Allow linkage to "Trusted Access for Cyber program"

While doing (non-security) patch review with copilot CLI and model Claude Opus
5, I received this error:

Execution failed: CAPIError: 422 This content was flagged for possible
cybersecurity risk. If this seem wrong, try rephrasing your request. To get
authorized for security work, join the Trusted Access for Cyber program:
https://chatgpt.com/cyber (Request ID: AAAE:E5BF3:76D4E1:87C02B:6A6D089A)

The request involved a multi-agent workflow, so Copilot may have routed some
part of it through an OpenAI model or service despite the selected session
model. In any case, the prescribed remedy does not appear usable: enrollment
at chatgpt.com/cyber is associated with my OpenAI identity, while this Copilot
request is associated with my GitHub identity.

Please provide a supported way for eligible Copilot users to obtain Trusted
Access for OpenAI-backed requests made through Copilot. This might involve a
Copilot-specific enrollment, linking an approved OpenAI identity, or another
GitHub-managed verification mechanism.

首先，在 Copilot CLI 多代理工作流中重现报告的 422 错误，并检查 Trusted Access 注册当前如何与身份关联。已批准的 Copilot 用户能够通过 GitHub 管理的受支持方式，为 OpenAI 支持的请求获取 Trusted Access，即表示完成。

由索引模型根据 Issue 内容生成。