Key Signals

• Model Context Protocol adoption accelerates across AI coding tools. Penpot's experimental MCP server implementation demonstrates how the protocol enables secure AI-powered workflows without exposing user data to third-party LLMs, allowing AI assistants like Claude to perform complex operations such as design-to-code conversion, component generation, and design system management through a structured API bridge. This approach represents a significant shift from the less sophisticated "describe and generate" workflows, offering LLM-agnostic integration across tools like Claude in VS Code, Cursor, and Claude Desktop. The MCP standard is emerging as a critical infrastructure layer for connecting AI capabilities to development tools while maintaining data privacy and security. [1]

• Cursor CLI gains comprehensive model and server management capabilities. The January 8 release introduces command-line controls for listing and switching between AI models via agent models, --list-models, and /models commands, alongside new /rules functionality for creating and editing coding rules directly from the terminal. The addition of /mcp enable and /mcp disable commands provides on-the-fly control over Model Context Protocol servers, while major hooks performance improvements address a critical bottleneck in CLI workflows. These enhancements signal Cursor's strategic investment in making its agentic coding capabilities accessible beyond the IDE, targeting power users and automation scenarios. [2]

• AWS Transform custom introduces agentic AI for large-scale code modernization. Amazon's new AWS Transform CLI leverages agentic AI to automate complex refactoring tasks like AWS SDK v1 to v2 migrations, reducing what traditionally requires weeks of manual work to minutes while maintaining code quality through automated build verification and transformation planning. The tool uses AI to analyze codebases, generate comprehensive transformation plans with detailed change proposals, and execute modifications based on user feedback and organizational requirements such as specific library versions. This represents a strategic move by AWS to position agentic AI as a critical tool for reducing technical debt at enterprise scale, with support for language upgrades, framework migrations, and organization-specific code patterns. [3]

Feature Update

• Cursor CLI introduces model management, MCP controls, and rules generation. The January 8, 2026 release adds three major command groups: model selection via agent models, --list-models, and /models slash commands for quickly switching between available AI models; /rules command for creating and editing project-specific coding rules directly from the terminal; and /mcp enable//mcp disable commands for runtime control of Model Context Protocol servers. The update also delivers major performance improvements to the hooks system, addressing latency issues in CLI workflows and bug fixes across the platform. [2]

• AWS Transform custom CLI enables automated SDK migrations and code modernization. AWS released its Transform custom tool powered by agentic AI to automate large-scale code transformations, demonstrated through AWS SDK for Java v1 to v2 migrations. The CLI provides AWS-Managed Transformations including AWS/java-aws-sdk-v1-to-v2 that analyze projects, generate comprehensive transformation plans with specific change details, accept user feedback and version requirements, verify builds automatically, and produce validation summaries. Key capabilities include support for Maven dependency updates, API migration patterns, builder pattern implementations, exception handling updates, and custom transformation definitions for organization-specific requirements. The tool accelerates migration timelines from weeks to minutes while reducing human error and ensuring code quality through automated verification. [3]

AI Coding News

• Penpot launches MCP server beta for AI-powered design workflows. The open-source design tool is experimenting with Model Context Protocol servers that enable AI assistants to interact with Penpot files through secure API bridges, supporting use cases like design-to-code conversion, code-to-design translation, design system documentation generation, component creation from scribbles, and automated design consistency checking. The implementation leverages Penpot's design-expressed-as-code architecture for granular programmatic control, includes a Python SDK, REST API, plugin system, and CLI tools, and works with any MCP-enabled AI assistant including Claude in VS Code, Cursor, and Claude Desktop. Penpot is actively seeking beta testers and community feedback to shape the development of what CEO Pablo Ruiz-Múzquiz describes as a more refined and adaptable approach compared to traditional "describe and generate" or "convert to code" AI workflows. [1]