@swmansion/argent 0.15.1-next.4 → 0.15.1-next.5

package/dist/tool-server.cjs

@@ -126102,7 +126102,10 @@ var sleep2 = (ms) => new Promise((r) => setTimeout(r, ms));
 var zodSchema16 = external_exports.object({
   udid: external_exports.string().describe("Target device id from `list-devices` (iOS UDID, Android serial, or Chromium id)."),
   x: external_exports.number().describe("Normalized horizontal position 0.0\u20131.0 (left=0, right=1), not pixels"),
-  y: external_exports.number().describe("Normalized vertical position 0.0\u20131.0 (top=0, bottom=1), not pixels")
+  y: external_exports.number().describe("Normalized vertical position 0.0\u20131.0 (top=0, bottom=1), not pixels"),
+  clickCount: external_exports.number().int().min(1).max(10).optional().describe(
+    "Number of taps/clicks dispatched as ONE multi-tap gesture (2 = double-tap / double-click). The taps land inside the OS double-tap window; on Chromium each click carries an escalating CDP clickCount so dblclick actually fires. Default 1."
+  )
 });
 var capability8 = {
   apple: { simulator: true, device: true },
@@ -126110,19 +126113,25 @@ var capability8 = {
   android: { emulator: true, device: true, unknown: true },
   chromium: { app: true }
 };
-async function tapChromium(api, x, y) {
+var TAP_HOLD_MS = 50;
+var MULTI_TAP_GAP_MS = 100;
+async function tapChromium(api, x, y, clickCount) {
   const vp = api.getViewport();
   const pxX = Math.max(0, Math.min(vp.width, x * vp.width));
   const pxY = Math.max(0, Math.min(vp.height, y * vp.height));
   await api.dispatchMouseEvent({ type: "mouseMoved", x: pxX, y: pxY });
-  await api.dispatchMouseEvent({ type: "mousePressed", x: pxX, y: pxY, clickCount: 1 });
-  await sleep2(50);
-  await api.dispatchMouseEvent({ type: "mouseReleased", x: pxX, y: pxY, clickCount: 1 });
+  for (let i = 1; i <= clickCount; i++) {
+    if (i > 1) await sleep2(MULTI_TAP_GAP_MS);
+    await api.dispatchMouseEvent({ type: "mousePressed", x: pxX, y: pxY, clickCount: i });
+    await sleep2(TAP_HOLD_MS);
+    await api.dispatchMouseEvent({ type: "mouseReleased", x: pxX, y: pxY, clickCount: i });
+  }
 }
 var gestureTapTool = {
   id: "gesture-tap",
   description: `Press the device screen (iOS simulator, Android emulator, or Chromium app) at normalized coordinates: x and y are fractions of screen width and height in 0.0\u20131.0 (not pixels).
 Sends a Down event followed by an Up event at the same point. For Chromium, this dispatches a CDP mouse-press/release on the renderer.
+Set clickCount: 2 for a double-tap / double-click \u2014 the taps are dispatched as one gesture with proper click counting, which two separate tap calls cannot guarantee.
 Use when you need to tap a button, link, or any tappable element on the screen.
 Returns { tapped: true, timestampMs }. Fails if the simulator-server / emulator backend / Chromium CDP is not reachable for the given device.
 Before tapping, determine the correct coordinates by using discovery tools \u2014 pick by platform: iOS / Android use \`describe\`, \`native-describe-screen\`, or \`debugger-component-tree\`; Chromium uses \`describe\` (the DOM walker), since the native and RN-specific discovery tools don't apply. More information in \`argent-device-interact\` skill`,
@@ -126140,29 +126149,33 @@ Before tapping, determine the correct coordinates by using discovery tools \u201
   async execute(services, params) {
     const device = resolveDevice(params.udid);
     const timestampMs = Date.now();
+    const clickCount = params.clickCount ?? 1;
     if (device.platform === "chromium") {
       const chromium = services.chromium;
-      await tapChromium(chromium, params.x, params.y);
+      await tapChromium(chromium, params.x, params.y, clickCount);
       return { tapped: true, timestampMs };
     }
     const api = services.simulatorServer;
-    sendCommand(api, {
-      cmd: "touch",
-      type: "Down",
-      x: params.x,
-      y: params.y,
-      second_x: null,
-      second_y: null
-    });
-    await sleep2(50);
-    sendCommand(api, {
-      cmd: "touch",
-      type: "Up",
-      x: params.x,
-      y: params.y,
-      second_x: null,
-      second_y: null
-    });
+    for (let i = 1; i <= clickCount; i++) {
+      if (i > 1) await sleep2(MULTI_TAP_GAP_MS);
+      sendCommand(api, {
+        cmd: "touch",
+        type: "Down",
+        x: params.x,
+        y: params.y,
+        second_x: null,
+        second_y: null
+      });
+      await sleep2(TAP_HOLD_MS);
+      sendCommand(api, {
+        cmd: "touch",
+        type: "Up",
+        x: params.x,
+        y: params.y,
+        second_x: null,
+        second_y: null
+      });
+    }
     return { tapped: true, timestampMs };
   }
 };
@@ -133567,7 +133580,7 @@ a prior tap), use individual tool calls instead.
 
 Allowed tools and their args (udid is auto-injected, do NOT include it in args):
 
-  gesture-tap:    { x: number, y: number }                                                                              [ios/android/chromium]
+  gesture-tap:    { x: number, y: number, clickCount?: number }                                                        [ios/android/chromium]
   gesture-swipe:  { fromX: number, fromY: number, toX: number, toY: number, durationMs?: number }                       [ios/android]
   gesture-scroll: { x: number, y: number, deltaX?: number, deltaY?: number, durationMs?: number }                       [chromium only]
   gesture-drag:   { fromX: number, fromY: number, toX: number, toY: number, durationMs?: number }                       [chromium only]
@@ -144544,7 +144557,12 @@ function toYamlStep(step) {
     case "run":
       return { run: step.flow };
     case "tap": {
-      const body = step.selector ? selectorToYaml(step.selector) : { x: step.x, y: step.y };
+      if (step.selector) {
+        const sel = selectorToYaml(step.selector);
+        return { tap: step.times !== void 0 ? { on: sel, times: step.times } : sel };
+      }
+      const body = { x: step.x, y: step.y };
+      if (step.times !== void 0) body.times = step.times;
       return { tap: body };
     }
     case "long-press": {
@@ -144839,6 +144857,50 @@ function parseLongPress(body, entry) {
   }
   return { kind: "long-press", selector: parseSelector(body, "long-press") };
 }
+function parseTapTimes(raw, entry) {
+  if (raw === void 0) return void 0;
+  if (typeof raw !== "number" || !Number.isInteger(raw) || raw < 1 || raw > 10) {
+    badEntry(entry, "tap.times must be an integer between 1 and 10 (2 = double-tap)");
+  }
+  return raw === 1 ? void 0 : raw;
+}
+function parseTap(body, entry) {
+  const obj = body !== null && typeof body === "object" ? body : {};
+  if (obj.x !== void 0 || obj.y !== void 0) {
+    if (obj.on !== void 0 || obj.text !== void 0 || obj.id !== void 0 || obj.identifier !== void 0 || obj.role !== void 0) {
+      badEntry(entry, "tap takes a selector or x/y coordinates, not both");
+    }
+    if (typeof obj.x !== "number" || typeof obj.y !== "number") {
+      badEntry(entry, "a coordinate tap needs numeric x and y");
+    }
+    if (!Object.keys(obj).every((k) => k === "x" || k === "y" || k === "times")) {
+      badEntry(entry, "a coordinate tap takes only { x, y, times }");
+    }
+    const step = { kind: "tap", x: obj.x, y: obj.y };
+    const times = parseTapTimes(obj.times, entry);
+    if (times !== void 0) step.times = times;
+    return step;
+  }
+  if (obj.on !== void 0 || obj.times !== void 0) {
+    if (obj.text !== void 0 || obj.id !== void 0 || obj.identifier !== void 0 || obj.role !== void 0) {
+      badEntry(
+        entry,
+        'the tap options form takes a nested selector \u2014 e.g. tap: { on: { text: "Photo" }, times: 2 }'
+      );
+    }
+    if (!Object.keys(obj).every((k) => k === "on" || k === "times")) {
+      badEntry(entry, "the tap options form accepts only { on, times }");
+    }
+    if (obj.on === void 0) {
+      badEntry(entry, 'tap with times needs a target \u2014 e.g. tap: { on: "Photo", times: 2 }');
+    }
+    const step = { kind: "tap", selector: parseSelector(obj.on, "tap.on") };
+    const times = parseTapTimes(obj.times, entry);
+    if (times !== void 0) step.times = times;
+    return step;
+  }
+  return { kind: "tap", selector: parseSelector(body, "tap") };
+}
 function fromYamlStep(raw) {
   const entry = raw;
   const kinds = STEP_DIRECTIVE_KEYS.filter((k) => k in entry);
@@ -144864,23 +144926,7 @@ function fromYamlStep(raw) {
   if ("echo" in raw) return { kind: "echo", message: String(raw.echo) };
   if ("launch" in raw) return { kind: "launch", app: parseLaunch(raw.launch) };
   if ("run" in raw) return { kind: "run", flow: String(raw.run) };
-  if ("tap" in raw) {
-    const body = raw.tap;
-    const obj = body !== null && typeof body === "object" ? body : {};
-    if (body !== null && typeof body === "object" && !Array.isArray(body)) {
-      rejectUnknownKeys(raw, obj, [...SELECTOR_KEYS, "x", "y"], "tap");
-    }
-    if (obj.x !== void 0 || obj.y !== void 0) {
-      if (obj.text !== void 0 || obj.id !== void 0 || obj.identifier !== void 0 || obj.role !== void 0) {
-        badEntry(raw, "tap takes a selector or x/y coordinates, not both");
-      }
-      if (typeof obj.x !== "number" || typeof obj.y !== "number") {
-        badEntry(raw, "a coordinate tap needs numeric x and y");
-      }
-      return { kind: "tap", x: obj.x, y: obj.y };
-    }
-    return { kind: "tap", selector: parseSelector(body, "tap") };
-  }
+  if ("tap" in raw) return parseTap(raw.tap, raw);
   if ("long-press" in raw) {
     return parseLongPress(raw["long-press"], raw);
   }
@@ -145719,13 +145765,15 @@ If a step was recorded by mistake, edit the .yaml file directly to remove it.`,
       const runTarget = params.command === RUN_TARGET_COMMAND && params.delayMs === void 0 ? await captureRunTarget(session, args) : void 0;
       const strippedArgs = stripDeviceKeys(args);
       const isLaunch = params.command === "restart-app" && params.delayMs === void 0 && typeof strippedArgs.bundleId === "string" && Object.keys(strippedArgs).length === 1;
+      const cc = args.clickCount;
+      const tapTimes = isTap && typeof cc === "number" && Number.isInteger(cc) && cc >= 2 && cc <= 10 ? { times: cc } : {};
       let step;
       let warning;
       if (captured?.selector) {
-        step = { kind: "tap", selector: captured.selector };
+        step = { kind: "tap", selector: captured.selector, ...tapTimes };
         warning = captured.warning;
       } else if (isTap) {
-        step = { kind: "tap", x: args.x, y: args.y };
+        step = { kind: "tap", x: args.x, y: args.y, ...tapTimes };
         warning = captured?.warning;
       } else if (isLaunch) {
         step = { kind: "launch", app: strippedArgs.bundleId };
@@ -146101,7 +146149,10 @@ async function runTap(env, target) {
   } else {
     return { ok: false, reason: "tap needs a selector or x/y coordinates" };
   }
-  await invokeOnDevice(env, "gesture-tap", point);
+  await invokeOnDevice(env, "gesture-tap", {
+    ...point,
+    ...target.times !== void 0 ? { clickCount: target.times } : {}
+  });
   return { ok: true };
 }
 var DEFAULT_LONG_PRESS_MS = 800;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@swmansion/argent",
-  "version": "0.15.1-next.4",
+  "version": "0.15.1-next.5",
   "description": "MCP server for iOS Simulator and Android Emulator control",
   "license": "Apache-2.0",
   "repository": {

package/skills/argent-create-flow/SKILL.md

@@ -23,7 +23,7 @@ Beyond raw `tool:` steps and `echo:`, flows support declarative directives inter
 | Directive    | YAML                                                                                                     | Meaning                                                                 |
 | ------------ | -------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------- |
 | `launch`     | `- launch: com.acme.app` or `- launch: { ios: …, android: … }`                                           | start the app from scratch (terminate + relaunch) and wait until ready  |
-| `tap`        | `- tap: Login` or `- tap: { x: 0.5, y: 0.57 }`                                                           | tap an element by selector (auto-waits), or a raw normalized point      |
+| `tap`        | `- tap: Login`, `- tap: { x: 0.5, y: 0.57 }`, `- tap: { on: Login, times: 2 }`                           | tap by selector (auto-waits) or raw point; `times: 2` = double-tap      |
 | `long-press` | `- long-press: Row 3` or `- long-press: { on: <sel>, duration: 1200 }`                                   | press and hold an element (default 800ms; Chromium: mouse press-hold)   |
 | `type`       | `- type: { into: email, text: "a@b.com" }`                                                               | focus a field, type, then press Enter to submit + dismiss the keyboard  |
 | `scroll-to`  | `- scroll-to: "Order #1234"` (scrolls down) or `- scroll-to: { target: …, direction: right, within: … }` | momentum-free scroll until the target is visible                        |