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
METR · March 26, 2026

Agent monitoring: test coverage gaps separately from the monitor’s judgment

analysis Agent SecurityAI Red TeamingModel Evaluation
Why it matters

METR’s embedded red-team exercise found routes around Anthropic’s internal monitoring, including traffic-identification configuration and an unmonitored subagent path that was fixed promptly. The article’s May 2026 update also describe monitor jailbreaks and possible unmonitored compute, with important limits: testers had infrastructure knowledge agents generally lacked, and contemporaneous models struggled to devise reliable covert attacks. Retrospective detection over days or weeks does not establish prevention of each action. These observations concern the assessed early-2026 systems, not their current security posture.

My takeaway: Trace monitoring coverage through parent agents, subagents and long-lived jobs. Test with informed adversaries, enforce monitoring outside the agent’s writable configuration, and measure per-action interception separately from eventual incident discovery.
Keep exploring

More curated notes connected through Agent Security 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.