feat(cua-driver-rs): screenshot defaults — jpeg @ 85 + max_image_dimension 1568

libs/cua-driver-rs/PARITY.md

@@ -855,11 +855,15 @@ Windows's `click` takes `{button: enum}` instead.  Rationale:
    Swift verbatim: `"✅ Window screenshot — WxH png [window_id: ID]"`
    (em-dash, checkmark, window-id suffix). Display fallback uses
    `"✅ Display screenshot — WxH png"` (Rust-only, see intentional below).
-2. **Default JPEG quality** — was 85, Swift defaults 95. Now 95.
-3. **Description** — multi-paragraph port from Swift adapted to Windows
+2. **Description** — multi-paragraph port from Swift adapted to Windows
    (BitBlt + PrintWindow transport; no permission gate needed).
-4. **`idempotent`** — was `true`; Swift uses `false` (a fresh pixel grab
+3. **`idempotent`** — was `true`; Swift uses `false` (a fresh pixel grab
    every call). Now matches Swift.
+4. **`max_image_dimension` default** — was 0 (no cap), Swift uses 1568.
+   Now 1568 on all 3 Rust platforms (matches
+   `CuaDriverConfig.defaultMaxImageDimension`). The 0 default was
+   producing 10MB screenshots on the Windows VM; 1568 caps the long
+   edge before encoding.
 
 ### Intentional Rust-only
 
@@ -868,6 +872,16 @@ Windows's `click` takes `{button: enum}` instead.  Rationale:
   Windows-only convenience that Swift can't easily provide because
   macOS Screen Recording requires per-window grants.  Schema accepts
   both shapes; description explains.
+- **Default `format`** — Swift defaults `png`; all 3 Rust platforms now
+  default `jpeg`. Rationale: agents typically want compact images for
+  vision-model context windows; PNG is lossless but multi-MB on screen
+  content. Schema still accepts both; callers wanting PNG pass
+  `{"format":"png"}`. Swift may follow; tracked as a follow-up parity
+  question.
+- **Default JPEG `quality`** — Swift defaults 95; Rust defaults 85
+  (already Linux's default and the macOS Claude-Code-compat tool's
+  default). 85 is the typical sweet spot for screen content. Diverges
+  from Swift only when both sides actually emit JPEG.
 
 ### Verified on Windows
 

libs/cua-driver-rs/crates/cua-driver/tests/mcp_protocol_test.rs

@@ -982,7 +982,7 @@ fn test_browser_eval_no_cdp_error() {
 #[test]
 #[cfg(target_os = "macos")]
 fn test_screenshot_no_window_id() {
-    //! Call screenshot without window_id — should capture the full display and return a PNG.
+    //! Call screenshot without window_id — should capture the full display and return image content.
     let binary = binary_path();
     if !binary.exists() { return; }
 

libs/cua-driver-rs/crates/platform-linux/src/tools/impl_.rs

@@ -17,7 +17,7 @@ pub struct DriverConfig {
 }
 
 impl Default for DriverConfig {
-    fn default() -> Self { Self { capture_mode: "som".into(), max_image_dimension: 0 } }
+    fn default() -> Self { Self { capture_mode: "som".into(), max_image_dimension: 1568 } }
 }
 
 pub struct ResizeRegistry {
@@ -953,12 +953,13 @@ impl Tool for ScreenshotTool {
         SS_DEF.get_or_init(|| ToolDef {
             name: "screenshot".into(),
             description: "Capture a screenshot via XGetImage or `import` (ImageMagick). \
-                Without window_id captures the full primary display. Supports png and jpeg formats.".into(),
+                Without window_id captures the full primary display. Supports png and jpeg formats \
+                (default jpeg, quality 85).".into(),
             input_schema: json!({
                 "type":"object","properties":{
                     "window_id":{"type":"integer"},
-                    "format":{"type":"string","enum":["png","jpeg"]},
-                    "quality":{"type":"integer","minimum":1,"maximum":95}
+                    "format":{"type":"string","enum":["png","jpeg"],"description":"Image format. Default: jpeg."},
+                    "quality":{"type":"integer","minimum":1,"maximum":95,"description":"JPEG quality 1-95; ignored for png. Default: 85."}
                 },"additionalProperties":false
             }),
             read_only: true, destructive: false, idempotent: true, open_world: false,
@@ -967,7 +968,7 @@ impl Tool for ScreenshotTool {
 
     async fn invoke(&self, args: Value) -> ToolResult {
         let xid_opt = args.get("window_id").and_then(|v| v.as_u64());
-        let format = args.get("format").and_then(|v| v.as_str()).unwrap_or("png").to_owned();
+        let format = args.get("format").and_then(|v| v.as_str()).unwrap_or("jpeg").to_owned();
         let quality = args.get("quality").and_then(|v| v.as_u64()).unwrap_or(85) as u8;
         let is_jpeg = format == "jpeg";
         let max_dim = self.state.config.read().unwrap().max_image_dimension;

libs/cua-driver-rs/crates/platform-macos/src/tools/mod.rs

@@ -109,14 +109,16 @@ pub struct DriverConfig {
     /// Default capture_mode for get_window_state when not specified per-call.
     pub capture_mode: String,
     /// Max screenshot dimension (0 = no limit). Applied during screenshot/zoom.
+    /// Default 1568 matches Swift's `CuaDriverConfig.defaultMaxImageDimension` —
+    /// the long edge is downscaled to this before encoding.
     pub max_image_dimension: u32,
 }
 
 impl Default for DriverConfig {
     fn default() -> Self {
         Self {
             capture_mode: "som".to_owned(),
-            max_image_dimension: 0,
+            max_image_dimension: 1568,
         }
     }
 }

libs/cua-driver-rs/crates/platform-macos/src/tools/screenshot.rs

@@ -16,7 +16,7 @@ fn def() -> &'static ToolDef {
         name: "screenshot".into(),
         description:
             "Capture a screenshot. Returns base64-encoded image data in the requested format \
-             (default png).\n\n\
+             (default jpeg, quality 85).\n\n\
              Without `window_id`, captures the full main display. With `window_id`, captures \
              just that window (pair with `list_windows` or `get_accessibility_tree` which return \
              window IDs).\n\n\
@@ -32,13 +32,13 @@ fn def() -> &'static ToolDef {
                 "format": {
                     "type": "string",
                     "enum": ["png", "jpeg"],
-                    "description": "Image format. Default: png."
+                    "description": "Image format. Default: jpeg."
                 },
                 "quality": {
                     "type": "integer",
                     "minimum": 1,
                     "maximum": 95,
-                    "description": "JPEG quality 1-95; ignored for png. Default: 95."
+                    "description": "JPEG quality 1-95; ignored for png. Default: 85."
                 }
             },
             "additionalProperties": false
@@ -56,8 +56,8 @@ impl Tool for ScreenshotTool {
 
     async fn invoke(&self, args: Value) -> ToolResult {
         let window_id = args.get("window_id").and_then(|v| v.as_u64()).map(|v| v as u32);
-        let format    = args.get("format").and_then(|v| v.as_str()).unwrap_or("png").to_owned();
-        let quality   = args.get("quality").and_then(|v| v.as_u64()).unwrap_or(95) as u8;
+        let format    = args.get("format").and_then(|v| v.as_str()).unwrap_or("jpeg").to_owned();
+        let quality   = args.get("quality").and_then(|v| v.as_u64()).unwrap_or(85) as u8;
         let use_jpeg  = format == "jpeg";
         let max_dim   = self.state.config.read().unwrap().max_image_dimension;
 

libs/cua-driver-rs/crates/platform-windows/src/tools/impl_.rs

@@ -43,7 +43,7 @@ pub struct DriverConfig {
 }
 
 impl Default for DriverConfig {
-    fn default() -> Self { Self { capture_mode: "som".into(), max_image_dimension: 0 } }
+    fn default() -> Self { Self { capture_mode: "som".into(), max_image_dimension: 1568 } }
 }
 
 pub struct ResizeRegistry {
@@ -2141,17 +2141,17 @@ impl Tool for ScreenshotTool {
             // ScreenCaptureKit).
             description: "Capture a screenshot of a single window via BitBlt + PrintWindow \
                 (no focus change). Returns base64-encoded image data in the requested format \
-                (default png).\n\n\
+                (default jpeg, quality 85).\n\n\
                 `window_id` is recommended (use `list_windows` to find it). When omitted, \
                 captures the full primary display — a Windows-only convenience for whole-\
                 screen snapshots without a per-window target.\n\n\
                 Requires no special permissions on Windows.".into(),
             input_schema: json!({
                 "type":"object","properties":{
                     "window_id":{"type":"integer","description":"HWND of the window to capture. When omitted, captures the full primary display (Windows-only)."},
-                    "format":{"type":"string","enum":["png","jpeg"],"description":"Image format. Default: png."},
+                    "format":{"type":"string","enum":["png","jpeg"],"description":"Image format. Default: jpeg."},
                     "quality":{"type":"integer","minimum":1,"maximum":95,
-                        "description":"JPEG quality 1-95; ignored for png."}
+                        "description":"JPEG quality 1-95; ignored for png. Default: 85."}
                 },"additionalProperties":false
             }),
             read_only: true, destructive: false, idempotent: false, open_world: false,
@@ -2160,8 +2160,8 @@ impl Tool for ScreenshotTool {
 
     async fn invoke(&self, args: Value) -> ToolResult {
         let hwnd_opt = args.get("window_id").and_then(|v| v.as_u64());
-        let format = args.get("format").and_then(|v| v.as_str()).unwrap_or("png").to_owned();
-        let quality = args.get("quality").and_then(|v| v.as_u64()).unwrap_or(95) as u8;
+        let format = args.get("format").and_then(|v| v.as_str()).unwrap_or("jpeg").to_owned();
+        let quality = args.get("quality").and_then(|v| v.as_u64()).unwrap_or(85) as u8;
         let is_jpeg = format == "jpeg";
         let max_dim = self.state.config.read().unwrap().max_image_dimension;
 

libs/cua-driver-rs/tests/integration/test_cursor_visibility.py

@@ -255,8 +255,10 @@ def test_custom_cursor_visible_after_click(self) -> None:
         # 7. Wait for cursor animation + render loop to settle.
         time.sleep(0.5)
 
-        # 8. Take a full-screen screenshot.
-        scr = client.call_tool("screenshot", {})
+        # 8. Take a full-screen screenshot. Force PNG so the per-pixel decode
+        # below works — the screenshot tool defaults to JPEG, which would be
+        # lossy and skew the cursor-colour search.
+        scr = client.call_tool("screenshot", {"format": "png"})
         b64 = scr.get("content", [{}])[0].get("data", "")
         self.assertTrue(b64, "screenshot returned no data")
         png_bytes = base64.b64decode(b64)