"""Content analysis prompts for MCP server."""
from __future__ import annotations
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from mcp.server.fastmcp import FastMCP
def register_analysis_prompts(mcp: FastMCP) -> None:
"""Register content analysis prompts on the MCP server.
Args:
mcp: The FastMCP server instance
"""
@mcp.prompt(title="Analyze Webpage")
def analyze_webpage(url: str, focus: str = "main_content") -> str:
"""Generate a structured analysis of a webpage.
Args:
url: The URL to analyze
focus: Analysis focus area (main_content, navigation, metadata, media, forms)
"""
return f"""Analyze the webpage at {url}.
## Instructions
1. Scrape the page: `scrape_url(urls=["{url}"])`
2. Review the markdown content
3. Focus your analysis on: **{focus}**
## Analysis Structure
Provide findings in this format:
### Page Overview
- URL: {url}
- Main purpose/topic
- Target audience
### Content Analysis ({focus})
- Key sections identified
- Main messages/themes
- Notable elements (CTAs, forms, media)
### Technical Observations
- Page structure quality
- Content organization
- Any issues noticed
## Focus Options
- `main_content`: Primary page content
- `navigation`: Site structure and links
- `metadata`: SEO tags and meta info
- `media`: Images, videos, embeds
- `forms`: Input fields and actions"""
@mcp.prompt(title="Summarize Content")
def summarize_content(url: str, length: str = "medium", style: str = "informative") -> str:
"""Generate a content summary with specified length and style.
Args:
url: The URL to summarize
length: Summary length (short, medium, long)
style: Writing style (informative, technical, casual)
"""
lengths = {
"short": "2-3 sentences",
"medium": "1-2 paragraphs",
"long": "comprehensive multi-paragraph",
}
return f"""Summarize the content at {url}.
## Instructions
1. Scrape: `scrape_url(urls=["{url}"], css_selector="article, main, .content")`
2. Generate a {lengths.get(length, lengths["medium"])} summary
3. Style: {style}
## Output Format
### Summary ({length})
[Your summary here]
### Key Points
- Point 1
- Point 2
- Point 3
### Source
- URL: {url}
- Scraped: [timestamp]"""
@mcp.prompt(title="Extract Data")
def extract_data(url: str, data_type: str = "general", selector: str = "") -> str:
"""Extract specific types of data from a webpage.
Args:
url: The URL to extract data from
data_type: Type of data (contact, pricing, dates, products, people, general)
selector: Optional CSS selector to target specific elements
"""
selector_hint = f', css_selector="{selector}"' if selector else ""
return f"""Extract {data_type} data from {url}.
## Instructions
1. Scrape: `scrape_url(urls=["{url}"]{selector_hint})`
2. Identify and extract: **{data_type}**
## Data Types
- `contact`: emails, phones, addresses
- `pricing`: prices, plans, costs
- `dates`: events, schedules, deadlines
- `products`: names, descriptions, specs
- `people`: names, titles, bios
- `general`: key facts and figures
## Output Format
### Extracted {data_type.title()} Data
| Field | Value | Confidence |
|-------|-------|------------|
| ... | ... | high/medium/low |
### Source Context
[Where this data was found on the page]"""
@mcp.prompt(title="Compare Pages")
def compare_pages(urls: str) -> str:
"""Compare content across multiple webpages.
Args:
urls: Comma-separated list of URLs to compare
"""
url_list = [u.strip() for u in urls.split(",")]
scrape_calls = "\n".join([f'- `scrape_url(urls=["{u}"])`' for u in url_list])
url_bullets = "\n".join([f"- {u}" for u in url_list])
return f"""Compare the following pages:
{url_bullets}
## Instructions
1. Scrape each page:
{scrape_calls}
2. Analyze each page's content
3. Compare across dimensions below
## Comparison Dimensions
- **Content**: Topics, depth, quality
- **Structure**: Organization, navigation
- **Purpose**: Goals, target audience
- **Features**: Unique elements, CTAs
## Output Format
### Summary Table
| Dimension | {" | ".join([f"Page {i + 1}" for i in range(len(url_list))])} |
|-----------|{"|".join(["---" for _ in url_list])}|
### Key Differences
[Bullet points]
### Key Similarities
[Bullet points]
### Recommendation
[Which page is better for what purpose]"""
