fix: block ALL LS follow-up requests across connections

Move the in-flight blocking check to the top of the LLM request flow,
BEFORE request modification. This catches follow-ups on ALL connections
(the LS opens multiple parallel TLS connections). Only the very first
modified request reaches Google — all others get fake STOP responses.

Previously, each new connection independently allowed one request
through before blocking, letting 4-5 requests leak per turn.

src/mitm/store.rs

@@ -2,11 +2,11 @@
 //!
 //! The MITM proxy writes usage data here; the API handlers read from it.
 
-use std::collections::HashMap;
-use std::sync::Arc;
-use std::sync::atomic::{AtomicBool, Ordering};
-use tokio::sync::RwLock;
 use serde::{Deserialize, Serialize};
+use std::collections::HashMap;
+use std::sync::atomic::{AtomicBool, Ordering};
+use std::sync::Arc;
+use tokio::sync::RwLock;
 use tracing::{debug, info};
 
 /// Token usage from an intercepted API response.
@@ -342,7 +342,9 @@ impl MitmStore {
 
     /// Record a captured function call from Google's response.
     pub async fn record_function_call(&self, cascade_id: Option<&str>, fc: CapturedFunctionCall) {
-        let key = cascade_id.map(|s| s.to_string()).unwrap_or_else(|| "_latest".to_string());
+        let key = cascade_id
+            .map(|s| s.to_string())
+            .unwrap_or_else(|| "_latest".to_string());
         info!(
             cascade = %key,
             tool = %fc.name,
@@ -377,7 +379,6 @@ impl MitmStore {
         self.awaiting_tool_result.store(false, Ordering::SeqCst);
     }
 
-
     /// Take any pending function calls (ignoring cascade ID).
     pub async fn take_any_function_calls(&self) -> Option<Vec<CapturedFunctionCall>> {
         let mut pending = self.pending_function_calls.write().await;
@@ -457,8 +458,6 @@ impl MitmStore {
 
     // ── Direct response capture (bypass LS) ──────────────────────────────
 
-
-
     /// Set (replace) the captured response text.
     pub async fn set_response_text(&self, text: &str) {
         *self.captured_response_text.write().await = Some(text.to_string());
@@ -484,8 +483,6 @@ impl MitmStore {
         self.response_complete.load(Ordering::SeqCst)
     }
 
-
-
     /// Async version of clear_response.
     pub async fn clear_response_async(&self) {
         self.response_complete.store(false, Ordering::SeqCst);