fix(context-dev): wire includeFrames, split crawl/extract maxPages, derive screenshot MIME

Addresses review feedback:
- includeFrames is now a block subblock + param for scrape_markdown/scrape_html
- crawl and extract use separate Max Pages fields (crawl 1-500, extract 1-50) so a
  crawl value can no longer be forwarded to extract beyond its limit
- screenshot file MIME type and extension are derived from the returned URL instead
  of being hardcoded to PNG

Author: waleedlatif1

apps/sim/blocks/blocks/context_dev.ts

@@ -238,6 +238,13 @@ Do not include any explanations, markdown formatting, or other text outside the
       type: 'switch',
       condition: { field: 'operation', value: ['scrape_markdown', 'crawl'] },
     },
+    {
+      id: 'includeFrames',
+      title: 'Include Frames',
+      type: 'switch',
+      mode: 'advanced',
+      condition: { field: 'operation', value: ['scrape_markdown', 'scrape_html'] },
+    },
     {
       id: 'fullScreenshot',
       title: 'Full Page Screenshot',
@@ -328,9 +335,17 @@ Do not include any explanations, markdown formatting, or other text outside the
       id: 'maxPages',
       title: 'Max Pages',
       type: 'short-input',
-      placeholder: 'crawl: 100, extract: 5',
+      placeholder: '100',
       mode: 'advanced',
-      condition: { field: 'operation', value: ['crawl', 'extract'] },
+      condition: { field: 'operation', value: 'crawl' },
+    },
+    {
+      id: 'extractMaxPages',
+      title: 'Max Pages',
+      type: 'short-input',
+      placeholder: '5',
+      mode: 'advanced',
+      condition: { field: 'operation', value: 'extract' },
     },
     {
       id: 'maxDepth',
@@ -560,13 +575,15 @@ Do not include any explanations, markdown formatting, or other text outside the
             setBool('useMainContentOnly')
             setBool('includeLinks')
             setBool('includeImages')
+            setBool('includeFrames')
             setNumber('maxAgeMs')
             setNumber('waitForMs')
             setNumber('timeoutMS')
             break
           case 'scrape_html':
             setString('url')
             setBool('useMainContentOnly')
+            setBool('includeFrames')
             setNumber('maxAgeMs')
             setNumber('waitForMs')
             setNumber('timeoutMS')
@@ -638,7 +655,7 @@ Do not include any explanations, markdown formatting, or other text outside the
             setString('instructions')
             setBool('factCheck')
             setBool('followSubdomains')
-            setNumber('maxPages')
+            setNumber('extractMaxPages', 'maxPages')
             setNumber('maxDepth')
             setNumber('maxAgeMs')
             setNumber('stopAfterMs')
@@ -755,6 +772,7 @@ Do not include any explanations, markdown formatting, or other text outside the
     useMainContentOnly: { type: 'boolean', description: 'Return only main content' },
     includeLinks: { type: 'boolean', description: 'Preserve hyperlinks' },
     includeImages: { type: 'boolean', description: 'Include image references' },
+    includeFrames: { type: 'boolean', description: 'Render iframe contents inline' },
     fullScreenshot: { type: 'boolean', description: 'Capture the full page' },
     handleCookiePopup: { type: 'boolean', description: 'Dismiss cookie banners' },
     markdownEnabled: { type: 'boolean', description: 'Scrape search results to markdown' },
@@ -766,7 +784,11 @@ Do not include any explanations, markdown formatting, or other text outside the
     queryFanout: { type: 'boolean', description: 'Expand query into variants' },
     factCheck: { type: 'boolean', description: 'Ground extracted values in page facts' },
     followSubdomains: { type: 'boolean', description: 'Follow subdomain links' },
-    maxPages: { type: 'number', description: 'Maximum pages to process' },
+    maxPages: { type: 'number', description: 'Maximum pages to crawl (1-500)' },
+    extractMaxPages: {
+      type: 'number',
+      description: 'Maximum pages to analyze for extraction (1-50)',
+    },
     maxDepth: { type: 'number', description: 'Maximum link depth' },
     maxProducts: { type: 'number', description: 'Maximum products to extract' },
     urlRegex: { type: 'string', description: 'Regex to filter URLs' },

apps/sim/tools/context_dev/screenshot.ts

@@ -12,6 +12,32 @@ import {
 } from '@/tools/context_dev/utils'
 import type { ToolConfig, ToolFileData } from '@/tools/types'
 
+/** Maps a lowercase image file extension to its MIME type. */
+const IMAGE_MIME_BY_EXTENSION: Record<string, string> = {
+  png: 'image/png',
+  jpg: 'image/jpeg',
+  jpeg: 'image/jpeg',
+  webp: 'image/webp',
+  gif: 'image/gif',
+  avif: 'image/avif',
+}
+
+/**
+ * Derives the file extension and MIME type for a stored screenshot from its URL,
+ * falling back to PNG when the URL has no recognizable image extension.
+ */
+function screenshotFileMeta(url: string): { extension: string; mimeType: string } {
+  try {
+    const ext = new URL(url).pathname.split('.').pop()?.toLowerCase() ?? ''
+    if (IMAGE_MIME_BY_EXTENSION[ext]) {
+      return { extension: ext, mimeType: IMAGE_MIME_BY_EXTENSION[ext] }
+    }
+  } catch {
+    // Unparseable URL — fall back to the default below.
+  }
+  return { extension: 'png', mimeType: 'image/png' }
+}
+
 export const contextDevScreenshotTool: ToolConfig<
   ContextDevScreenshotParams,
   ContextDevScreenshotResponse
@@ -100,10 +126,11 @@ export const contextDevScreenshotTool: ToolConfig<
     const screenshotUrl: string = data.screenshot ?? ''
     const domain: string | null = data.domain ?? null
 
+    const { extension, mimeType } = screenshotFileMeta(screenshotUrl)
     const file: ToolFileData | undefined = screenshotUrl
       ? {
-          name: `${domain ?? 'screenshot'}.png`,
-          mimeType: 'image/png',
+          name: `${domain ?? 'screenshot'}.${extension}`,
+          mimeType,
           url: screenshotUrl,
         }
       : undefined