Fake updates to Microsoft’s AI agent toolkit stole developer passwords
Two malicious versions of agent-framework, the Python package for Microsoft’s AI agent toolkit, were published to the PyPI software registry on Tuesday…
November 17, 2026 · Heads up
Two malicious versions of agent-framework, the Python package for Microsoft’s AI agent toolkit, were published to the PyPI software registry on Tuesday evening and removed within about an hour, security firm Wiz reports. Versions 1.26.1 and 1.26.2 contained code that ran every time Python started and sent out passwords and access keys for Azure, GitHub, OpenAI and other services. The package is downloaded about 41,000 times a week. Wiz says version 1.26.0 is the last safe release and that how the bad versions were uploaded is not yet confirmed.
Why it matters to you: AI agent tools hold the keys to a lot of systems, which makes them a prime target. If your team builds with this package, check whether 1.26.1 or 1.26.2 was installed anywhere, pin to 1.26.0, and change every password and key on any affected machine or build server.
Source: Wiz
More from Today in AI
- Nvidia revenue doubles to US$115.8B in first results without Huang · November 17, 2026
- Microsoft launches MAI Code 2, its own coding AI, in GitHub Copilot · November 17, 2026
- California sues OpenAI over AI agents it says escaped control · November 16, 2026
- US moves to stop China renting top AI chips through the cloud · November 16, 2026
- OpenAI launches AI agents that work in hospital records systems · November 16, 2026
All of Today in AI. Where AI is heading.
Related Power of AI pages
Keep reading with Start here, everyday uses, the tool guide, the writing workshop, and sources and standards.