Inside Out - The Cloud has Never been so Close
summary
Claims that mapping real AWS IAM, Lambda, EC2 and S3 metadata into a graph uncovers concrete privilege escalation chains, for example stolen long‑term keys on a workstation -> lambda modify code -> fetch instance metadata -> assume role -> read S3. Implementation uses AWS APIs and IAM simulation, Python to export CSV then write into Neo4j, and aggregates low‑level relations into attack connections with numeric weights to rank easy, medium, hard paths. Tool is predictive, not preventive, and tested on decent sized environments.
AI-written summary · about
chapters (8)
- 0:00Introduction to cloud security
- 2:29API pitfalls and AWS overview
- 5:25Attack vectors and simulation
- 8:07Graph-based security solution
- 10:30Graph data modeling
- 14:19Risk analysis and visualization
- 17:26Future improvements
- 19:19Q&A session
related talks
26mAWS (mis)configuration from an attacker’s-eye view
56mPivoting in Amazon Clouds
40mBringing a Machete to the Amazon
43mThink like a HACKER
46mAttacking AWS: The Full Cyber Kill Chain
47mBringing a Machete to the Amazon
47mBreaking the Cloud to Rebuild It: A Tale of 3 ☁️ Breaches!
40mWeak Links in Cloud IAM - Never Trust. Always Verify!