diff --git a/src/scheduler.rs b/src/scheduler.rs
index 9fddb08..215680b 100644
--- a/src/scheduler.rs
+++ b/src/scheduler.rs
@@ -432,6 +432,33 @@ pub fn send_after_named<M: Send + 'static>(
     with_runtime(|inner| inner.timers.lock().unwrap().insert_send(deadline, armed_by, fire))
 }
 
+/// Deliver `msg` onto a channel the caller owns after `after`, rather than to a
+/// registry address. The sibling of [`send_after`] used by the gen_server timer
+/// layer (RFC 015 §5): arming a server timer must land the fire on the loop's
+/// own `Sys` channel — giving it the loop's arm position — not in the inbox.
+///
+/// Same substrate as [`send_after`]: a `Reason::Send` entry, the same `armed`
+/// set, the same [`TimerId`](crate::timer::TimerId) for [`cancel_timer`]. Only
+/// the fire thunk differs — `tx.send(msg)` instead of a registry resolve — so a
+/// send onto a channel whose receiver is gone is dropped, exactly as a failed
+/// address resolve is (Erlang `send_after` semantics). `pid` is informational
+/// (who armed it); delivery lives entirely in the thunk.
+// Wired into `TimerHandle::arm_after` in the RFC 015 timer chunk; until then it
+// is referenced only from tests, so the non-test build would flag it dead.
+#[allow(dead_code)]
+pub(crate) fn send_after_to<T: Send + 'static>(
+    after: std::time::Duration,
+    tx: Sender<T>,
+    msg: T,
+) -> crate::timer::TimerId {
+    let deadline = crate::timer::deadline_from_now(after);
+    let armed_by = current_pid().unwrap_or(Pid::new(0, 0));
+    let fire = Box::new(move || {
+        let _ = tx.send(msg);
+    });
+    with_runtime(|inner| inner.timers.lock().unwrap().insert_send(deadline, armed_by, fire))
+}
+
 /// Cancel a timer armed by [`send_after`] / [`send_after_named`]. Returns
 /// `true` if it was still pending (delivery now prevented), `false` if it had
 /// already fired or been cancelled.
@@ -701,3 +728,54 @@ pub fn register_supervisor_channel(pid: Pid, sender: Sender<Signal>) {
 pub fn run<F: FnOnce() + Send + 'static>(f: F) {
     crate::runtime::init(crate::runtime::Config::exact(1)).run(f);
 }
+
+#[cfg(test)]
+mod send_after_to_tests {
+    use super::*;
+    use crate::channel::channel;
+    use std::sync::atomic::{AtomicBool, Ordering};
+    use std::sync::Arc;
+    use std::time::Duration;
+
+    // A fired send_after_to lands its message on the caller's own channel.
+    #[test]
+    fn delivers_onto_the_channel() {
+        run(|| {
+            let (tx, rx) = channel::<u32>();
+            send_after_to(Duration::from_millis(10), tx, 42);
+            // recv parks until the scheduler fires the timer thunk.
+            assert_eq!(rx.recv().unwrap(), 42);
+        });
+    }
+
+    // cancel_timer before the deadline prevents delivery and reports the race
+    // win; the channel then closes with no message once the lone sender (moved
+    // into the now-discarded thunk) is gone.
+    #[test]
+    fn cancel_prevents_delivery() {
+        run(|| {
+            let (tx, rx) = channel::<u32>();
+            let id = send_after_to(Duration::from_millis(50), tx, 7);
+            assert!(cancel_timer(id), "cancel before fire should win the race");
+            // No delivery: the discarded thunk drops the only sender, so recv
+            // sees a closed channel rather than the value.
+            assert!(rx.recv().is_err());
+        });
+    }
+
+    // The thunk runs on the scheduler thread; a closed receiver makes the send
+    // a harmless no-op (Erlang send_after semantics) rather than a panic.
+    #[test]
+    fn send_to_closed_channel_is_harmless() {
+        let reached = Arc::new(AtomicBool::new(false));
+        let r2 = reached.clone();
+        run(move || {
+            let (tx, rx) = channel::<u32>();
+            send_after_to(Duration::from_millis(10), tx, 1);
+            drop(rx); // receiver gone before the timer fires
+            crate::sleep(Duration::from_millis(30));
+            r2.store(true, Ordering::SeqCst);
+        });
+        assert!(reached.load(Ordering::SeqCst), "runtime survived the dead-channel fire");
+    }
+}