About
This skill provides a robust Python wrapper for interacting with local Ollama instances, specifically optimized for the Phi-4 model with a 16K context window. It streamlines complex LLM tasks such as clinical section detection, summarization, and recommendation generation while offering built-in reliability features like timeout handling, automatic retries, and structured logging. Designed for precision-heavy workflows, it ensures deterministic outputs through configurable sampling parameters and includes health check capabilities to verify server availability before executing critical inference tasks.
