diff --git a/docs/explanation/repository-lifecycle.md b/docs/explanation/repository-lifecycle.md
index 7a18db7..349fdb2 100644
--- a/docs/explanation/repository-lifecycle.md
+++ b/docs/explanation/repository-lifecycle.md
@@ -328,11 +328,16 @@ When `deletion_request_disrespector = true`:
 ```
 1. Kind 5 deletion request arrives
    ↓
-2. Store deletion request event in main database
+2. Check idempotency:
+   - if every actionable `e`/`a` target is already covered by an existing
+     same-author kind-5 request, return a duplicate success response without
+     storing the new deletion event
    ↓
-3. Do NOT process deletion
+3. Store deletion request event in main database
    ↓
-4. Repository and events remain fully accessible
+4. Do NOT process deletion
+   ↓
+5. Repository and events remain fully accessible
    ↓
 Result: Archival relay preserves all content
 ```
@@ -349,9 +354,11 @@ NIP-62 vanish requests. It does NOT prevent blacklist-triggered deletions.
 **Implementation Note:** Implemented. The LMDB backend's automatic NIP-09 and
 NIP-62 processing is disabled (`process_nip09(false)`, `process_nip62(false)`);
 ngit-grasp owns deletion handling in relay policy code, which short-circuits
-when `deletion_request_disrespector` is set — storing but not acting on kind-5
-or kind-62 requests. NIP-09 and NIP-62 are omitted from the NIP-11
-`supported_nips` list in this mode.
+already-covered NIP-09 deletions before the `deletion_request_disrespector`
+archival-mode branch. When `deletion_request_disrespector` is set, new kind-5
+requests are stored but not acted on, and kind-62 requests are stored but not
+acted on. NIP-09 and NIP-62 are omitted from the NIP-11 `supported_nips` list
+in this mode.
 
 ## Recovery Mechanism
 
diff --git a/src/nostr/lifecycle/deletion/policy.rs b/src/nostr/lifecycle/deletion/policy.rs
index 35e2c0a..7b75aa7 100644
--- a/src/nostr/lifecycle/deletion/policy.rs
+++ b/src/nostr/lifecycle/deletion/policy.rs
@@ -73,9 +73,32 @@ impl DeletionPolicy {
     /// main database) so clients see an OK and the request is preserved, but it
     /// is NOT acted upon — no purgatory eviction, no main-DB deletion, and no
     /// tombstone recording. The targeted events therefore remain fully
-    /// accessible. The same archival-mode contract is applied to NIP-62 vanish
-    /// requests in [`DeletionService::handle_vanish`](super::service::DeletionService::handle_vanish).
+    /// accessible. Already-covered kind-5 requests are still treated as
+    /// duplicates before archival-mode storage to avoid polluting the served
+    /// deletion-request stream. The same archival-mode contract is applied to
+    /// NIP-62 vanish requests in [`DeletionService::handle_vanish`](super::service::DeletionService::handle_vanish).
     pub async fn handle(&self, event: &Event) -> WritePolicyResult {
+        match self
+            .ctx
+            .tombstones
+            .deletion_targets_already_covered(event)
+            .await
+        {
+            Ok(true) => {
+                tracing::info!(
+                    event_id = %event.id.to_hex(),
+                    author = %event.pubkey.to_hex(),
+                    "Skipping duplicate NIP-09 deletion request; all actionable targets are already covered"
+                );
+                return duplicate("deletion target(s) already covered");
+            }
+            Ok(false) => {}
+            Err(e) => {
+                tracing::warn!(error = %e, "Tombstone lookup failed during duplicate deletion check");
+                return reject_error(format!("internal error checking deletion coverage: {e}"));
+            }
+        }
+
         // Archival mode: store the deletion request but do not process it.
         if self.ctx.config.deletion_request_disrespector {
             tracing::info!(
@@ -104,27 +127,6 @@ impl DeletionPolicy {
             return reject_invalid("too many deletion targets");
         }
 
-        match self
-            .ctx
-            .tombstones
-            .deletion_targets_already_covered(event)
-            .await
-        {
-            Ok(true) => {
-                tracing::info!(
-                    event_id = %event.id.to_hex(),
-                    author = %event.pubkey.to_hex(),
-                    "Skipping duplicate NIP-09 deletion request; all actionable targets are already covered"
-                );
-                return duplicate("deletion target(s) already covered");
-            }
-            Ok(false) => {}
-            Err(e) => {
-                tracing::warn!(error = %e, "Tombstone lookup failed during duplicate deletion check");
-                return reject_error(format!("internal error checking deletion coverage: {e}"));
-            }
-        }
-
         // Validate authorship of all targets first. If any `e`-tag target exists
         // in the main DB and is owned by someone else, the whole deletion is
         // invalid (mirrors the backend's `handle_deletion_event` returning