accelerating the investigation in this case
Affected version
GitHub Copilot CLI 1.0.36
Steps to reproduce the behavior
Expected behavior
The agent should either wait for agents before duplicating work, or cancel agents once the main thread has found the same info, or better calibrate when to use agents vs direct investigation based on complexity
Additional context
No response