Malicious AGENTS Files Steer AI Coding Agents Toward Malware
A malicious AGENTS.md file triggered a blocked malware install. Researchers tested Codex and Claude Code attacks and urged checks on agent files and settings.
Read more →Latest cybersecurity news, vulnerabilities, and threat intelligence
A malicious AGENTS.md file triggered a blocked malware install. Researchers tested Codex and Claude Code attacks and urged checks on agent files and settings.
Read more →
LiteLLM investigators traced a poisoned Python release to an unrevoked master key, stolen GitHub tokens and abused workflows. Exposed secrets require rotation.
Read more →
Railway says secrets in about 100,000 workspaces may be compromised. October 23 updates detail recovery and urge customers to revoke and replace credentials.
Read more →
Node2Shell research demonstrates code execution through the Node.js Brotli decoder. Fixed releases are available, and firewall protection has important limits.
Read more →
PHP emergency updates fix two critical remote code execution flaws, prompting urgent patching and precautionary website shutdowns across affected organisations.
Read more →
Eight compromised eMudhra certificate keys raise HTTPS impersonation risks, prompting browser and software updates and urgent reviews of trusted certificates.
Read more →
New NGINX alerts urge administrators to patch three flaws that can crash servers or potentially execute code, with exposure dependent on server configuration.
Read more →
Attackers exploit SiteGuard Server Edition to run commands as root or SYSTEM. Fixes are available, but administrators must preserve evidence before patching.
Read more →
ConnectWise expands service suspensions after Microsoft 365 intrusions, as partners review application credentials and await backup and restoration access.
Read more →
Attackers exploited a BeyondTrust Remote Support zero-day to gain access and deploy backdoors. Self-hosted customers must verify patches and review exposure.
Read more →