DeepSeek Harness plugins — Page 365

Quiet progress narration for DeepSeek Harness 0.2, with native persistent settings and independent Reader compatibility.

User-approved Goal, Milestone, and Turn checkpoints for solo long-running DeepSeek Harness projects.

DSH 持久化项目大脑：理解代码架构、记住项目历史、跨 Session 保持上下文

DeepSeek Harness plugins for project-level persistent context: memory consolidation (MEMORY.md + CONTEXT.md), low-frequency autolearn for project skills, session archives with a mechanical index, and automatic handoff.

DSH plugin: owns the sidebar workspace browser (vendored from the official plugin) so project-grouping logic can be layered onto the real UI instead of a reimplementation.

中文 DSH 本地知识复习插件：零配置资料库、证据优先问答、Markdown 摘要与二次分类。

Project board for DeepSeek Harness: kanban card groups, cards linked to landfill-project memory, mounted repos, a pinned 项目区 workspace, and project sessions that auto-load card memory.

Per-project MCP loading for DeepSeek Harness: sessions of a project automatically get the MCP servers declared in its .dsh/mcp.json (mcpServers JSON shape, shared with Claude Code / Cursor / VS Code). Client bridge — consumes MCP servers, not an MCP server, not an official DeepSeek package.

Evidence-backed, human-approved project memory for DeepSeek Harness

Project-scoped AI memory: records pitfalls and user habits per project, injects them into future conversations

Project Context Engineering for AI agents: turn messy project context into a clear, actionable workspace.

Private, per-project system-prompt rules for DeepSeek Harness — matched by git remote URL, repo path, or cwd prefix; worktree-aware; never committed to the repository.

Local project, requirement-group, and conversation workbench for DeepSeek Harness Web and Desktop

DeepSeek Harness Web plugin for date-organized projectless sessions under Documents/DSH

DeepSeek Harness 项目组插件：在一个会话中跨多个项目文件夹工作——读写组内文件、运行命令、查看各项目 AGENTS.md；每个项目可带一句说明，AI 据此快速定位目录。项目组跟随工作区（每个工作区独立默认选择），组内文件路径默认允许读写。官方安装：dsh plugin --profile web add。

PRO-LONG-style programmatic memory for DeepSeek Harness: appends every session event to a per-workspace log.txt that the agent retrieves with grep/python. Built-in write probe, /prolong status command and permission-deny counter.

Privacy-conscious Prometheus metrics and a Grafana dashboard for DeepSeek Harness

全局提示注入 / 上下文病毒（context virus）防御插件：tools/pre-execute 扫工具参数、tools/post-execute 扫工具结果、agent/pre-step 扫进入模型前的消息并注入金丝雀守卫、llm/stream 出站消毒；支持 block / quarantine / monitor 三模式、磁盘化可演进病毒库（学习/导入/导出）与审计日志。纯 JS 无原生依赖，profile 层挂载，对所有预设与子代理全局生效。

Trae-style input booster for the dsh web GUI: a sparkle (✨) button beside the model selector optimizes the typed draft into a high-quality, faithful, directly executable prompt via the session's default model; after optimization the button becomes an undo (↩) icon that restores the original input. Zero-build pure-JS plugin: the host half serves /api/prompt-boost/optimize (llm.stream), the browser half registers the conversation.input.right slot.

DSH 提示词工作台：看清 system prompt 的每一段、写下你自己的家规、体检重复与矛盾，并记下每次改动。

给 DeepSeek Harness 用的个人提示词管理器：把「个人提示词 + 系统提示词改动」存成预设，按会话或全局挂载，并能改写、关闭、还原 DSH 的原生段落，还能看清模型实际收到的最终提示词。

Standalone dsh web plugin: an input-area "Prompt Enhance" button that improves the current draft with GitHub prompt-engineering templates and fills it back after confirmation, with its own on/off switch in Settings.

DeepSeek Harness 输入框一键增强提示词:保留原句、一键撤回、模型不可用自动降级。One-click prompt enhance with undo.

DSH web plugin: a star button beside the composer that rewrites the current draft into a clearer prompt, combining PromptForge rule diagnosis with the session's own default model.