Multi-agent orchestration and LLM model routing for DeepSeek Harness (DSH): semantic AI agent profiles, exact model routes, declarative teams and strategies, and bounded subagent delegation - a TypeScript plugin that adds no second runtime.
DSH（DeepSeek Harness）实时重复输出守卫：模型复读时立即停止生成（默认同一字符串连续重复 ≥10 次；阈值、单元长度、检测窗口、白名单等全部参数可在设置页动态调整）。已生成内容正常提交，不丢消息、不污染会话日志。Real-time repetition guard plugin for DSH, with a graphical settings page and npm one-command install.
DeepSeek Harness (DSH) plugin: live task completion time, token & per-model cost estimator in the session header, with LLM-powered PERT estimates. / DSH 插件：会话标题栏实时任务耗时、token 与花费估算。
LLM Wiki workbench for dsh-better-sidebar: browse and lint a Karpathy-style interlinked markdown wiki, open pages with the native markdown viewer, clickable [[wikilinks]] and provenance/relative .md links, and backlinks. Ships the llm-wiki agent skill.
DeepSeek Harness agent preset for adversarial review: presumes the artifact is wrong, demands a citable authority and file:line for every finding, refuses to edit tests to pass, and asks before changing anything.DeepSeek Harness「对抗式审查」Agent 预设：预设被审查对象是错的，每条结论必须给出权威依据与文件行号；严禁改测试让其通过，修改任何文件前先征求同意。
This plugin was developed using DeepSeek Harness and the DeepSeek V4 Flash 0731 / Pro 0813 models, with multiple rounds of refinement. It enhances the existing sub‑agent functionality and adds several other features.
DSH (DeepSeek Harness) plugin: delegate read-only collection tasks to a local Ollama model — zero cloud token cost, zero API key. 把只读采集任务交给本机 Ollama 模型：含设置面板 + 选型尺 + AI 可读说明书。
