diff --git a/hub/package-lock.json b/hub/package-lock.json
index 225dfb7..4a1b665 100644
--- a/hub/package-lock.json
+++ b/hub/package-lock.json
@@ -1,12 +1,12 @@
 {
   "name": "@paradigm/hub",
-  "version": "0.0.40",
+  "version": "0.0.41",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@paradigm/hub",
-      "version": "0.0.40",
+      "version": "0.0.41",
       "dependencies": {
         "@alicloud/credentials": "^2.4.5",
         "@alicloud/docmind-api20220711": "^1.4.15",
diff --git a/hub/package.json b/hub/package.json
index f52ffb0..f44d64f 100644
--- a/hub/package.json
+++ b/hub/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@paradigm/hub",
-  "version": "0.0.40",
+  "version": "0.0.41",
   "private": true,
   "type": "module",
   "engines": {
diff --git a/hub/src/feishu/trigger.ts b/hub/src/feishu/trigger.ts
index 4647315..d8cce1e 100644
--- a/hub/src/feishu/trigger.ts
+++ b/hub/src/feishu/trigger.ts
@@ -1,7 +1,8 @@
 /**
  * Sends a "processing" reaction immediately, then streams a single
  * interactive card through the full agent run lifecycle: thinking → tool
- * calls (with trace panel) → streaming answer text → final card. The card
+ * calls (with trace panel) → streaming answer text → final card. On finish,
+ * replaces Typing with CheckMark (success) or CrossMark (failure). The card
  * shows a collapsible tool-use panel, a collapsible reasoning panel, and
  * the markdown answer text. Throttled to ~2.5 patches/sec to avoid
  * spamming the Feishu API.
@@ -680,7 +681,16 @@ export function makeTriggerHandler(deps: TriggerDeps): TriggerHandler {
             initializedSkills: [...(result.initializedSkillIds ?? [])],
           },
         });
-        await removeProcessingReaction();
+        // Mirror the start "Typing" reaction: drop processing, then stamp a
+        // terminal emoji so teachers see done/failed without reading the card.
+        const removedProcessingReaction = await removeProcessingReaction();
+        if (removedProcessingReaction) {
+          await addReaction(
+            rt,
+            msg.message_id,
+            outcome.isError ? "CrossMark" : "CheckMark",
+          );
+        }
       })
       .catch(async (e) => {
         const removedProcessingReaction = await removeProcessingReaction();
diff --git a/hub/test/unit/feishu-reactions.test.ts b/hub/test/unit/feishu-reactions.test.ts
index 4d52fde..72b2f20 100644
--- a/hub/test/unit/feishu-reactions.test.ts
+++ b/hub/test/unit/feishu-reactions.test.ts
@@ -52,7 +52,7 @@ describe("Feishu reactions", () => {
     await expect(removeReaction(rt, "message-1", "reaction-1")).resolves.toBe(false);
   });
 
-  it("adds Typing on start and removes it on success", async () => {
+  it("adds Typing on start and replaces it with CheckMark on success", async () => {
     const run = deferred<RunResult>();
     const rt = mockRuntime();
     const runAgent = vi.fn((req: RunRequest) => {
@@ -71,6 +71,7 @@ describe("Feishu reactions", () => {
       expect(rt.reactionRequests).toEqual([
         { kind: "add", messageId: "message-1", emoji: "Typing", reactionId: "reaction-1" },
         { kind: "remove", messageId: "message-1", reactionId: "reaction-1" },
+        { kind: "add", messageId: "message-1", emoji: "CheckMark", reactionId: "reaction-2" },
       ]);
     });
   });