生産性ツール · DeepSeek Harness プラグイン — 63 ページ

Session-backed writing pad for the DeepSeek Harness web GUI with Markdown preview, complete-draft XML requests, human-readable transcript projections, and focused AI writing tools.

ZCode (Z.ai) desktop-app connector for DeepSeek Harness: an account/quota panel plus a live capability view, derived from the running GUI client over CDP — not from an API key.

DeepSeek Harness plugin: connects the local dsh instance outbound to a central relay server over a fixed HTTP + WebSocket API, streaming live work status and workspace/session state and accepting remote control operations.

DeepSeek Harness Core（DHC）人格核心进化插件：心脏/情感/身体/大脑/灵魂四件套自主进化、三层记忆、自动抓取消化、心跳体检、HTTP API、内嵌鲸系桌面宠物（启动 harness 自动拉起、双向接口），并整体编译五器官（ModLens 眼睛 / at-file 左手 / genui 右手 / automation 双脚 / better-sidebar 工作台）为单插件

GateCraft（门控工艺）：门控式数学建模 skill 套件 for DeepSeek Harness——五阶段流水线、阶段门、国奖写作范式与脚本化质检；agent 求解质检，人在阶段门思考决策。

Six-stage WeChat Official Account content-publishing workflow with quality-gate tools and a bundled skill, installable as a DeepSeek Harness (DSH) plugin

Native VS Code client for DeepSeek Harness. Connect to your existing local Harness instance and continue the same workspaces and sessions directly inside VS Code.

Web UI polish layer for DeepSeek Harness: normalizes unfinished or self-contradictory official UI, reconciles style conflicts between installed plugins, and unifies the visual language — with live-customizable chat width, markdown font size, workspace scale and UI font stack from Settings → General, all following official design tokens. Client-only dsh bundle, zero model cost, removable without residue.

Mode menu (agent preset) enhancements for ezdsh: a hub link pinned in the menu footer and tuned typography on the new-session screen.

Stateless dialogue compression that mimics human memory — forgetting-curve bounded context for LLM conversations. DeepSeek Harness (DSH) plugin backend included.

Stateless dialogue compression that mimics human memory — forgetting-curve bounded context for LLM conversations. DeepSeek Harness (DSH) plugin backend included.

项目文档体系初始化 skill（DeepSeek Harness 插件包 / Reasonix 技能）：提问驱动落实文档规范 + 固化 AI 协作工作流，以「存量完善」为核心场景，跨平台可用（DSH / Reasonix / ZCode / Claude Code / Codex）。

会话额度监控（官方 dsh bundle）：按 DeepSeek 官方价格记账、额度耗尽拦截 + 输入框上方进度条 UI

A structured multi-panelist review workflow for the DeepSeek Harness: N panelists score in isolated subagents blind to each other, a chair reconciles disagreement, an independent critic audits the process.

SpecPowers for the DeepSeek Harness and Claude Code: a spec-driven + test-driven engineering skill group fusing OpenSpec and Superpowers

算账 (suanzhang) — DeepSeek Harness 计费插件：侧边栏余额/今日消费、算账页签按步骤费用（交易终端风格）、官方价自动同步、跨会话跨天汇总。

微信 iLink ClawBot 远程桥接插件（DeepSeek Harness）—— 官方 iLink 协议直连，带 Web 设置界面（扫码登录/模型/预设/权限/会话管理）与 24h 自动续期

DSH bundle that packages the Tabbit Browser skill and background installer

DSH 右侧仪表盘 UI 插件：上下文占用（三色圆环）、Token 用量与人民币费用估算、会话统计、目标、待办、后台任务、工作区；支持右侧栏（details 槽位）与会话标签页（conversation.view）两种形态。

DSH context-pressure guard: when a session's next request would carry a dangerously large prompt, tell the user at turn end to write the handoff and start a new session. Consumes the official token-meter contextPressure projection; injects no prompt, registers no model-facing tool, so its context tax is zero. v0.3: at 45% occupancy the host half auto-drafts a machine-facts handover block into the session working directory; v0.2 added the host state route and an opt-in (default off) WeChat push; v0.3.2 onwards ships bilingual (zh/en, auto-detected per half) runtime copy; v0.3.3 was a docs-only release that fixed the install instructions once the package was live on npm; v0.3.4 fixes cross-platform absolute-path detection in resolveHandoverPath (a Windows-style path was treated as relative on Linux) and moves publishing to CI with npm trusted publishing (OIDC).

Generic file upload compatibility layer for DeepSeek Harness web

DeepSeek Harness bundle for the aiworkskills WeChat article workflow

Lightweight zero-dependency memory plugin for DeepSeek Harness: Markdown-file memory store, L0 catalog injection, and 5 model-facing tools.

Image preview in the dsh session: read_image tool results render as a small thumbnail by default and open full-size in the built-in lightbox on click