llama-cpp

为DSH接入本地大模型能力：在「设置→插件」页一键启停本地 llama.cpp 大模型（双槽x双模态x双预设），卡片内配置、一条命令安装、自动注册，装完即用| Enable local large model capabilities for DSH: One-click start/stop for local llama.cpp models (dual‑slot × dual‑modal × dual‑preset) right in Settings → Plugins; configure within the card, install with a single command, automatically register, and ready to use

Local OpenAI-compatible model-server discovery + tiered reasoning-metadata resolution for DeepSeek Harness (dsh).

DeepSeek Harness plugin: run subagents on a local llama.cpp model via the OpenAI Responses (/v1/responses) protocol — LLM adapter + delegation skill

Manage local LLM inference servers in DSH: GPU registry, parameter profiles, VRAM validation, and tok/s benchmarks for llama.cpp, SGLang, and vLLM

DeepSeek Harness plugin: OpenAI-compat server protocol card (loopback probe, no GGUF)

DeepSeek Harness LLM adapter for a local llama.cpp OpenAI-compatible server.