ChatGPT 代码泄露了 dots 的永久电子邮件地址和固定工作账户句柄
OpenAI 正在为 ChatGPT dots 构建一个原生的电子邮件地址设置流程，包含永久个人标识、固定的工作账号地址以及在代理现有权限下激活电子邮件的控制，根据 RuntimeWire 在官方 Windows 应用中发现的代码。
By Ryan Merket · Published · Updated
RUNTIMEWIRE INVESTIGATION — Original analysis
Original reporting by RuntimeWire, based on reverse engineering.
Reporting record
Finding
Code in the official Windows app shows a native email setup flow under development with permanent personal handles, fixed work addresses, and consent and permission controls for activation.
How we verified
Methods: reverse engineering.
The app’s bundled client code contains personal-handle validation and permanence copy, separate work-address logic, claim and link requests recording consent version 1, and controls for activating, disconnecting, and reconnecting an address.
RuntimeWire inspected the email-related components and branch logic in the official Windows app, then prepared labeled reconstructions of the personal-claim and work-activation dialogs using the code’s copy, control order, and states. The addresses and styling were fictional or approximate; RuntimeWire did not claim an address or send a message.
Reproduction
RuntimeWire partially reproduced the finding.
Reverse engineer the binary and supporting files.
Company response
The company was not contacted before publication.
OpenAI 正在为 ChatGPT dots 构建原生电子邮件地址设置流程，包含永久个人标识符、固定的工作账号地址，以及在代理现有权限下启用电子邮件的控制，这些信息来自 RuntimeWire 在官方 Windows 应用中发现的代码。
该实现超出了 dots 推出前报道的电子邮件后缀引用。它包括由服务器提供的地址选项、一个认领请求、地址与特定 dot 的关联、发送同意处理，以及用于断开和重新连接已拥有地址的界面。RuntimeWire 尚未确认该流程是否向普通账户开放、是否有人认领了地址或通过该地址发送过消息。…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.