The Archive

Anthropic's VS Code extension on Cursor marketplace was compromised for ~12 hours; attack vector remains unclear.

Cursor marketplace Claude Code VS Code extension contained embedded backdoor vulnerability.

Security backdoor discovered in Claude Code VS Code extension on Cursor Marketplace.

Learn how to move an agent session from the GitHub Copilot app into VS Code without losing the conversation, code changes, or working context. The post How to continue a GitHub Copilot app session in VS Code appeared first on The GitHub Blog.

Microsoft announced two new text LLMs this morning - MAI-Thinking-1 (reasoning, 35B parameters, available to "select early partners") and MAI-Code-1-Flash (5B parameters, "purpose-built for GitHub Copilot and VS Code to deliver high performance and lower cost [...] rolling out to GitHub Copilot individual users in Visual Studio Code"). I've not been able to try either of them just yet. It's very interesting to see Microsoft releasing models with such low parameter counts, especially given how expensive larger models are to access right now. They claim MAI-Thinking-1 "is preferred to Sonnet 4....

Mistral launches Vibe, a unified agent for long-horizon productivity and coding tasks with Work/Code modes and VS Code integration.

*Everything I learned the hard way — 6 weeks, no sleep :), two environments, one agent that actually works.* # The Story I spent six weeks building a personal AI agent from scratch — not a chatbot wrapper, but a persistent assistant that manages tasks, tracks deals, reads emails, analyzes business data, and proactively surfaces things I'd otherwise miss. It started in the cloud (Claude Projects — shared memory files, rich context windows, custom skills). Then I migrated to Claude Code inside VS Code, which unlocked local file access, git tracking, shell hooks, and scheduled headless tasks....

VS Code adds Agents window for local AI models but requires internet + GitHub Copilot subscription, limiting offline/open-source utility.

User reports successful local inference with Qwen3.6-35B on AMD R9700 GPU, generating functional code and tests via llama-cpp.

Empirical comparison of Qwen3.6-27B and Coder-Next models across 40 test cases shows statistical parity with task-dependent tradeoffs.