fix: map KeyError->404 and ValueError->400 for MCP clients

Tool functions raise KeyError for unknown ids and ValueError for invalid
state transitions. Without handlers, FastAPI returned a bare 500
"Internal Server Error" to MCP clients — useless for an agent that needs
to know whether the id was wrong vs the cluster was down.

KeyError -> 404 with the message ("Unknown job_id: foo").
ValueError -> 400 with the message (e.g. "pending_id p_xyz is in status
'CANCELLED', not PENDING").

tests/integration/test_mcp_routes.py

@@ -110,3 +110,69 @@ def test_list_connections_works_with_empty_body():
     r = c.post("/list_connections", json={})
     assert r.status_code == 200
     assert r.json() == []
+
+
+# --- Exception handlers: KeyError -> 404, ValueError -> 400 ---
+
+def test_unknown_job_id_returns_404():
+    c = TestClient(app)
+    r = c.post("/get_job_status", json={"job_id": "missing"})
+    assert r.status_code == 404
+    assert "missing" in r.json()["detail"]
+
+
+def test_unknown_pending_id_returns_404():
+    c = TestClient(app)
+    r = c.post("/get_pending_job", json={"pending_id": "p_nope"})
+    assert r.status_code == 404
+    assert "p_nope" in r.json()["detail"]
+
+
+def test_unknown_connection_name_returns_404():
+    c = TestClient(app)
+    r = c.post("/get_connection", json={"name": "nope"})
+    assert r.status_code == 404
+    assert "nope" in r.json()["detail"]
+
+
+def test_unknown_connection_in_prepare_returns_404():
+    c = TestClient(app)
+    r = c.post(
+        "/prepare_submit_job",
+        json={"connection": "nope", "script_path": "/tmp/x.py"},
+    )
+    assert r.status_code == 404
+    assert "nope" in r.json()["detail"]
+
+
+def test_confirm_non_pending_returns_400():
+    """A CANCELLED pending should refuse confirm; FastAPI should surface ValueError as 400."""
+    c = TestClient(app)
+    c.post("/save_connection", json={"name": "prod", "master": "yarn"})
+    prep = c.post(
+        "/prepare_submit_job",
+        json={"connection": "prod", "script_path": "/tmp/x.py"},
+    ).json()
+    pid = prep["pending_id"]
+    c.post("/cancel_pending_job", json={"pending_id": pid})
+    r = c.post("/confirm_submit_job", json={"pending_id": pid})
+    assert r.status_code == 400
+    assert "CANCELLED" in r.json()["detail"]
+
+
+def test_cancel_already_submitted_returns_400():
+    """Cancelling a SUBMITTED pending should refuse with ValueError -> 400."""
+    c = TestClient(app)
+    c.post("/save_connection", json={"name": "prod", "master": "yarn"})
+    prep = c.post(
+        "/prepare_submit_job",
+        json={"connection": "prod", "script_path": "/tmp/x.py"},
+    ).json()
+    pid = prep["pending_id"]
+    # Simulate a SUBMITTED state by mutating the pending directly
+    p = pending_store.store.get(pid)
+    p.status = "SUBMITTED"
+    pending_store.store.save(p)
+    r = c.post("/cancel_pending_job", json={"pending_id": pid})
+    assert r.status_code == 400
+    assert "SUBMITTED" in r.json()["detail"]