diff --git a/.changeset/fix-serialization-error-exit.md b/.changeset/fix-serialization-error-exit.md
new file mode 100644
index 00000000..d1e9cb9d
--- /dev/null
+++ b/.changeset/fix-serialization-error-exit.md
@@ -0,0 +1,5 @@
+---
+"@googleworkspace/cli": patch
+---
+
+Exit non-zero with a clear error message when JSON serialization fails instead of printing empty stdout
diff --git a/crates/google-workspace-cli/src/executor.rs b/crates/google-workspace-cli/src/executor.rs
index 46f31ac4..07a6652f 100644
--- a/crates/google-workspace-cli/src/executor.rs
+++ b/crates/google-workspace-cli/src/executor.rs
@@ -303,11 +303,13 @@ async fn handle_json_response(
             println!(
                 "{}",
                 crate::formatter::format_value_paginated(&json_val, output_format, is_first_page)
+                    .map_err(|e| GwsError::Other(anyhow::Error::from(e)))?
             );
         } else {
             println!(
                 "{}",
                 crate::formatter::format_value(&json_val, output_format)
+                    .map_err(|e| GwsError::Other(anyhow::Error::from(e)))?
             );
         }
 
@@ -379,7 +381,11 @@ async fn handle_binary_response(
         return Ok(Some(result));
     }
 
-    println!("{}", crate::formatter::format_value(&result, output_format));
+    println!(
+        "{}",
+        crate::formatter::format_value(&result, output_format)
+            .map_err(|e| GwsError::Other(anyhow::Error::from(e)))?
+    );
 
     Ok(None)
 }
@@ -428,6 +434,7 @@ pub async fn execute_method(
         println!(
             "{}",
             crate::formatter::format_value(&dry_run_info, output_format)
+                .map_err(|e| GwsError::Other(anyhow::Error::from(e)))?
         );
         return Ok(None);
     }
diff --git a/crates/google-workspace-cli/src/formatter.rs b/crates/google-workspace-cli/src/formatter.rs
index 08d4d287..c935d0ed 100644
--- a/crates/google-workspace-cli/src/formatter.rs
+++ b/crates/google-workspace-cli/src/formatter.rs
@@ -58,12 +58,12 @@ impl OutputFormat {
 }
 
 /// Format a JSON value according to the specified output format.
-pub fn format_value(value: &Value, format: &OutputFormat) -> String {
+pub fn format_value(value: &Value, format: &OutputFormat) -> Result<String, serde_json::Error> {
     match format {
-        OutputFormat::Json => serde_json::to_string_pretty(value).unwrap_or_default(),
-        OutputFormat::Table => format_table(value),
-        OutputFormat::Yaml => format_yaml(value),
-        OutputFormat::Csv => format_csv(value),
+        OutputFormat::Json => serde_json::to_string_pretty(value),
+        OutputFormat::Table => Ok(format_table(value)),
+        OutputFormat::Yaml => Ok(format_yaml(value)),
+        OutputFormat::Csv => Ok(format_csv(value)),
     }
 }
 
@@ -76,14 +76,18 @@ pub fn format_value(value: &Value, format: &OutputFormat) -> String {
 /// For JSON the output is compact (one JSON object per line / NDJSON).
 /// For YAML each page is prefixed with a `---` document separator so the
 /// combined stream is a valid YAML multi-document file.
-pub fn format_value_paginated(value: &Value, format: &OutputFormat, is_first_page: bool) -> String {
+pub fn format_value_paginated(
+    value: &Value,
+    format: &OutputFormat,
+    is_first_page: bool,
+) -> Result<String, serde_json::Error> {
     match format {
-        OutputFormat::Json => serde_json::to_string(value).unwrap_or_default(),
-        OutputFormat::Csv => format_csv_page(value, is_first_page),
-        OutputFormat::Table => format_table_page(value, is_first_page),
+        OutputFormat::Json => serde_json::to_string(value),
+        OutputFormat::Csv => Ok(format_csv_page(value, is_first_page)),
+        OutputFormat::Table => Ok(format_table_page(value, is_first_page)),
         // Prefix every page with a YAML document separator so that the
         // concatenated stream is parseable as a multi-document YAML file.
-        OutputFormat::Yaml => format!("---\n{}", format_yaml(value)),
+        OutputFormat::Yaml => Ok(format!("---\n{}", format_yaml(value))),
     }
 }
 
@@ -469,11 +473,18 @@ mod tests {
     #[test]
     fn test_format_json() {
         let val = json!({"name": "test"});
-        let output = format_value(&val, &OutputFormat::Json);
+        let output = format_value(&val, &OutputFormat::Json).unwrap();
         assert!(output.contains("\"name\""));
         assert!(output.contains("\"test\""));
     }
 
+    #[test]
+    fn test_format_value_json_non_empty() {
+        let val = json!({"key": "value"});
+        let result = format_value(&val, &OutputFormat::Json).unwrap();
+        assert!(!result.is_empty());
+    }
+
     #[test]
     fn test_format_table_array_of_objects() {
         let val = json!({
@@ -482,7 +493,7 @@ mod tests {
                 {"id": "2", "name": "world.txt"}
             ]
         });
-        let output = format_value(&val, &OutputFormat::Table);
+        let output = format_value(&val, &OutputFormat::Table).unwrap();
         assert!(output.contains("id"));
         assert!(output.contains("name"));
         assert!(output.contains("hello.txt"));
@@ -494,7 +505,7 @@ mod tests {
     #[test]
     fn test_format_table_single_object() {
         let val = json!({"id": "abc", "name": "test"});
-        let output = format_value(&val, &OutputFormat::Table);
+        let output = format_value(&val, &OutputFormat::Table).unwrap();
         assert!(output.contains("id"));
         assert!(output.contains("abc"));
     }
@@ -512,7 +523,7 @@ mod tests {
                 "usage": "500"
             }
         });
-        let output = format_value(&val, &OutputFormat::Table);
+        let output = format_value(&val, &OutputFormat::Table).unwrap();
         // Should contain dot-notation keys
         assert!(
             output.contains("user.displayName"),
@@ -539,7 +550,7 @@ mod tests {
             {"id": "1", "owner": {"name": "Alice"}},
             {"id": "2", "owner": {"name": "Bob"}}
         ]);
-        let output = format_value(&val, &OutputFormat::Table);
+        let output = format_value(&val, &OutputFormat::Table).unwrap();
         assert!(
             output.contains("owner.name"),
             "expected flattened column:\n{output}"
@@ -555,7 +566,7 @@ mod tests {
         let long_emoji = "😀".repeat(70); // each emoji is 4 bytes
         let val = json!([{"col": long_emoji}]);
         // Should not panic
-        let output = format_value(&val, &OutputFormat::Table);
+        let output = format_value(&val, &OutputFormat::Table).unwrap();
         assert!(output.contains("col"), "column name must appear:\n{output}");
     }
 
@@ -563,7 +574,7 @@ mod tests {
     fn test_format_table_multibyte_exact_boundary() {
         // Multi-byte chars at various positions must not panic or produce garbled output.
         let val = json!([{"name": "café résumé naïve"}]);
-        let output = format_value(&val, &OutputFormat::Table);
+        let output = format_value(&val, &OutputFormat::Table).unwrap();
         assert!(output.contains("name"), "column must appear:\n{output}");
     }
 
@@ -575,7 +586,7 @@ mod tests {
                 {"id": "2", "name": "world"}
             ]
         });
-        let output = format_value(&val, &OutputFormat::Csv);
+        let output = format_value(&val, &OutputFormat::Csv).unwrap();
         assert!(output.contains("id,name"));
         assert!(output.contains("1,hello"));
         assert!(output.contains("2,world"));
@@ -591,7 +602,7 @@ mod tests {
                 ["Andrew", "Male", "1. Freshman"]
             ]
         });
-        let output = format_value(&val, &OutputFormat::Csv);
+        let output = format_value(&val, &OutputFormat::Csv).unwrap();
         let lines: Vec<&str> = output.lines().collect();
         assert_eq!(lines[0], "Student Name,Gender,Class Level");
         assert_eq!(lines[1], "Alexandra,Female,4. Senior");
@@ -602,7 +613,7 @@ mod tests {
     fn test_format_csv_flat_scalars() {
         // Flat array of non-object, non-array values → one value per line
         let val = json!(["apple", "banana", "cherry"]);
-        let output = format_value(&val, &OutputFormat::Csv);
+        let output = format_value(&val, &OutputFormat::Csv).unwrap();
         let lines: Vec<&str> = output.lines().collect();
         assert_eq!(lines.len(), 3);
         assert_eq!(lines[0], "apple");
@@ -614,7 +625,7 @@ mod tests {
     fn test_format_csv_flat_scalars_with_escaping() {
         // Scalars that contain commas/quotes must be CSV-escaped
         let val = json!(["plain", "has,comma", "has\"quote"]);
-        let output = format_value(&val, &OutputFormat::Csv);
+        let output = format_value(&val, &OutputFormat::Csv).unwrap();
         let lines: Vec<&str> = output.lines().collect();
         assert_eq!(lines.len(), 3);
         assert_eq!(lines[0], "plain");
@@ -632,7 +643,7 @@ mod tests {
     #[test]
     fn test_format_yaml() {
         let val = json!({"name": "test", "count": 42});
-        let output = format_value(&val, &OutputFormat::Yaml);
+        let output = format_value(&val, &OutputFormat::Yaml).unwrap();
         assert!(output.contains("name: \"test\""));
         assert!(output.contains("count: 42"));
     }
@@ -641,7 +652,7 @@ mod tests {
     fn test_format_table_empty_array() {
         let val = json!({"files": []});
         // No items to extract, falls back to single-object table
-        let output = format_value(&val, &OutputFormat::Table);
+        let output = format_value(&val, &OutputFormat::Table).unwrap();
         assert!(output.contains("files"));
     }
 
@@ -666,7 +677,7 @@ mod tests {
         // `drive#file` contains `#` which is a YAML comment marker; the
         // serialiser must quote it rather than emit a block scalar.
         let val = json!({"kind": "drive#file", "id": "123"});
-        let output = format_value(&val, &OutputFormat::Yaml);
+        let output = format_value(&val, &OutputFormat::Yaml).unwrap();
         // Must be a double-quoted string, not a block scalar (`|`).
         assert!(
             output.contains("kind: \"drive#file\""),
@@ -681,7 +692,7 @@ mod tests {
     #[test]
     fn test_format_yaml_colon_in_string_is_quoted() {
         let val = json!({"url": "https://example.com/path"});
-        let output = format_value(&val, &OutputFormat::Yaml);
+        let output = format_value(&val, &OutputFormat::Yaml).unwrap();
         assert!(
             output.contains("url: \"https://example.com/path\""),
             "expected double-quoted url, got:\n{output}"
@@ -692,7 +703,7 @@ mod tests {
     #[test]
     fn test_format_yaml_multiline_still_uses_block() {
         let val = json!({"body": "line one\nline two"});
-        let output = format_value(&val, &OutputFormat::Yaml);
+        let output = format_value(&val, &OutputFormat::Yaml).unwrap();
         // Multi-line content should still use block scalar.
         assert!(
             output.contains("body: |"),
@@ -710,7 +721,7 @@ mod tests {
                 {"id": "2", "name": "b.txt"}
             ]
         });
-        let output = format_value_paginated(&val, &OutputFormat::Csv, true);
+        let output = format_value_paginated(&val, &OutputFormat::Csv, true).unwrap();
         let lines: Vec<&str> = output.lines().collect();
         assert_eq!(lines[0], "id,name", "first page must start with header");
         assert_eq!(lines[1], "1,a.txt");
@@ -723,7 +734,7 @@ mod tests {
                 {"id": "3", "name": "c.txt"}
             ]
         });
-        let output = format_value_paginated(&val, &OutputFormat::Csv, false);
+        let output = format_value_paginated(&val, &OutputFormat::Csv, false).unwrap();
         let lines: Vec<&str> = output.lines().collect();
         // The first (and only) line must be a data row, not the header.
         assert_eq!(lines[0], "3,c.txt", "continuation page must have no header");
@@ -740,7 +751,7 @@ mod tests {
                 {"id": "1", "name": "foo"}
             ]
         });
-        let output = format_value_paginated(&val, &OutputFormat::Table, true);
+        let output = format_value_paginated(&val, &OutputFormat::Table, true).unwrap();
         assert!(
             output.contains("id"),
             "table header must appear on first page"
@@ -755,7 +766,7 @@ mod tests {
                 {"id": "2", "name": "bar"}
             ]
         });
-        let output = format_value_paginated(&val, &OutputFormat::Table, false);
+        let output = format_value_paginated(&val, &OutputFormat::Table, false).unwrap();
         assert!(output.contains("bar"), "data row must be present");
         assert!(
             !output.contains("──"),
@@ -766,8 +777,8 @@ mod tests {
     #[test]
     fn test_format_value_paginated_yaml_has_document_separator() {
         let val = json!({"files": [{"id": "1", "name": "foo"}]});
-        let first = format_value_paginated(&val, &OutputFormat::Yaml, true);
-        let second = format_value_paginated(&val, &OutputFormat::Yaml, false);
+        let first = format_value_paginated(&val, &OutputFormat::Yaml, true).unwrap();
+        let second = format_value_paginated(&val, &OutputFormat::Yaml, false).unwrap();
         assert!(
             first.starts_with("---\n"),
             "first YAML page must start with ---"
diff --git a/crates/google-workspace-cli/src/helpers/calendar.rs b/crates/google-workspace-cli/src/helpers/calendar.rs
index cf28b249..ac45a72d 100644
--- a/crates/google-workspace-cli/src/helpers/calendar.rs
+++ b/crates/google-workspace-cli/src/helpers/calendar.rs
@@ -415,6 +415,7 @@ async fn handle_agenda(matches: &ArgMatches) -> Result<(), GwsError> {
     println!(
         "{}",
         crate::formatter::format_value(&output, &output_format)
+            .map_err(|e| GwsError::Other(anyhow::Error::from(e)))?
     );
     Ok(())
 }
diff --git a/crates/google-workspace-cli/src/helpers/gmail/triage.rs b/crates/google-workspace-cli/src/helpers/gmail/triage.rs
index 3c275e1f..2c34e4dc 100644
--- a/crates/google-workspace-cli/src/helpers/gmail/triage.rs
+++ b/crates/google-workspace-cli/src/helpers/gmail/triage.rs
@@ -173,6 +173,7 @@ pub async fn handle_triage(matches: &ArgMatches) -> Result<(), GwsError> {
     println!(
         "{}",
         crate::formatter::format_value(&output, &output_format)
+            .map_err(|e| GwsError::Other(anyhow::Error::from(e)))?
     );
 
     Ok(())
diff --git a/crates/google-workspace-cli/src/helpers/workflows.rs b/crates/google-workspace-cli/src/helpers/workflows.rs
index b921d864..36f0e925 100644
--- a/crates/google-workspace-cli/src/helpers/workflows.rs
+++ b/crates/google-workspace-cli/src/helpers/workflows.rs
@@ -259,12 +259,17 @@ async fn get_json(
         .map_err(|e| GwsError::Other(anyhow::anyhow!("JSON parse failed: {e}")))
 }
 
-fn format_and_print(value: &Value, matches: &ArgMatches) {
+fn format_and_print(value: &Value, matches: &ArgMatches) -> Result<(), GwsError> {
     let fmt = matches
         .get_one::<String>("format")
         .map(|s| crate::formatter::OutputFormat::from_str(s))
         .unwrap_or_default();
-    println!("{}", crate::formatter::format_value(value, &fmt));
+    println!(
+        "{}",
+        crate::formatter::format_value(value, &fmt)
+            .map_err(|e| GwsError::Other(anyhow::Error::from(e)))?
+    );
+    Ok(())
 }
 
 async fn handle_standup_report(matches: &ArgMatches) -> Result<(), GwsError> {
@@ -361,7 +366,7 @@ async fn handle_standup_report(matches: &ArgMatches) -> Result<(), GwsError> {
         "date": now_in_tz.format("%Y-%m-%d").to_string(),
     });
 
-    format_and_print(&output, matches);
+    format_and_print(&output, matches)?;
     Ok(())
 }
 
@@ -405,7 +410,7 @@ async fn handle_meeting_prep(matches: &ArgMatches) -> Result<(), GwsError> {
 
     if items.is_empty() {
         let output = json!({ "message": "No upcoming meetings found." });
-        format_and_print(&output, matches);
+        format_and_print(&output, matches)?;
         return Ok(());
     }
 
@@ -438,7 +443,7 @@ async fn handle_meeting_prep(matches: &ArgMatches) -> Result<(), GwsError> {
         "attendeeCount": attendee_list.len(),
     });
 
-    format_and_print(&output, matches);
+    format_and_print(&output, matches)?;
     Ok(())
 }
 
@@ -528,7 +533,7 @@ async fn handle_email_to_task(matches: &ArgMatches) -> Result<(), GwsError> {
         "sourceMessageId": message_id,
     });
 
-    format_and_print(&output, matches);
+    format_and_print(&output, matches)?;
     Ok(())
 }
 
@@ -613,7 +618,7 @@ async fn handle_weekly_digest(matches: &ArgMatches) -> Result<(), GwsError> {
         "periodEnd": time_max,
     });
 
-    format_and_print(&output, matches);
+    format_and_print(&output, matches)?;
     Ok(())
 }
 
@@ -686,7 +691,7 @@ async fn handle_file_announce(matches: &ArgMatches) -> Result<(), GwsError> {
         "space": space,
     });
 
-    format_and_print(&output, matches);
+    format_and_print(&output, matches)?;
     Ok(())
 }