Just implemented credential protection in greyproxy. Instead of seeing real API keys values, the agent gets opaque `greyproxy:credential:v1:xxx` placeholders.
The proxy swaps them back before forwarding upstream. Agent never touches the actual key. #greywall #greyproxy
Currently wired OpenAI's /v1/responses into greyproxy, and we can now watch Anthropic + OpenAI traffic through the same MITM proxy now. Their streaming formats differ more than I expected - tool calls especially. I guess i was also out of the loop seeing that openai thinking is encrypted...
Mathieu Virbel