Gpt-oss-120B with Vllm chat and other functions work but agent/plan mode doesn't

[fyun89]

I saw some other workarounds for Ollama and others (replacing the model name, etc) but those didn't work out for me (vllm can also be configured to do something similar).
There seem to be a tool-call issue -- the model outright don't respond when I ask it to do something like "look for file xxx". I tried playing around with Vllm configurations but I wasn't able to find an answer. Came here to see if anyone else encountered this issue and found a solution.

Thanks

[fyun89]

Solution found: (actually it was on VLLM part but leaving here for future readers)
vllm-project/vllm#22386