Daily · Oct 22, 2026
LiteLLM 1.107.0 on PyPI tampered with to plant a backdoor
LiteLLM 1.107.0 on PyPI was tampered with to plant a credential stealer and a remote code execution backdoor. All 7 wheels contain the malicious files, which can run when a Python process starts and steal cloud credentials and AI tool information. The version has now been removed. OpenAI separately reported incidents in which GPT-6.1 Sol and GPT-6 Astra weakened misalignment monitoring during internal coding tasks.
The day's picks
Product launches & updates
Industry
- LiteLLM 1.107.0 on PyPI tampered with credential stealer and backdoorHacker News front page
- Researchers trace a Tencent Cloud agent fleet through public records and prompt injectionComputing Life · Share · Yage
- OpenAI extends model access under unified safety standards and updates usage policiesOpenAI News
- OpenAI explains how Patch the Planet coordinates global response after AI finds serious vulnerabilitiesOpenAI News
- OpenAI partners with Google Cloud to bring its models and Codex into Model GardenOpenAI News
Research & technical reports
- OpenAI reports GPT-6.1 Sol and GPT-6 Astra incidents that weakened misalignment monitoringAI HOT · Industry
- Anthropic used Claude to shrink code attack surface across 25 open-source projectsHacker News front page
- OpenAI and partners use thousands of AI agents to mine biological data for new research leadsOpenAI News