Hi! Could you clarify whether your gateway handles authentication and authorization for interactions between AI agents and MCP services and/or LLMs?
I’m particularly interested in whether there is built-in support for securing agent-to-service communication, or if this is expected to be implemented externally.
Performance is indeed better than the standard Go regexp package in many cases. However, I noticed RSS degradation in profiling when using very complex patterns.
I used it for masking sensitive data in logs inside a custom Open Policy Agent (OPA) plugin, and it worked well for that use case overall.
Really useful write-up — thanks. Worth a related pointer: the experimental jsonv2 (https://github.com/go-json-experiment) applies the same inliner trick from Filippo Valsorda's "Efficient Go APIs with the inliner" — see https://github.com/go-json-experiment/json/blob/00ed12fed2a6548f4995c2c849fe69d7fd7629ff/jsontext/token.go#L282.
Hi! Could you clarify whether your gateway handles authentication and authorization for interactions between AI agents and MCP services and/or LLMs?
I’m particularly interested in whether there is built-in support for securing agent-to-service communication, or if this is expected to be implemented externally.
Performance is indeed better than the standard Go regexp package in many cases. However, I noticed RSS degradation in profiling when using very complex patterns.
I used it for masking sensitive data in logs inside a custom Open Policy Agent (OPA) plugin, and it worked well for that use case overall.