Rust e2e: skip the two multi-client BYOK provider tests in-process

The runtime registers the LLM inference provider per connection and, by design,
never releases the slot on disconnect (runtime shared_api/llm_inference.rs). Over
the in-process transport every client shares this process's runtime, so a second
provider-registering client is refused ('Another client is already the LLM
inference provider') — deterministically failing reacquires_a_fresh_token_for_each_
request and dispatches_token_acquisition_per_provider when they run after the first
BYOK test. stdio spawns a separate child per test, so all three run there.

The BYOK bearer-token path over the in-process transport is validated by
callback_token_is_applied_as_authorization_header (still runs in-process); the
per-request / per-provider dispatch these two exercise is transport-agnostic and
fully covered over stdio. Guard them with skip_inprocess, matching the branch's
pattern for genuine in-process runtime limitations.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: SteveSandersonMS

rust/tests/e2e/byok_bearer_token_provider.rs

@@ -189,6 +189,18 @@ async fn callback_token_is_applied_as_authorization_header() {
 
 #[tokio::test]
 async fn reacquires_a_fresh_token_for_each_request() {
+    // The runtime registers the LLM inference provider per connection and, by design,
+    // never releases the slot on disconnect (runtime `shared_api/llm_inference.rs`). Over
+    // the in-process transport every client shares this process's runtime, so a second
+    // provider-registering client is refused ("Another client is already the LLM
+    // inference provider"). The BYOK bearer-token behavior over the in-process transport
+    // is covered by `callback_token_is_applied_as_authorization_header`; this scenario's
+    // provider-dispatch logic is transport-agnostic and is covered over stdio.
+    if super::support::skip_inprocess(
+        "llmInference.setProvider is process-global in-process; a second provider client is refused",
+    ) {
+        return;
+    }
     with_e2e_context_no_snapshot(|ctx| {
         Box::pin(async move {
             ctx.set_default_copilot_user();
@@ -252,6 +264,16 @@ async fn reacquires_a_fresh_token_for_each_request() {
 
 #[tokio::test]
 async fn dispatches_token_acquisition_per_provider() {
+    // See `reacquires_a_fresh_token_for_each_request`: in-process, the process-global LLM
+    // inference provider registration is not released on disconnect, so this additional
+    // provider-registering client is refused. The BYOK transport path is covered in-process
+    // by `callback_token_is_applied_as_authorization_header`; the per-provider dispatch
+    // logic exercised here is transport-agnostic and covered over stdio.
+    if super::support::skip_inprocess(
+        "llmInference.setProvider is process-global in-process; a second provider client is refused",
+    ) {
+        return;
+    }
     with_e2e_context_no_snapshot(|ctx| {
         Box::pin(async move {
             ctx.set_default_copilot_user();