DSH Effort Tweak: Add per-model inference effort settings to DSH Web UI

DSH Effort Tweak is an MIT-licensed DSH plugin maintained by Toukaiteio, designed to address the cumbersome task of configuring model reasoning effort levels within the DeepSeek Harness custom provider. This plugin introduces a new "Settings → Reasoning effort" page in the DSH Web UI, allowing users to intuitively select various reasoning levels—from `off` to `max`—for individual configured models, without the need to manually edit the `llm-pi-ai` configuration file. Configuration changes are saved via the DSH settings service, preserving the model's other fields and provider-specific values. Regarding installation, it supports Windows PowerShell.