Collapse five identical _ollama_models case arms into one, add timeout
to prevent shell hang when remote ollama host is unreachable, replace
launch integrations function with a plain variable, and remove
redundant comments.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Ollama doesn't expose a "completion" subcommand yet, so this provides
manual completions for subcommands, flags, model names (from ollama list),
running models (from ollama ps), and launch integrations.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
