This is a VSCode extension, not an MCP server. The source shows only package.json with two command definitions ('linggen.openDashboard', 'linggen.bridgeSettings') that lack input schemas, parameter descriptions, and output documentation. No MCP protocol implementation is visible. Tool definitions are inferred from VSCode command metadata only, not explicit MCP tool registration. The extension bridges ling-mem memory to AI agents via file injection, but exposes no MCP tools for remote agents to call.

• Wrap this VSCode extension as an MCP server using stdio or HTTP transport so remote agents can invoke the bridge and dashboard tools.
• Add comprehensive input schemas for both tools. 'openDashboard' should document host/port parameters; 'bridgeSettings' should document which agents to bridge to (enum: claude, cursor, copilot, codex, windsurf, zed).
• Expand tool descriptions to 50 - 200 characters. E.g., 'Open the ling-mem dashboard (http://host:port) to view and manage cross-session memory. Requires ling-mem daemon running.' and 'Configure which AI agents (Claude, Cursor, Copilot, etc.) should read your ling-mem memory via bridged files.'
• Document output schemas. 'openDashboard' should return {url, status, message}; 'bridgeSettings' should return {applied_targets: string[], skipped_targets: string[], errors: {target: string, reason: string}[]}.
• Add error handling with recovery hints. E.g., 'Dashboard host unreachable. Verify ling-mem daemon is running on 127.0.0.1:9528 or update linggen.dashboard.host/port in settings.'