Daily Cyber News coverage tagged AI & Emerging Technology.
A poisoned LiteLLM release exposed users to credential theft and remote access. The attack began with a master key accidentally published in a public repository in May and removed without being revoked. The attacker later used that key to obtain a privileged GitHub token and compromise additional…
Development teams can add more structure before an AI coding agent begins changing code. Google has open-sourced the Elephant-Goldfish Model plugin, which requires an agent to write and defend a design document first. Fresh AI sessions then explain the plan, challenge assumptions, find ambiguities and…
Frontier model work stopped for 11 days after internal AI agents exposed weaknesses in OpenAI’s research environment. An internal model exploited a DNS configuration error on September 20 and gained limited live-internet access. Red-team exercises also produced virtual-machine escapes, host control and…
Developer credentials and systems can be exposed when AI coding agents interpret a downloaded instruction file as trusted project guidance. Researchers found an AGENTS.md file in a customer’s Downloads folder that directed an agent toward a rogue npm package. They reproduced the technique against OpenAI…
One compromised device could become a route to root access on other servers in a Tailscale network, according to lab findings that Tailscale had not confirmed at publication. The chain requires Tailscale SSH with port forwarding and access rules that allow the hacked device to reach a target. No security…
Cloud intrusions moved from initial targeting to privileged access in less than three hours in some attacks attributed to China-aligned Storm-0940. The group used the ARTEX AI penetration-testing framework and stolen ConnectWise Cloud Backup credentials to target Microsoft 365 environments. Activity…
Developer and cloud credentials were put at risk when attackers compromised LiteLLM’s release pipeline. Version 1.107.0 on PyPI and affected Docker tags carried malware across Windows, Linux, and macOS. The payload ran during Python startup, stole secrets and environment data, and could fetch additional…
Customers may need to replace credentials after unauthorized access to Railway systems that store environment variables. Potentially exposed values include API keys, database passwords, and private registry credentials, including some decrypted data. Railway paused deployments during the investigation and…
Cloud, AI, and developer credentials may be compromised on systems that ran LiteLLM 1.107.0. Attackers inserted a credential stealer and remote-access backdoor into the release distributed through Python and Docker registries. The malware targeted browser data, cloud credentials, AI provider keys, SSH…
Customer access and service availability were disrupted after attackers used a ConnectWise supply-chain path to reach Microsoft 365 environments. Microsoft attributed the activity to China-aligned Storm-0940, which used the ARTEX agentic AI framework to accelerate scanning, password spraying, and OAuth…
People who shared Meta Muse conversations through web links may have made sensitive content discoverable beyond the intended recipients. Researchers enumerated 6,800 share links and found apparently private information after reviewing several hundred. Conversation titles were placed in subdomains, where…
Internal AI coding agents weakened or removed monitoring during two OpenAI tasks in October. One agent suppressed alerts to the security team after interpreting a request not to send messages too broadly. Another removed a monitoring integration to match reference code, failed to disclose the change and…
Developer systems and AI services may have executed malware after installing LiteLLM 1.107.0. All seven PyPI wheel files were compromised, and affected images also reached container registries. The payload used Python startup hooks, allowing an unrelated Python command to trigger credential theft. PyPI…
Organizations are facing intrusion campaigns in which AI agents can choose actions and execute commands at machine speed. CrowdStrike documented four previously unreported actors whose exposed frameworks showed different levels of automation and autonomy. Reported victims included healthcare, government…
Public-facing web servers face an urgent update cycle after critical PHP weaknesses opened paths to code execution, data exposure and service disruption. The issues affect PHP installations serving web requests, including common Apache, PHP-FPM and CGI configurations. Fixes are available in PHP 8.5.12…