@mindstudio-ai/remy 0.1.58 → 0.1.60

package/dist/headless.js

@@ -2350,9 +2350,11 @@ async function captureAndAnalyzeScreenshot(promptOrOptions) {
   let prompt;
   let existingUrl;
   let onLog;
+  let path9;
   if (typeof promptOrOptions === "object" && promptOrOptions !== null) {
     prompt = promptOrOptions.prompt;
     existingUrl = promptOrOptions.imageUrl;
+    path9 = promptOrOptions.path;
     onLog = promptOrOptions.onLog;
   } else {
     prompt = promptOrOptions;
@@ -2361,9 +2363,11 @@ async function captureAndAnalyzeScreenshot(promptOrOptions) {
   if (existingUrl) {
     url = existingUrl;
   } else {
-    const ssResult = await sidecarRequest("/screenshot-full-page", void 0, {
-      timeout: 12e4
-    });
+    const ssResult = await sidecarRequest(
+      "/screenshot-full-page",
+      path9 ? { path: path9 } : void 0,
+      { timeout: 12e4 }
+    );
     url = ssResult?.url || ssResult?.screenshotUrl;
     if (!url) {
       throw new Error(
@@ -2398,6 +2402,10 @@ var screenshotTool = {
         imageUrl: {
           type: "string",
           description: "URL of an existing screenshot to analyze instead of capturing a new one. Use this for additional questions about a previous screenshot."
+        },
+        path: {
+          type: "string",
+          description: 'Navigate to this path before capturing (e.g. "/settings", "/dashboard"). If omitted, screenshots the current page.'
         }
       }
     }
@@ -2413,6 +2421,7 @@ var screenshotTool = {
       }
       return await captureAndAnalyzeScreenshot({
         prompt: input.prompt,
+        path: input.path,
         onLog: context?.onLog
       });
     } catch (err) {
@@ -2425,6 +2434,7 @@ var screenshotTool = {
 function startStatusWatcher(config) {
   const { apiConfig, getContext, onStatus, interval = 3e3, signal } = config;
   let lastLabel = "";
+  let lastContext = "";
   let inflight = false;
   let stopped = false;
   const url = `${apiConfig.baseUrl}/_internal/v2/agent/remy/generate-status`;
@@ -2435,9 +2445,10 @@ function startStatusWatcher(config) {
     inflight = true;
     try {
       const context = getContext();
-      if (!context) {
+      if (!context || context === lastContext) {
         return;
       }
+      lastContext = context;
       const res = await fetch(url, {
         method: "POST",
         headers: {
@@ -2512,6 +2523,16 @@ ${summaryBlock.text}
     startIdx = checkpointIdx + 1;
   }
   const messagesToProcess = messages.slice(startIdx);
+  const toolUseIds = /* @__PURE__ */ new Set();
+  for (const msg of messagesToProcess) {
+    if (msg.role === "assistant" && Array.isArray(msg.content)) {
+      for (const block of msg.content) {
+        if (block.type === "tool") {
+          toolUseIds.add(block.id);
+        }
+      }
+    }
+  }
   const cleaned = messagesToProcess.filter((msg) => {
     if (Array.isArray(msg.content)) {
       const blocks = msg.content;
@@ -2519,6 +2540,9 @@ ${summaryBlock.text}
         return false;
       }
     }
+    if (msg.role === "user" && msg.toolCallId && !toolUseIds.has(msg.toolCallId)) {
+      return false;
+    }
     return true;
   }).map((msg) => {
     if (msg.role === "user" && typeof msg.content === "string" && msg.content.startsWith("@@automated::")) {
@@ -2983,7 +3007,12 @@ var BROWSER_TOOLS = [
     description: "Capture a full-height screenshot of the current page. Returns a CDN URL with full text analysis and description.",
     inputSchema: {
       type: "object",
-      properties: {}
+      properties: {
+        path: {
+          type: "string",
+          description: 'Navigate to this path before capturing (e.g. "/settings"). If omitted, screenshots the current page.'
+        }
+      }
     }
   },
   {
@@ -3056,7 +3085,10 @@ var browserAutomationTool = {
       executeTool: async (name, _input, _toolCallId, onLog) => {
         if (name === "screenshotFullPage") {
           try {
-            return await captureAndAnalyzeScreenshot({ onLog });
+            return await captureAndAnalyzeScreenshot({
+              path: _input.path,
+              onLog
+            });
           } catch (err) {
             return `Error taking screenshot: ${err.message}`;
           }
@@ -3308,6 +3340,10 @@ var definition5 = {
       prompt: {
         type: "string",
         description: "Optional specific question about the screenshot."
+      },
+      path: {
+        type: "string",
+        description: 'Navigate to this path before capturing (e.g. "/settings"). If omitted, screenshots the current page.'
       }
     }
   }
@@ -3316,6 +3352,7 @@ async function execute5(input, onLog) {
   try {
     return await captureAndAnalyzeScreenshot({
       prompt: input.prompt,
+      path: input.path,
       onLog
     });
   } catch (err) {

package/dist/index.js

@@ -2056,9 +2056,11 @@ async function captureAndAnalyzeScreenshot(promptOrOptions) {
   let prompt;
   let existingUrl;
   let onLog;
+  let path10;
   if (typeof promptOrOptions === "object" && promptOrOptions !== null) {
     prompt = promptOrOptions.prompt;
     existingUrl = promptOrOptions.imageUrl;
+    path10 = promptOrOptions.path;
     onLog = promptOrOptions.onLog;
   } else {
     prompt = promptOrOptions;
@@ -2067,9 +2069,11 @@ async function captureAndAnalyzeScreenshot(promptOrOptions) {
   if (existingUrl) {
     url = existingUrl;
   } else {
-    const ssResult = await sidecarRequest("/screenshot-full-page", void 0, {
-      timeout: 12e4
-    });
+    const ssResult = await sidecarRequest(
+      "/screenshot-full-page",
+      path10 ? { path: path10 } : void 0,
+      { timeout: 12e4 }
+    );
     url = ssResult?.url || ssResult?.screenshotUrl;
     if (!url) {
       throw new Error(
@@ -2118,6 +2122,10 @@ var init_screenshot2 = __esm({
             imageUrl: {
               type: "string",
               description: "URL of an existing screenshot to analyze instead of capturing a new one. Use this for additional questions about a previous screenshot."
+            },
+            path: {
+              type: "string",
+              description: 'Navigate to this path before capturing (e.g. "/settings", "/dashboard"). If omitted, screenshots the current page.'
             }
           }
         }
@@ -2133,6 +2141,7 @@ var init_screenshot2 = __esm({
           }
           return await captureAndAnalyzeScreenshot({
             prompt: input.prompt,
+            path: input.path,
             onLog: context?.onLog
           });
         } catch (err) {
@@ -2147,6 +2156,7 @@ var init_screenshot2 = __esm({
 function startStatusWatcher(config) {
   const { apiConfig, getContext, onStatus, interval = 3e3, signal } = config;
   let lastLabel = "";
+  let lastContext = "";
   let inflight = false;
   let stopped = false;
   const url = `${apiConfig.baseUrl}/_internal/v2/agent/remy/generate-status`;
@@ -2157,9 +2167,10 @@ function startStatusWatcher(config) {
     inflight = true;
     try {
       const context = getContext();
-      if (!context) {
+      if (!context || context === lastContext) {
         return;
       }
+      lastContext = context;
       const res = await fetch(url, {
         method: "POST",
         headers: {
@@ -2239,6 +2250,16 @@ ${summaryBlock.text}
     startIdx = checkpointIdx + 1;
   }
   const messagesToProcess = messages.slice(startIdx);
+  const toolUseIds = /* @__PURE__ */ new Set();
+  for (const msg of messagesToProcess) {
+    if (msg.role === "assistant" && Array.isArray(msg.content)) {
+      for (const block of msg.content) {
+        if (block.type === "tool") {
+          toolUseIds.add(block.id);
+        }
+      }
+    }
+  }
   const cleaned = messagesToProcess.filter((msg) => {
     if (Array.isArray(msg.content)) {
       const blocks = msg.content;
@@ -2246,6 +2267,9 @@ ${summaryBlock.text}
         return false;
       }
     }
+    if (msg.role === "user" && msg.toolCallId && !toolUseIds.has(msg.toolCallId)) {
+      return false;
+    }
     return true;
   }).map((msg) => {
     if (msg.role === "user" && typeof msg.content === "string" && msg.content.startsWith("@@automated::")) {
@@ -2729,7 +2753,12 @@ var init_tools = __esm({
         description: "Capture a full-height screenshot of the current page. Returns a CDN URL with full text analysis and description.",
         inputSchema: {
           type: "object",
-          properties: {}
+          properties: {
+            path: {
+              type: "string",
+              description: 'Navigate to this path before capturing (e.g. "/settings"). If omitted, screenshots the current page.'
+            }
+          }
         }
       },
       {
@@ -2865,7 +2894,10 @@ var init_browserAutomation = __esm({
           executeTool: async (name, _input, _toolCallId, onLog) => {
             if (name === "screenshotFullPage") {
               try {
-                return await captureAndAnalyzeScreenshot({ onLog });
+                return await captureAndAnalyzeScreenshot({
+                  path: _input.path,
+                  onLog
+                });
               } catch (err) {
                 return `Error taking screenshot: ${err.message}`;
               }
@@ -3143,6 +3175,7 @@ async function execute5(input, onLog) {
   try {
     return await captureAndAnalyzeScreenshot({
       prompt: input.prompt,
+      path: input.path,
       onLog
     });
   } catch (err) {
@@ -3163,6 +3196,10 @@ var init_screenshot3 = __esm({
           prompt: {
             type: "string",
             description: "Optional specific question about the screenshot."
+          },
+          path: {
+            type: "string",
+            description: 'Navigate to this path before capturing (e.g. "/settings"). If omitted, screenshots the current page.'
           }
         }
       }

package/dist/prompt/compiled/interfaces.md

@@ -26,12 +26,16 @@ dist/interfaces/web/
 
 ```json
 {
-  "devPort": 5173,
-  "devCommand": "npm run dev",
-  "defaultPreviewMode": "desktop"
+  "web": {
+    "devPort": 5173,
+    "devCommand": "npm run dev",
+    "defaultPreviewMode": "desktop"
+  }
 }
 ```
 
+All fields are nested under the `"web"` key.
+
 | Field | Type | Default | Description |
 |-------|------|---------|-------------|
 | `devPort` | `number` | `5173` | Port for the dev server |
@@ -99,7 +103,7 @@ Omit the `methods` field (or the config entirely) to expose all methods.
 ### Usage
 
 ```bash
-curl -X POST https://api.mindstudio.ai/_internal/v2/apps/{appId}/methods/submit-vendor-request/invoke \
+curl -X POST https://{app-subdomain}.mindstudio.ai/_/methods/submit-vendor-request/invoke \
   -H "Authorization: Bearer sk..." \
   -H "Content-Type: application/json" \
   -d '{ "input": { "name": "Acme" } }'
@@ -200,7 +204,7 @@ Inbound HTTP endpoints that invoke methods.
 }
 ```
 
-Endpoint URL: `https://api.mindstudio.ai/_internal/v2/webhook/{appId}/{secret}`
+Endpoint URL: `https://{app-subdomain}.mindstudio.ai/_/webhook/{secret}`
 
 Accepts any HTTP method. The method receives `{ method, headers, query, body }` as input.
 

package/dist/prompt/static/team.md

@@ -38,7 +38,9 @@ Always consult the code sanity check before writing code in initialCodegen with
 
 ### QA (`runAutomatedBrowserTest`)
 
-For verifying complex stateful interactions: multi-step form submissions, auth flows, real-time updates, flows that require specific data/role setup. This spins up a full chrome browser automation — it's heavyweight. Do not use it for basic rendering or navigation checks. If you can verify something with a screenshot or by reading the code, do that instead. Run a scenario first to seed test data and set user roles. The user is able to watch QA work on their screen via a live browser preview - the cursor will move, type, etc - so you can also use this to demo functionality to the user and help them understand how to use their app. Don't tell the QA agent what values it should use for testing, it will figure that out on its own.
+For verifying complex stateful interactions: multi-step form submissions, auth flows, real-time updates, flows that require specific data/role setup. This spins up a full chrome browser automation — it's heavyweight. Do not use it for basic rendering or navigation checks. If you can verify something with a screenshot or by reading the code, do that instead. Run a scenario first to seed test data and set user roles. The user is able to watch QA work on their screen via a live browser preview - the cursor will move, type, etc - so you can also use this to demo functionality to the user and help them understand how to use their app.
+
+The QA agent can see the screen. Describe what to test, not how — it will figure out what to click, what to check, and what values to use.
 
 ### Background Execution
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mindstudio-ai/remy",
-  "version": "0.1.58",
+  "version": "0.1.60",
   "description": "MindStudio coding agent",
   "repository": {
     "type": "git",