AI Red Team Engineering notes, research references, and field notes on AI systems, responsible AI, governance, and compliance.
Global search

Search AI Red Team

Search across the complete knowledge base.

Enter at least two characters to begin.
↑↓ select Enter open Browse the full research archive
The Hacker News AI Security · August 18, 2026

Attackers Exploit MLflow SSRF Flaw to Steal Cloud Credentials and Secrets

news AI EngineeringAI Red Teaming
Why it matters

CVE-2026-64849 is an unauthenticated SSRF flaw in MLflow model-registry webhooks that lets anyone reaching a vulnerable Tracking Server proxy requests to internal services and cloud metadata endpoints. The redirect handling bypasses earlier fixes, and honeypot telemetry showed indiscriminate scanning within hours of disclosure aimed at stealing cloud credentials and secrets. MLflow fixed the issue in 3.15.0.

My takeaway: Upgrade MLflow to 3.15.0 or later and remove Tracking Servers from public reach. Review webhook and network logs for metadata-service access, rotate every credential reachable from the affected workload, restrict instance metadata and egress, and require authentication and network policy around experiment infrastructure even when the upstream project assumes a trusted deployment boundary.
Keep exploring

More curated notes connected through AI Engineering and AI Red Teaming.

OpenAI News · framework

Pacing model development in an era of cyber-critical capabilities

OpenAI says preliminary evidence that Astra may meet its Critical cybersecurity threshold led it to pause frontier reinforcement-learning work for two weeks and keep its largest planned run on hold. New safeguards include stronger workload and network isolation, continuous boundary testing, token-level monitoring that escalates suspicious tool activity, and broader alignment checks for deception, reward hacking, and unauthorized access.

OWASP GenAI Security Project · guide

OWASP Top 10 for Agentic Applications for 2026

OWASP's community guide organizes agentic-system risk into ten categories, including goal hijacking, tool misuse, identity and privilege abuse, memory poisoning, insecure inter-agent communication, cascading failures, and rogue-agent behavior. It provides a shared taxonomy and mitigation starting point rather than a certification checklist or evidence that a deployed system is secure.

OpenAI News · framework

A blueprint for democratic governance of frontier AI

OpenAI proposes a three-part U.S. frontier-AI governance model: harmonize emerging state safety laws into a federal baseline, strengthen CAISI as an evaluation and standards institution, and coordinate a broader resilience program. Proposed controls include severe-risk evaluations, transparency reports, independent audits, safety-incident reporting, model-weight security, whistleblower protection, and periodic technical assessments.