Firewall Coverage

This page renders route-level Prompt Firewall coverage from the public API artifact. It is meant to help customers understand where text inspection runs in the governed execution path.

This projection does not expose detector rules, extractor names, bypass-oriented payload details, or internal enforcement wiring.

Prompt Firewall route coverage.
Route	Provider	Firewall	Customer Scope
POST /v1/proxy/openai	openai	Evaluated	Evaluated before provider execution when inspectable text is present.
POST /v1/proxy/anthropic	anthropic	Evaluated	Evaluated before provider execution when inspectable text is present.
POST /v1/proxy/google	google	Evaluated	Evaluated before provider execution when inspectable text is present.
POST /v1/proxy/xai	xai	Evaluated	Evaluated before provider execution when inspectable text is present.
POST /v1/proxy/meta	meta	Evaluated	Evaluated before provider execution when inspectable text is present.
POST /v1/executions	*	Evaluated	Evaluated before provider execution when inspectable text is present.
POST /v1/execute	*	Evaluated	Evaluated before provider execution when inspectable text is present.
POST /v1/permits	*	Not at submission	Permit-first mode operates on metadata only
POST /v1/jobs	*	Not at submission	Async jobs do not run firewall at submission

Coverage is route- and payload-dependent. Treat the Prompt Firewall as one layer in the governance path, not as a universal content-security boundary.