생산성 도구 · DeepSeek Harness 플러그인 — 197 페이지

Resume-capable CLI runner for DeepSeek Harness: conversational text or machine-readable JSON from one plugin

Command palette for DeepSeek Harness (Cmd+K) — new chat, jump to session, switch model and reasoning effort, permissions, slash commands, theme, language and settings, all from one fuzzy-searchable bar.

CoAligne for DeepSeek Harness: shared team project context - synced files, Agent sessions, comments and teammate activity via the CoAligne MCP server, plus the coaligne-workflow skill

Import OpenAI Codex conversation history into DeepSeek Harness sessions, from a /codex-import slash command

Codex-style sidebar skin for the DeepSeek Harness Web GUI: a DeepSeek Harness wordmark, a single New chat row that starts a real session without picking a project (allocating a fresh working directory, registering it as a real workspace, and reclaiming it again as soon as the chat is abandoned unused), and two collapsible sections — Projects and Recents — in place of the shipped panel rows.

DSH 花费显示：每轮对话花了多少钱、整个会话累计多少钱，按官方价格表精确计价（自动刷新、高峰/低谷自动同步、多供应商多币种），并附带账户余额与可视化设置面板；余额与会话累计始终同币种。

Session cost estimate for the DSH Web UI: tokenUsage projection × configurable price table, with a one-click refresh of official DeepSeek prices

Persistent cost ledger for the DeepSeek Harness: every priced model call lands in a monthly JSONL file; /ledger renders monthly reports, /ledger-export writes CSV, and the agent can query its own spend via ledger_query.

Cost Lens for dsh web: DeepSeek API peak/off-peak rates, per-session cost tracking, schedule-to-off-peak sends, and cost displays in the sidebar, session cards, and chat stats

Cron scheduler plugin for DeepSeek Harness: schedule agent tasks via natural language or a sidebar UI, with fixed/new session strategies, run history, manual trigger, and live sync.

Cross-disciplinary association popup for DeepSeek Harness: pick the disciplines you care about, then draw a card — the LLM offers one concrete structural link between another field and what you're currently working on. A tool for thought (a hook, never a finished analysis), surfaced faint and ignorable over an in-memory Connection RPC channel, never the session log.

Drive a Windows 10 interactive desktop from DeepSeek Harness: window/UI automation, trusted screenshots, offline OCR, and Chrome CDP — as tools, slash commands and skills.

Companion plugin for DSH-desktop: current-session / usage / prompt / approval APIs, turn stats (TTFT/speed/cache), error surfacing, in-page context menu + pinned sessions, right-side file panel, and the /api/shell PTY terminal bridge.

对话上下文桥接：为大语言模型(LLM)/AI 智能体的对话做跨会话上下文桥接的 DSH 插件，把一次对话的共识编译为可移植三层快照，一键引入新对话。

滴答清单 (DIDA365 / TickTick) plugin for DeepSeek Harness: workspace-bound list, @-mention tasks in chat, and agent control via the dida365 MCP server

dsh plugin: cross-session fact notepad — save tagged one-liners, find them back in any session via /fact and the fact_find tool.

DSH 通用桌宠「菲比」：悬浮在 DSH 网页界面的动画精灵，随会话状态与鼠标互动。

DeepSeek Harness 飞花令游戏插件：简易/古法严格双模式，按会话独立维护令字、得分、已用诗句与剩余提示次数；提供 feihualing_start / feihualing_status / feihualing_stop 三个工具与关键词指令，检测到 shell/文件写入类工具调用时自动暂停并保存现场。

Workspace file explorer docked in the DeepSeek Harness sidebar: a lazy directory tree of the current session's working directory, a Settings on/off switch, and drag-to-resize.

A session-aware workspace file tree, text and media preview, and Git status panel for DeepSeek Harness.

Installation safety gate & data-protection guard for DeepSeek Harness: 60 static signature rules scan plugin sources for malicious install scripts, credential theft, obfuscation and network callbacks before 'dsh plugin add', with scan baselines (gate_diff), structured reports (gate_report), 12 destructive-command patterns and workspace-boundary checks.

DeepSeek Harness plugin: a read-only workspace Git status panel for the composer tool row and a git_status model tool.

Run validated GraphGPT YAML workflows as a native DeepSeek Harness tool

Verification-gap probe for DeepSeek Harness: a backend plugin that watches each user turn, judges steer/ride/neutral with a cheap LLM, and accumulates how far the AI's autonomous progress has outrun what the user has actually verified — surfacing a faint, ignorable prompt (over an in-memory Connection RPC channel, never the session log) when they've drifted out of grasp.