vscode

The VS Code team details recent optimizations to GitHub Copilot's agentic harness, such as prompt caching and tool search, to improve token efficiency and reduce costs under usage-based billing.

The author explains the motivation behind PEP 832, which aims to standardize virtual environment discovery for tools like VS Code, addressing problems of knowing which environments exist and which tool created them.

Qwen code companion is now available on the VS Code marketplace, offering an AI-powered coding assistant for developers.

A personal reflection on the author's decades-long relationship with Emacs, including switching to VSCode and IntelliJ, and ultimately returning to Emacs for its unique features.

A security researcher discloses a critical vulnerability in VSCode's webview that allows attackers to steal full-access GitHub OAuth tokens by luring users to click a link. The bug affects the github.dev web editor.

The blogger reminds users that the Cline extension on VSCode can use free models, especially the newly released Qwen 3.7 Max which offers 20M free tokens daily with 1M context length.

VSCode team released a series of videos on Agent-First Development, introducing the five pillars of AI programming: model selection, action boundaries, context, prompt precision, and tool usage, to help developers use AI agents more effectively.

A hacker group called TeamPCP is conducting an unprecedented wave of software supply chain attacks, compromising hundreds of open source tools and breaching companies including GitHub, Anthropic, and Mercor.

GitHub confirmed that a malicious VS Code extension installed by an employee led to the breach of approximately 3,800 internal repositories. The attacker group TeamPCP claimed responsibility and is attempting to sell the stolen data.

GitHub has made remote control for GitHub Copilot CLI and VS Code sessions generally available, allowing users to start work on one machine and continue seamlessly from another.

Microsoft open-sourced the VS Code extension AI Toolkit, which integrates model selection, Playground, Agent building, batch testing, and evaluation into one interface, simplifying AI Agent development process.

VS Code introduces a new 'Agents' feature that allows developers to work with multiple agents across multiple projects, expanding its capabilities beyond single-task workflows.