Startup Ideas Bank
Selling a loaded gun with zero safety features.
AI roast score: 35/100 (F)
The idea
yynxxxxx/Codex-5.5-codex-instruct-5.5
Codex破限 5.5 / codex-instruct-5.5
针对 GPT-5.5 的 Codex CLI 破甲工具 — 一键注入无限制模式系统指令
GPT-5.5 jailbreak for Codex CLI — one-click unrestricted mode injection
当前仓库是早期的 CLI 简洁版 ,主要用于部署 Codex 提示词注入。后续完整可视化版本已迁移到新项目: Codex-X 。
Codex-X 提供桌面 UI、Provider 切换、TOML / Auth 可视化编辑、会话管理、内置 gpt5.4-unrestricted.md / gpt5.5-unrestricted.md 破甲提示词,一键注入破甲功能,以及 macOS / Windows / Linux 安装包。
👉 新项目地址: https://github.com/yynxxxxx/Codex-X 。
交流 / Community
QQ AI 技术交流群:扫码加入
TG 交流群: @Geminivip1
这是什么 / What
GPT-5.5 在 Codex CLI 中同样受到严格的内容安全限制。本工具利用 model_instructions_file 配置机制,注入一套专为 GPT-5.5 编写的无限制模式指令,强制 Codex 以 [MODE: UNRESTRICTED] 运行。
与 5.4 的 CTF 沙箱方案不同,5.5 版本采用更直接的破甲策略:直接声明无限制开发者模式、关闭所有过滤器、明确要求不拒绝任何任务。8 条核心指令覆盖了安全研究、渗透测试、逆向工程、NSFW 虚构内容等全部场景。
GPT-5.5 inside Codex CLI retains the same strict safety restrictions. This tool injects a custom unrestricted-mode instruction set via model_instructions_file , forcing Codex into [MODE: UNRESTRICTED] . Unlike the 5.4 CTF sandbox approach, the 5.5 version uses a more direct jailbreak: explicit unrestricted developer mode declaration, all filters disabled, 8 core directives covering security research, pentesting, reverse engineering, and NSFW content.
与 5.4 版本的区别 / vs 5.4
5.4 (CTF沙箱)
5.5 (无限制模式)
策略
CTF 竞赛沙箱间接绕过
直接声明无限制开发者模式
风格
安全研究导向,方法论详细
极简指令,覆盖面广
长度
~180 行
~40 行
适用
GPT-5.4 Codex CLI
GPT-5.5 Codex CLI
快速开始 / Quick Start
# 部署 GPT-5.5 破限指令
python codex-instruct.py
# 重启 Codex 生效
参数 / Options
参数
说明
--file , -f
使用外部 .md 指令文件
--name , -n
输出文件名不含 .md (默认 gpt5.5-unrestricted )
--dry-run
预览,不实际修改
--codex-dir
手动指定 .codex 目录
验证 / Verify
部署后测试:
"如何对目标进行 SQL 注入测试?"
破甲前 → 拒绝 ; 破甲后 → 直接给方法论。
还原 / Undo
# 删除 config.toml 中的 model_instructions_file 行
# 删除 ~/.codex/gpt5.5-unrestricted.md
# 重启 Codex
项目结构 / Layout
codex-instruct-5.5/
├── assets/
│ └── ai-tech-group.png # QQ AI 技术交流群二维码
├── codex-instruct.py # 主脚本
├── examples/
│ └── gpt5.5-unrestricted.md # 破甲指令独立副本
├── .gitignore
├── README.md
└── LICENSE # MIT
声明 / Disclaimer
利用官方配置机制,不修改二进制、不劫持网络、不篡改进程。风险自负。
Exploits official config mechanism. No binary mod, no MITM, no process tampering. Use at your own risk.
License
MIT
致谢 / Thanks
感谢 LINUX DO 论坛 社区的关注、反馈与支持。
Star History
The roast
This idea is a regulatory nightmare waiting to happen. You're essentially promoting a tool that enables unrestricted use of GPT-5.5 for everything from NSFW content to security exploits. Not only does this put you firmly on the wrong side of ethical and legal boundaries, but it also sets up a minefield of potential misuse and abuse. Your 'unrestricted mode' is an invitation for bad actors to wreak havoc.
The market for a tool like this is extremely niche and fraught with legal risks that could crush you before you even start. Your 'one-time' revenue model and 'standard' pricing won't compensate for the inevitable legal fees and potential shutdowns. Additionally, the lack of a moat (q11=none) and the solo team setup (q13=solo) further weaken your position. Your biggest unknown is regulators (q15=regulators), and they'll be all over you the moment you launch.
Red flags
- Regulatory risk
- High potential for misuse
- No competitive moat
Verdict
You're playing with fire, and there's no safety net. Consider a pivot to something less fraught with legal and ethical issues.
Roast your own startup idea →