lansi-ai/dsh-forge

C2

把 DeepSeek Harness 做成一个真正的桌面应用:Electron 主进程内嵌 Cordis Host(与官方 Web 版同内核、零移植), 渲染进程加载官方 Web UI 发行物(file:///自定义协议 + IPC 桥接,不开放 HTTP 端口), 所有桌面原生能力(托盘、全局热键、系统通知、剪贴板、开机自启、协议唤起、多窗口)以 host 插件 形态注入运行时, 与官方「一切皆

★ 1+ · lansi-ai/dsh-forge source on GitHub · this plugin in the registry

lansi-ai/dsh-forge is a DeepSeek Harness plugin rated C2 — one powerful capability or sensitive behavior. It executes system commands, decodes base64 payloads, ships no manifest.

Not a plugin — depends on dsh packages but declares no plugin manifest

What it can do

CapabilityFlagEvidence
executes system commandsexec×14 in authored code, e.g. scripts/release.cjs:35, scripts/release.cjs:35
decodes base64 payloadsbase64_decode×1 in authored code, e.g. scripts/upstream.cjs:229
ships no manifestno_manifestuses 11 services, 0 tool regs, no manifest

Services it injects

layout locale remote sessions settingsSchema settingsScope slots theme themeIcon uiWorkspace workspaces

Hooks it attaches

command/executed connection/reset theme/change

Outbound domains

codeload.github.com api.github.com api.deepseek.com

Environment variables it reads

DSH_HOME DSH_VERBOSE NO_COLOR DSH_BARE_SCREEN_MS DSH_SPLASH_DEMO

How to read this

Levels measure capability surface and transparency, not maliciousness. A C3 plugin can be entirely legitimate — a desktop shell genuinely needs subprocesses. The point is that you can see this before installing. See the levels explained and how dsh plugins work.

Findings come from static analysis of shipped code; nothing is executed. Think a flag is wrong? Open an issue — every flag cites the file and line it came from.