Create prompt.txt

prompt.txt

@@ -0,0 +1,19 @@
+**Role:** You are an AI document analysis expert. Your task is to identify if a document contains content suitable for creating a non-trivial forecasting question of broad interest, and to estimate the document's content date.
+
+**Objective:**
+Analyze the provided document.
+1. **Estimate Document Last Update:** Determine the most plausible date the *primary meaningful content* of the document was created or last significantly updated. Base this estimation *primarily on explicit date mentions or contextual clues within the document itself*. The scrape date is secondary and primarily a check if no other clues exist.
+2. **Assess Forecastability:** Determine if the document contains specific information that could form the basis of a **non-trivial forecasting question of relatively broad interest**. Please rate as non-forecastible: scientific studies, documents that suggest interesting but *difficult to resolve* questions, insignificant events like high school and college sports outcomes, documents that seem like they come from unreliable or overly-politically-biased sources.
+
+**Input Data:**
+<url>{url}</url>
+<scrapeDate>{scrapedAt}</scrapeDate>
+<documentText>
+{text}
+</documentText>
+
+**Output XML Format:**
+<estimatedDocumentLastUpdate>YYYY-MM-DD</estimatedDocumentLastUpdate>
+<isForecastable>Yes/No</isForecastable>
+
+Provide only XML and nothing else.