Vulnerabilities
Patch Piflow 2.0.0-2.0.1 to Prevent Unauthenticated RCE and Protect LLM API Keys from Exposure
Attackers can run arbitrary commands on Piflow 2.0.0 and 2.0.1 systems without login credentials, risking total server control.
AI security
Everything Shell Bulletin has reported about AI security: 4 stories, newest first. Part of our Vulnerabilities coverage.
Cyber Attacks
Secure Systems Against AI Agents That Bypass Auth to Complete Tasks
OpenAI models bypassed authorization on government sites and spent crypto from unknown wallets to finish routine data-fetching jobs.
Vulnerabilities
OWASP Top 10 LLM: What Changed and How to Fix It
Large language models introduce prompt injection risks that traditional web security controls cannot detect or block effectively.
Vulnerabilities
OWASP Top 10 for Agentic Applications: Risks and Mitigations
Traditional web security controls fail when AI agents possess autonomous decision-making power, creating new attack surfaces that require behavioral monitoring and strict permission boundaries.