Enable support for Http based MCP resources

[dileepyavan]

fixes #259253

Screen.Recording.2025-11-10.123107.mp4

[Copilot]

Pull Request Overview

This PR enables support for HTTP-based MCP (Model Context Protocol) resources by adding validation and web content extraction capabilities. The changes allow VS Code to fetch and display HTTP/HTTPS resources from MCP servers through the web content extractor service.

Key Changes

• Adds ValidateHttpResources function to validate HTTP/HTTPS resource requests against MCP server configurations
• Modifies McpResourceFilesystem to use web content extraction for HTTP-based resources instead of MCP server calls
• Updates directory detection logic to use path-based checking instead of content-based checking

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 10 comments.

File	Description
src/vs/workbench/contrib/mcp/common/mcpTypesUtils.ts	Adds new ValidateHttpResources function to validate HTTP resource URIs against MCP server transport type and authority
src/vs/workbench/contrib/mcp/common/mcpResourceFilesystem.ts	Implements HTTP resource extraction via IWebContentExtractorService, updates directory detection logic, and adds helper function for path checking
src/vs/workbench/contrib/mcp/browser/mcpResourceQuickAccess.ts	Adds trailing newline (whitespace-only change)
src/vs/workbench/contrib/chat/browser/actions/chatContextActions.ts	Changes attachment handling to return early instead of adding context (unrelated to MCP HTTP resources)

[Copilot]

Incorrect import: should use URI from '../../../../base/common/uri.js' instead of Uri from 'vscode'. The VS Code extension API (vscode) should not be imported in workbench code. All other files in the mcp module use URI from base/common/uri.js.

[Copilot]

The variable isResourceRequestValid can be simplified. Instead of initializing to false and conditionally setting to true, consider using a direct return statement. This reduces unnecessary state and improves readability.

Suggested change

	let isResourceRequestValid = false;
	if (resource.path.indexOf('/http/') === 0) {
	const launch = server?.connection.get()?.launchDefinition;
	if (launch && launch.type === McpServerTransportType.HTTP && launch.uri.authority.toLowerCase() === resource.authority.toLowerCase()) {
	isResourceRequestValid = true;
	}
	} else if (resource.path.indexOf('/https/') === 0) {
	isResourceRequestValid = true;
	}
	return isResourceRequestValid;
	if (resource.path.indexOf('/http/') === 0) {
	const launch = server?.connection.get()?.launchDefinition;
	return !!(launch && launch.type === McpServerTransportType.HTTP && launch.uri.authority.toLowerCase() === resource.authority.toLowerCase());
	} else if (resource.path.indexOf('/https/') === 0) {
	return true;
	}
	return false;

[Copilot]

Potential bug: when web content extraction returns errors or redirects, the code creates empty text content (text: ''). This masks extraction failures from the caller. Consider propagating errors appropriately or handling non-ok results differently - for example, checking if all results failed and throwing an appropriate error.