DeepSeek Harness 플러그인 — 295 페이지

Audio Copilot for DeepSeek Harness: transcribe audio (ASR) and synthesize speech (TTS) — gives text-only agents ears and a voice. Windows-local SAPI TTS out of the box; OpenAI-compatible ASR/TTS endpoints configurable. Includes an in-composer voice-input (mic) button.

Dub video and audio into 10 languages with voice cloning, from a DeepSeek Harness agent — one tool call, local file or URL

Ecosystem-wide plugin health audit for DeepSeek Harness: syncs the GitHub dsh-plugin topic into a local catalog, probes npm, static-scans plugin files for security, scores every plugin (maintenance / docs / npm / ecosystem + security veto), keeps score history with movers, and exposes a web leaderboard plus agent tools.

Security auditing and session forensics for DeepSeek Harness (dsh): full tool-invocation-chain recording (who/what/files/status/duration) from the session/event stream and tool-dispatch hook, high-risk pattern flagging, SQLite (WAL) persistence with privacy masking, filter/query with JSON/Markdown/JSONL exports, plain-text session replay, and an audit policy tool-set plus a standalone CLI.

Aurora theme for DeepSeek Harness Web (persistent edition): teal palette via theme.overrideTokens.

DeepSeek 官网风格动态极光背景：深蓝流体光斑 + 金色流纹 + 粒子网格，随 dsh 启动自动加载

Aurora Glass — an original background & liquid-glass appearance plugin for the DeepSeek Harness Web UI (custom wallpapers and frosted translucent panels).

Caddy-fronted administrator authentication bundle for DeepSeek Harness

Import local Codex / Grok / Claude / Copilot / Gemini login state into DeepSeek Harness

DSH Web 用户授权插件 — 用户名密码 + 通用 OAuth 2.0（内置 GitHub/Google/Discord 等模板）+ Solana/EVM 钱包登录。零依赖核心，纯配置接入

A middle permission tier for DeepSeek Harness between workspace-write and danger-full-access: auto-approves harmless commands and operations targeting configured trusted areas, beyond the current workspace.

Autonomous (auto) mode permission classifier for DeepSeek Harness: a Claude-Code-auto-mode-like classifier over tools/pre-execute and approval/request, a selectable 'auto' permission preset, LLM semantic judge, git checkpointing, agent discipline guidance, and a web control page in Settings → Plugins.

Exit the DSH CLI process automatically when the Web UI closes (equivalent to Ctrl+C). DSH plugin: 关闭 Web UI 自动退出 CLI 进程.

DSH Auto Guard: a Claude Code Auto Mode-like command approval mechanism that adds an LLM safety net on top of full access, with rules, caches, file tracker, and sensitive path gates.

DSH Auto Maintenance System - Diagnosis, Fix, Backup, Monitor in one solution

Claude Code style auto-memory plugin for DeepSeek Harness: typed memory files + MEMORY.md index auto-injected into the system prompt

CC-style auto-approval for DeepSeek Harness: deterministic rules + model-agnostic classifier as final authority. Robust verdict parser, non-bash pre-execute escalation envelope, circuit breaker, fail-closed semantics.

DSH plugin: auto mode that routes permission-gated tool calls through an LLM review before approving, blocking, or asking for confirmation.

Auto model selection for DeepSeek Harness: routes each turn to V4 Flash or V4 Pro through a classifier call

Open the DSH Web GUI in an app-style WebView2 window on profile start, with a settings card for browser paths

Route the first prompt to one of DeepSeek Harness's four shipped Agent Presets with DeepSeek V4 Flash.

LLM auto-review approval answerer for DeepSeek Harness — decides sandbox escalations without a human prompt via a deterministic filter and a clean-context LLM safety review. REQUIRES a patched harness core (see core-patches/)

Auto-translate for the DeepSeek Harness Web GUI: when the assistant's reply is in English, a Chinese translation appears below the turn; each tool call in the turn gets a one-line Chinese gloss. The translation is a standalone provider call that never enters the session transcript.

DeepSeek Harness 视觉桥：自动发现你已配置的多模态模型，给纯文本主模型装上 vision 工具，识别结果以纯文本返回。零配置，一条命令安装。