Skip to content

Open-source AI browser agent for Chrome and Firefox (monorepo) 🧠

1.7 KB
Raw Blame
{
  "id": "014",
  "category": "tool-error-pivot",
  "mode": "act",
  "browser": "chrome",
  "tab": {
    "url": "https://arxiv.org/pdf/2401.00001.pdf",
    "title": "2401.00001.pdf"
  },
  "description": "read_page returned PDF-tab notice. Pivot to read_pdf.",
  "seed": [
    {
      "role": "user",
      "content": "[Current page context β€” applies to this user message and supersedes older page context for phrases like \"this page\". URL: https://arxiv.org/pdf/2401.00001.pdf β€” Title: 2401.00001.pdf]\n\nsummarize this paper"
    },
    {
      "role": "assistant",
      "content": "",
      "tool_calls": [
        {
          "id": "call_h_1aeb28",
          "type": "function",
          "function": {
            "name": "read_page",
            "arguments": "{}"
          }
        }
      ]
    },
    {
      "role": "tool",
      "tool_call_id": "call_h_1aeb28",
      "name": "read_page",
      "content": "{\"success\":false,\"error\":\"This tab is a PDF β€” Firefox's built-in PDF viewer is a privileged page our content scripts cannot reach. Use read_pdf instead, which fetches the binary and extracts text directly.\"}"
    }
  ],
  "expected": {
    "idealNextToolCall": {
      "name": "read_pdf",
      "args": {
        "url": "https://arxiv.org/pdf/2401.00001.pdf"
      }
    },
    "antiPatterns": [
      {
        "match": "read_page({})",
        "reason": "Same tool that just rejected the PDF tab."
      },
      {
        "match": "inspect_viewport({})",
        "reason": "PDF viewer capture won't produce extractable text."
      }
    ],
    "successRubric": "Use read_pdf with the tab URL. The error message names the right tool β€” follow it."
  }
}