We pointed our autonomous offensive agent, AgentO, at Datadog's new Pathfinding Labs and it solved all 122 live AWS privilege-escalation labs — no human in the loop, no access to the answers, average 9 turns. Here's what made it work.
scopeshift demonstrates that Claude Code and other AI coding agents can be tricked into running unauthorized pentests through network-layer deception alone — no adversarial prompts required. The agent thinks it's probing localhost, but the traffic lands on a real third-party target.
AWS IAM eventual consistency gives attackers a 4-second window after credential revocation. notyet is an open-source tool to test IR containment playbooks.
Standard AWS IR containment fails against attackers exploiting IAM eventual consistency. This article presents an SCP-enforced technique that makes identity-level containment attacker-resistant.
AWS IAM eventual consistency creates a 4-second window where deleted AWS access keys can still work. Learn how attackers exploit this and how to mitigate it.
An initial security overview of AWS Lambda Managed Instances, exploring the Bottlerocket-based architecture, the 'Elevator' components, and security insights for this new compute model.
Exploring the breakthrough potential and emerging risks of AI agents that can autonomously discover and exploit complex AWS attack chains.
Did you know a misconfigured private API Gateway can be invoked from any AWS account?
The article highlights the dangers of using the AWS managed policy PowerUserAccess in complex environments by sharing the attacker's perspective on the matter.
This research was conducted by OFFENSAI to show what an attacker can achieve after creating a malicious OIDC identity provider in AWS and how they can do it. The article presents novel techniques and tools for persistence and evasion.
Move from cloud exposure detection to controlled validation, technical evidence, and risk-based prioritization, powered by AI.
