Describe the bug
When I ask an agent to rubber-duck or if a subagent decides to rubber-duck, I without fail always get 4 rubber-duck instances. They all launch at separate times but consistently show up eventually. The all arrive at same conclusions so they don't add any extra value. I asked copilot why it launched four and it said it only called the tool once and there is no way to specify how many instances launched. It thought that perhaps each one was too slow to start responding so the code tried to launch another as a backup and so on until there were four.
Affected version
1.0.22
Steps to reproduce the behavior
Ask an AI to rubber-duck something. Watch as four background tasks slowly accumulate.
Expected behavior
Only one rubber duck instance should launch at a time/per-request.
Additional context
- Windows Terminal on Windows
Describe the bug
When I ask an agent to rubber-duck or if a subagent decides to rubber-duck, I without fail always get 4 rubber-duck instances. They all launch at separate times but consistently show up eventually. The all arrive at same conclusions so they don't add any extra value. I asked copilot why it launched four and it said it only called the tool once and there is no way to specify how many instances launched. It thought that perhaps each one was too slow to start responding so the code tried to launch another as a backup and so on until there were four.
Affected version
1.0.22
Steps to reproduce the behavior
Ask an AI to rubber-duck something. Watch as four background tasks slowly accumulate.
Expected behavior
Only one rubber duck instance should launch at a time/per-request.
Additional context