AI security
LLM security starts with architecture
A technical explanation of token streams, prompt boundaries and the control layers around LLM applications.
Read articleResearch
Notes from the work.
Technical walkthroughs and lab research on application security, cloud environments, AI systems and defensive controls.
AI security
A multi-turn attack through the LLM tool layer
A controlled-lab attack chain showing how per-message filters, tool parameters and unsafe template context combine into cross-scope data exposure.
Read articleEndpoint security
Testing EDR controls with custom shellcode runners
A lab study of DLL unhooking, AMSI bypass and payload transformation against an enterprise EDR trial.
Read articleCloud security
AWS privilege escalation through SSRF and IAM
A lab walkthrough tracing SSRF, exposed credentials, S3 access and IAM privilege escalation in AWSGoat.
Read articleCloud security
Azure privilege escalation through automation accounts
An AzureGoat walkthrough covering SSRF, exposed application settings, RBAC discovery and runbook abuse.
Read articleCloud architecture
Cross-account AWS database access
A practical guide to VPC peering, IAM role assumption and Secrets Manager across AWS accounts.
Read articleAdversary simulation
Building a reproducible red team lab
A Windows and Active Directory lab for testing multi-step attack paths without relying on destructive techniques.
Read article