home/categories/debugging/nvidia-openshell-agents-skills-debug-inference-skill-md
debuggingtools

debug-inference

Debug why inference.local or external inference setup is failing. Use when the user cannot reach a local model server, has provider base URL issues, sees inference verification failures, hits protocol mismatches, or needs to diagnose inference on local vs remote gateways. Trigger keywords - debug inference, inference.local, local inference, ollama, vllm, sglang, trtllm, NIM, inference failing, model server unreachable, failed to verify inference endpoint, host.openshell.internal.

NVIDIA
maintainer
NVIDIA
更新于 3/16/2026
星标
4745
分支
497
quick start

Installation and usage

Debug why inference.local or external inference setup is failing. Use when the user cannot reach a local model server, has provider base URL issues, sees inference verification failures, hits protocol mismatches, or needs to diagnose inference on local vs remote gateways. Trigger keywords - debug inference, inference.local, local inference, ollama, vllm, sglang, trtllm, NIM, inference failing, model server unreachable, failed to verify inference endpoint, host.openshell.internal.

安装
$ install --globalskills.sh
使用

安装后,您可以通过在终端运行以下命令来使用此技能:

skills use debug-inference