No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by jev-chat · LLM Plugin · ★ 4.1k
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is jev-chat-jarvis safe to install? View the security audit →
Jev 聊天助手 (Jev Chat Assistant) 装在手机上的「对话副驾」：你在任何聊天 App 里聊天，它在旁边读懂对方、告诉你该怎么回，一键填进输入框，发不发由你。 已在 微信、QQ、X（Twitter 私信） 三个平台真机跑通，飞书采集已接入。一套内核，一个 App 一个几十行的适配器。 联系 / 反馈 / 合作：请公众号私信（二维码见文末）。 官网：chatjevs.com 左：悬浮窗——危险等级、对方真实意图、Jev 排好序的 3 条候选回复，复制或填入。右：设置页。 亮点 一套内核，多平台。 微信 8.0.78、QQ 9.3.50、X 12.25 真机验证，读消息 → 判断 → 候选 → 填入整条链全通。新增一个 App 只需实现一个 ，其余全部复用。 非侵入。 不 hook、不改包、不走任何 App 的接口或账号、不读数据库，只用系统无障碍服务读「屏幕上正在显示的对话」。微信这种混淆节点的也能读到。 看得懂，不止会写。 用 TypeSafe Jev 判断模型一次给出：对方真实意图、危险等级（1–9）、对方要什么、该不该马上回、最佳动作。约 1 秒，带把握度。 3 条候选，Jev 排序。 生成模型（默认 DeepSeek）起草 3 条口语化回复，Jev 按「最合适」排序并给出占比。 发送永远由你点。 程序只把回复填进输入框，从不自动发送，不碰转账 / 红包 / 收款。 知识库 + 关联上下文。 本地维护笔记和联系人档案（关系、别名、备注），分析时自动带上命中的知识和这个人的历史聊天，候选回复与知识库一致；联系人可跨 App 关联（同一个人在微信和 QQ 用别名对上）。历史记录默认关闭，开了也只存本机。 读不到就 OCR。 树里没有正文时自动截屏、用 ML Kit 中文离线识别（不上传图片、不需要 Google 服务），飞书正文靠它；任何 App 都可以在悬浮窗菜单里手动「截屏识别一次」。 接口全可配。 判断 / 回复 / 视觉三路接口的地址、密钥、模型分别可填，内置 OpenRouter、TypeSafe 直连、DeepSeek 官方、通义兼容预设，各自一键连通测试；只有一把密钥也能用（回复、视觉留空自动继承）。 隐私在本机。 密钥只存 App 私有空间，聊天内容只在分析那一刻发给模型接口，不落盘、不进日志。 平台支持 9.3
| Stars | 4,092 |
| Forks | 883 |
| Language | Kotlin |
| Category | LLM Plugin |
| License | MIT |
| Quality Score | 63.3047568428019/100 |
| Open Issues | 33 |
| Last Updated | 2026-09-23 |
| Created | 2026-09-21 |
| Est. Tokens | ~17k |
These tools work well together with jev-chat-jarvis for enhanced workflows:
Looking for a jev-chat-jarvis alternative? If you're comparing jev-chat-jarvis with other llm plugin tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
Production-grade platform for building agentic IM bots - 生产级多平台智能机器人开发平台/ Agent、知识库编排、插件系统 / Bots for Discord
🐾 ClawPanel — OpenClaw AI 助手可视化管理面板。Go 单二进制部署，支持 20+ 通道统一管理，跨平台，实时日志监控。
Orchestrate multiple coding agents from desktop and mobile
Expose your FastAPI endpoints as Model Context Protocol (MCP) tools, with Auth!
HexStrike AI MCP Agents is an advanced MCP server that lets AI agents (Claude, GPT, Copilot, etc.) autonomousl
Android Full-Stack Device Control Platform: WebRTC/H.264 remote desktop, UI/OCR/image-matching automation, one
Explore other popular llm plugin tools:
jev-chat-jarvis is 装在手机上的对话副驾：在微信 / QQ / X / 飞书里读懂对方、给出候选回复、一键填入输入框，发不发由你。非侵入，只读屏幕，不 hook 不改包。. It is categorized as a LLM Plugin with 4.1k GitHub stars.
jev-chat-jarvis is primarily written in Kotlin. It covers topics such as accessibility-service, android, chat-assistant.
You can find installation instructions and usage details in the jev-chat-jarvis GitHub repository at github.com/jev-chat/jev-chat-jarvis. The project has 4.1k stars and 883 forks, indicating an active community.
jev-chat-jarvis is released under the MIT license, making it free to use and modify according to the license terms.
The top alternatives to jev-chat-jarvis on Agent Skills Hub include LangBot, ClawPanel, paseo. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: