MCP Data Fetch Server

""" Tool definitions for the MCP server. """ from models import Tool def get_tool_definitions(): """Return all tool definitions.""" return [ Tool( name="fetch_webpage", description="Securely fetch and parse web pages. Returns clean content in markdown/text/html, handles cookies and banners. NO JavaScript execution.", input_schema={ "type": "object", "properties": { "url": { "type": "string", "description": "URL to fetch (http/https only)" }, "format": { "type": "string", "enum": ["markdown", "text", "html"], "description": "Output format", "default": "markdown" }, "include_links": { "type": "boolean", "description": "Include extracted links", "default": True }, "include_images": { "type": "boolean", "description": "Include image URLs", "default": False }, "remove_banners": { "type": "boolean", "description": "Attempt to remove cookie banners and popups", "default": True } }, "required": ["url"], }, ), Tool( name="extract_links", description="Extract and categorize all links from a webpage (internal, external, resources).", input_schema={ "type": "object", "properties": { "url": { "type": "string", "description": "URL to analyze" }, "filter": { "type": "string", "enum": ["all", "internal", "external", "resources"], "description": "Filter link types", "default": "all" } }, "required": ["url"], }, ), Tool( name="download_file", description="Safely download files with size limits and type validation. Sanitizes filenames to prevent path traversal.", input_schema={ "type": "object", "properties": { "url": { "type": "string", "description": "File URL" }, "filename": { "type": "string", "description": "Optional filename (will be sanitized)" } }, "required": ["url"], }, ), Tool( name="get_page_metadata", description="Extract structured metadata (title, description, Open Graph, Twitter Cards, schema.org).", input_schema={ "type": "object", "properties": { "url": { "type": "string", "description": "URL to analyze" } }, "required": ["url"], }, ), Tool( name="check_url", description="Check URL status without full download (HEAD request). Returns status code, headers, size.", input_schema={ "type": "object", "properties": { "url": { "type": "string", "description": "URL to check" } }, "required": ["url"], }, ), ]