Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mrpulor-gh/nuphus --skill nuphus-modulesgit clone --depth 1 https://github.com/mrpulor-gh/nuphusWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mrpulor-gh/nuphus/nuphus-modules)<a href="https://agentmods.dev/skills/mrpulor-gh/nuphus/nuphus-modules"><img src="https://agentmods.dev/badge/skills/mrpulor-gh/nuphus/nuphus-modules/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mrpulor-gh/nuphus/nuphus-modules"><img src="https://agentmods.dev/badge/skills/mrpulor-gh/nuphus/nuphus-modules.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.04387 |
| Opus 5.5 | $0.00000 | $0.01755 |
| Sonnet 5.5 | $0.00000 | $0.00877 |
| Haiku 4.5 | $0.00000 | $0.00439 |
Grade A, and why
nuphus-modules scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 173 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Nuphus 项目模块知识
自包含的 Nuphus 项目模块摘要库:32 个模块的结构与作用,一问即答。 数据在本技能
data/目录(与技能同目录,自包含)——不引用仓库docs/(该目录不被 git 跟踪、不在发布包内)。
0. 使用协议(给 AI 执行者)
skill_query命中本技能 → 命中条目即答案骨架(定位 / 结构 / 功能作用 / 规则与坑),先答再补;- 需要行级细节或「现在怎样」→ 读源码(提示词「代码地图」已给目录:
src/src-tauri/frontend/src/);摘要内已带的关键文件:行号可直接转引; - 答案引用
源码路径:行号;摘要为 2026-11-04 提炼,可能与最新代码有漂移,行级断言以源码复核为准; - 本技能随版本发布(
plugin/skills/builtin/,编译期内嵌进二进制、首启落盘到用户数据目录),任何装机环境可用。
1. 项目骨架速答(高频问题的快速答案)
1.1 Nuphus 是什么
Tauri v2 + Rust + React 18 的桌面 AI Agent(协同共生桌面助手),「统一 Runtime + ReactAgent」架构。三层职责边界:
| 层 | 目录 | 职责 |
|---|---|---|
| 前端 | frontend/ |
UI、交互、状态展示、i18n、主题 |
| 壳 | src-tauri/ |
窗口、IPC 命令注册、AppState、事件转发(无业务逻辑) |
| 核心 | src/ |
Runtime、Agent、LLM、Memory、Tools、Security(全部业务逻辑) |
IPC 层约 177 个命令(实测 268 个命令属性 / 308 注册项);核心引擎 32+ 子模块。外部能力:LLM Provider 12+ 家多协议、桌面自动化(Win32/UIA/OCR)、浏览器 CDP(chromiumoxide)、外部 Agent(门铃 HTTP)。
1.2 Cargo workspace:5 个 crate
| Crate | 路径 | 角色 |
|---|---|---|
nuphus |
src/ |
核心引擎(lib),32+ 模块 |
nuphus-desktop |
src-tauri/ |
Tauri 桌面壳(bin nuphus) |
desktop-api |
src-tauri/crates/desktop-api/ |
底层桌面 API(鼠标/键盘/窗口/截图/OCR/输入模拟) |
nuphus-index |
crates/nuphus-index/ |
知识库内存索引引擎(扫描/解析/检索) |
nuphus-browser |
crates/nuphus-browser/ |
浏览器自动化(Chrome DevTools Protocol) |
前端是独立 npm 工程(frontend/,pnpm):React 18.3 + TypeScript 5.8 + Vite 6 + @xyflow/react 12(工作流画布)+ pdfjs-dist + Vitest 4。
1.3 一次用户请求的生命周期
用户输入(主窗口 ChatInputBar)
→ invoke("send_message_cmd") [src-tauri/commands/process.rs]
→ busy 并发锁 + 读取 soul/relation/known_paths
→ 注入记忆上下文(memory → store,FTS5 + 语义双通路)
→ run_leader_with_config(复用或新建 ReactAgent) [process/leader.rs]
→ Runtime::react_loop() —— ReAct 主循环 [src/runtime/react_loop.rs]
├─ llm → transports → Provider(SSE 流式)
├─ agent::exec_tool —— 五层安全链
│ ├─ permissions(ToolCategory 授权)
│ ├─ security(SecurityGuard / 注入检测 / 审批)
│ └─ tools(ToolRegistry 分发执行)
├─ session(上下文窗口管理)
├─ memory → store(双通路检索 + 沉淀)
├─ skill(按需知识检索,不占 prompt)
├─ workflow / browser / desktop / mcp / handoff
└─ protection(ProtectionGuard 断路器)
→ 流式事件 → event_seq 推送前端(useEvents hook 消费)
What ships with it
33 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- data/core-agent.md 3.0 KB
- data/core-browser.md 3.0 KB
- data/core-config.md 3.1 KB
- data/core-crosscut.md 3.8 KB
- data/core-desktop.md 3.0 KB
- data/core-knowledge.md 2.4 KB
- data/core-llm.md 2.3 KB
- data/core-memory.md 2.4 KB
- data/core-runtime.md 2.4 KB
- data/core-security.md 3.0 KB
- data/core-skill.md 2.3 KB
- data/core-tools.md 2.3 KB
- data/core-workflow.md 2.8 KB
- data/fe-app-shell.md 3.3 KB
- data/fe-chat.md 3.4 KB
- data/fe-hud-market-capture.md 3.0 KB
- data/fe-i18n-styles.md 2.5 KB
- data/fe-mobile.md 3.7 KB
- data/fe-overview.md 2.9 KB
- data/fe-pages.md 4.0 KB
- data/fe-testing.md 2.1 KB
- data/fe-tools.md 3.1 KB
- data/fe-workflow-canvas.md 3.8 KB
- data/fe-workflow-panels.md 3.0 KB
- data/tauri-commands.md 2.9 KB
- data/tauri-desktop-api.md 3.5 KB
- data/tauri-mobile.md 3.4 KB
- data/tauri-models.md 2.6 KB
- data/tauri-render.md 2.3 KB
- data/tauri-shell.md 3.7 KB
- data/tauri-speech.md 2.8 KB
- data/tauri-video.md 2.2 KB
- skill.json 1.5 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 173 lines · 0 tokens per session scan A 5f464766fc18
nuphus-modules is a skill published in the GitHub repository mrpulor-gh/nuphus (103 stars, last pushed yesterday), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 4,387 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-05.
nuphus-modules compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| nuphus-modulesmrpulor-gh/nuphus | Skill | 103 | yesterday | Apache-2.0 | A | 0 | Claude Code, Codex |
| browserwing-executorMemTensor/MemOS | Skill | 12k | 2d ago | Apache-2.0 | A | 42 | Claude Code, Codex |
| dev-browserMemTensor/MemOS | Skill | 12k | 2d ago | Apache-2.0 | A | 84 | Claude Code, Codex |
| browserwing-adminMemTensor/MemOS | Skill | 12k | 2d ago | Apache-2.0 | B | 47 | Claude Code, Codex |
| computer-useMemTensor/memmy-agent | Skill | 2.2k | yesterday | MIT | A | 37 | Claude Code, Codex |
| Web2SkillzhimaAi/chatwiki | Skill | 2.1k | 7d ago | none | A | 67 | Codex |
| agent-browserEvovexAI/EvoFlow | Skill | 631 | today | none | A | 68 | Claude Code, Codex |