Plugins

Control por ronda del nivel de pensamiento (reasoning_effort) para DeepSeek Harness: Auto programa low/high/max según el historial reciente de tool calls, o permite fijar manualmente un wire level (off/on/minimal/low/medium/high/xhigh/max), con mapeo wire personalizado, guard de capacidades por modelo y presets de ventana de contexto (64K–1M); su capability card se combina con [dsh-llm-openai-completions](https://github.com/drscrewdriver/dsh-llm-openai-completions) (instalable desde dsh-market) para probar reasoning en gateways compatibles con OpenAI (vLLM / LM Studio / autoalojados). La card escribe capacidades llm-pi-ai, el adapter conduce el wire (compat.thinkingFormat) y el plugin mantiene la lista de takeover.

Adds three visual pages to DSH Settings: MCP servers (managed by rewriting the profile patch managed block from a JSON source of truth, with YAML validation and timestamped backups), skills (browse, create, edit and delete across DSH global, shared .agents and Claude / Codex directories) and the global AGENTS.md persona (with one-click restore of the last three backups).