EthanYoQ/AI-Novel-Writer

C3

本地优先 AI 小说创作工作台,提供 Windows/macOS 桌面版与 DeepSeek Harness 插件开发预览,支持角色、大纲、章节蓝图、审稿修稿和本地模型。

★ 100+ · EthanYoQ/AI-Novel-Writer source on GitHub · this plugin in the registry

EthanYoQ/AI-Novel-Writer is a DeepSeek Harness plugin rated C3 — powerful capability combined with sensitive behavior. It patches the dsh runtime, can rewrite the system prompt, gates tool execution, executes system commands.

What it can do

CapabilityFlagEvidence
patches the dsh runtimeruntime_patch./cordis.patch.yml
can rewrite the system promptprompt_surfaceapproval, systemPrompt, tools/pre-execute
gates tool executiontool_gatehook: tools/pre-execute
executes system commandsexec×47 src, e.g. scripts/prepare-native-for-node.mjs:1, scripts/prepare-native-for-node.mjs:1
decodes base64 payloadsbase64_decode×1 src, e.g. electron/security/windows-safe-file-system.ts:545
starts a network servernet_server×2 src, e.g. plugins/dsh-ai-novel-writer/scripts/qualify-release.mjs:373, plugins/dsh-ai-novel-writer/scripts/qualify-release.mjs:375
reads credential-class env varstoken_envAI_NOVEL_RELEASE_SMOKE_TOKEN, GITHUB_TOKEN

Services it injects

agentLoop agentPresets agents approval connection llm sessions slots systemPrompt tools workspaceRegistry workspaces

Hooks it attaches

agent-preset/selected approval/request connection/reset llm/stream session/event tools/pre-execute

Outbound domains

api.openai.com api.deepseek.com api.github.com api-docs.deepseek.com api.x.ai docs.x.ai ai.google.dev generativelanguage.googleapis.com cloud.siliconflow.cn docs.siliconflow.cn

Environment variables it reads

APP_ROOT AI_NOVEL_RELEASE_EVIDENCE_ROOT GITHUB_TOKEN VITE_PUBLIC npm_execpath HTTP_PROXY HTTPS_PROXY http_proxy https_proxy AI_NOVEL_SMOKE_OPEN_PROJECT

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.