插件

子智能体矩阵蜂群：把异质任务路由到最合适的模型（OpenRouter 类网关 + cfgpu.com/llm/square），通过进程内子智能体或直接 LLM 调用下放，并按模型统计 token 消耗、用真实反馈驱动排名。

Two-tier model router for DeepSeek Harness — LLM-Judge routing, multi-model fallback chains, exponential-backoff failover, and task-level orchestration (DSH adaptation of pi-shift-router)

DeepSee: DeepSeek Harness vision, multi-model routing, and Gemini visual self-checks before delivery

DeepSeek Harness 可复用 Markdown Prompt Profile，支持单轮模型选择、参数替换和状态恢复。

Smart provider routing: composer-button one-click enable/disable, any channel error (including 4xx) auto-falls-back to the next channel, chain names shown in the model selector without exposing internal channels/models, multi-chain management with settings UI.

Reasoning-mode routing plugin for V4 Flash on headless: w7 persona anchoring, first-turn tool anchoring and complexity-dispatched guidance.

Google Antigravity / Cloud Code Assist model provider for DSH with native Web OAuth, real-time quota tracking, and dynamic reasoning effort routing.

Unified static-image generation router for DeepSeek Harness: one image_generate tool with standard parameters, config-driven OpenAI-compatible engine routing (OpenAI Images, Zhipu CogView, and any compatible endpoint) with health-aware fallback, durable w

Adaptive task and risk routing, loop budgets, and progressive tool-result pruning for DeepSeek Harness

Native DeepSeek Harness tools for an Obsidian wiki vault: wiki_query, wiki_write, and wiki_lint implement the mechanical core (path routing, delta tracking, index/log bookkeeping, health checks) of the wiki skill suite.

基于飞书官方 WebSocket 长连接的 DSH 双向聊天桥：消息路由到所选工作区的最新会话并自动回复，可在设置页切换机器人凭证与工作区。

Provider discovery, matching, health checks, and pre-output failover for DeepSeek Harness.