AWS Attack & Defend Security for Startups
The AWS mistakes small teams actually ship: over-permissive IAM, public S3, exposed metadata and long-lived keys. Attack each one, fix it without a platform team, then close with the AWS Security Agent and Bedrock AI security.
- Format
- Online, live
- Labs
- Your own AWS account
- Next batch
- Not scheduled
We will write once, when a date is set.
About this training
AWS runs a large share of production web application infrastructure today, which means the attack surface is broad and the risk is real. This workshop is a structured entry point for engineers starting AWS security, and a practical attacker-to-defender path for teams that already run production workloads, closing with the AWS Security Agent and AI security on Amazon Bedrock.
Across two days you make a complete pass through AWS security with hands-on labs that mirror production. The emphasis is on assessing managed AWS services, real adversary tradecraft, and defenses that hold up afterwards. You start with IAM fundamentals and privilege escalation patterns, then move through serverless and data, network, and container infrastructure on ECS and ECR. You chain attacks end to end: overly permissive roles, exposed APIs, misconfigured security groups, stale DNS, and risky storage policies. From there you apply exploitation techniques including Lambda RCE paths, ECR image backdooring, and LFI against EC2 behind WAF. Defense is not an afterthought: you harden services, generate and consume telemetry, and operationalise response with GuardDuty and Security Hub, including SBOM generation and image scanning. Day two ends on AI: securing Amazon Bedrock with guardrails, building a defended GenAI application, and using AI-assisted tooling for penetration testing and code review.
Labs are step by step. The trainer runs each one first, then participants repeat it. Every block ends with a clean teardown to control cost and drift.
Key aspects of the workshop include:
- Map the AWS kill chain across identity, network, data and containers
- Escalate with wildcard PassRole and AssumeRole, using Pacu
- Exploit Lambda RCE from API Gateway and pivot with SSM
- Expose and validate RDS through a misconfigured security group
- Take over subdomains via stale Route53 entries pointing at S3
- Probe EC2 behind AWS WAF and exploit LFI
- Deploy and attack a vulnerable ECS service, then backdoor an ECR image
- Detect with signal, not noise: GuardDuty findings and suppression rules
- Feed findings into Security Hub, run automation rules, and analyse WAF logs in Athena
- Draft practical SCPs and permission boundaries that teams can actually live with
- Secure Amazon Bedrock with guardrails, then build and defend a GenAI application
Who this is for
Does one of these read like your own week?
- You run production workloads on AWS and have never attacked your own account.
- You are a founder or lead engineer at a small team with no dedicated platform or security hire yet.
- You own IAM, and inherited roles and policies from someone who has since left.
- You want the AI security layer too: Bedrock guardrails, a defended GenAI app, and AI-assisted pentesting and code review.
- You can read a CloudFormation or Terraform file and want the attacker's view of what it actually grants.
If so, this training is for you.
What you’ll do
56 labs and 2 theory sessions across 11 sections. Theory appears where it is needed for the next lab, not as a lecture block.
01Setup and Fundamentals
- theoryWelcome, agenda and prerequisites
- labLab environment setup: accounts and tooling (Pacu, AWS CLI)
02AWS IAM Fundamentals and Attacks
- labIAM users and groups
- labIAM SCPs and RCPs
- labS3 resource policy for fine-grained access
- labIAM least privilege
- labPrivilege escalation via wildcard IAM PassRole
- labPrivilege escalation via wildcard IAM AssumeRole, using Pacu
- labCleanup
03Serverless and Compute Attacks
- labDeploy Lambda with API Gateway
- labExploiting Lambda RCE via API Gateway
- labLambda privilege escalation via SSM
- labExploiting exposed RDS via a security group with password auth
- labExploiting the OWASP Serverless Top 10
04Messaging, DNS and Storage Attacks
- labDeploy SQS and SNS
- labPublic SQS queue and SNS topic with a wildcard principal
- labUnencrypted SNS
- labS3 subdomain takeover via a stale Route53 entry
- labExploiting the subdomain takeover
- labPublicly exposed S3 bucket via a misconfigured ACL
05Network and WAF Attacks
- labDeploying EC2 and WAF
- labReviewing security group and VPC misconfiguration
- theoryAWS WAF rule sets
- labExploiting a vulnerable EC2 instance via LFI
- labCleanup
06Container Attacks
- labDeploy a vulnerable ECS service
- labTask role and execution role privilege escalation
- labMisconfigured, public ECR repository
- labBackdooring Docker image layers with Dive and Crane
- labCleanup
07Defense and Monitoring, Part 1
- labAWS S3 security
- labS3 signed URLs
- labS3 origin access identity
- labAWS GuardDuty for threat detection
- labGuardDuty suppression rules
- labGenerating an SBOM with Amazon Inspector and Syft
- labBasic VPC and NACL security
- labNetwork flow logs using VPC traffic mirroring
08Defense and Monitoring, Part 2, and AWS WAF
- labAWS Security Hub threat detection
- labAutomation rules for Security Hub findings
- labSuppressing findings in Security Hub
- labAutomated weekly vulnerability summary email
- labWeb application security with AWS WAF
- labRate limiting bots with a custom response
- labAnalysing AWS WAF logs with Athena
- labCleanup of Security Hub and AWS WAF
09Amazon Bedrock Security and Guardrails
- labCreate guardrails using the AWS console
- labCreate guardrails using the AWS SDK and CLI
- labApply guardrails while invoking models
- labContent filtering and PII redaction with guardrails
- labDefend against prompt injection attacks
10Building Secure GenAI Applications
- labArchitecture overview of a secure GenAI application
- labSet up the frontend and backend
- labIntegrate Bedrock guardrails into the application
11AI-Powered AppSec with AWS Security Agent
- labConfigure a penetration testing environment
- labAI-assisted penetration testing
- labSecurity design review using AI
- labAI-assisted security code review
Read the prerequisites before you book.
Some trainings require you to bring your own cloud account (AWS, GCP, or both) with billing enabled and admin privileges. Labs will not run without it, so confirm you meet every prerequisite below before the session.
Hard requirement for corporate laptops: admin permission to install tools, endpoint security and antivirus disabled, and the required lab domains not blocked on your network. Confirm this with your IT team before you register - troubleshooting a locked-down laptop mid-session is not possible.
What you need for this training
- Your own AWS account, free tier is enough, with a valid card on file and MFA enabled.
- AWS CLI installed and configured. Administrative privileges on your laptop are recommended.
- A laptop that can reach the internet without a restrictive corporate proxy, and can install a lightweight VPN client.
- Comfort with the command line. No prior AWS security experience required.
- Bedrock model access requested and approved in your own account before day 2, using instructions sent after registration.
Why your own account
- A mistake in one account cannot touch anyone else's resources.
- Billing stays traceable per person.
- The experience matches a real account, not a shared sandbox.
- A trainer can see directly which account still needs cleanup.
- Bedrock and the AI Security Agent labs need account-level quotas a shared account cannot give everyone.
Estimated AWS cost
- $35-50 per student for the two days.
- Mostly Bedrock, GuardDuty, Security Hub and compute time.
- Set a billing alert at $30.
- Set a hard budget alert at $50 before day 1.
- Run the cleanup script at the end of each day.
How our trainings run
- Live and instructor-led, with the trainers in the room rather than a recording.
- Labs run on your own account and your own laptop, with cleanup steps provided at the end of each block.
- Attack first, then fix. Every exploited flaw is followed by the control that stops it.
- You keep the material, the recording, and a certificate of completion.
Who runs it
Anjali Shukla
Senior Security Engineer · Black Hat and Nullcon trainer
Founder, Container Security Village. OWASP EKS Goat lead. AWS Community Builder. Women Influencer in Cloud Security, CSA Bangalore 2023. Leads the W3-CS Bengaluru chapter.
Want a seat when it runs?
This batch is not scheduled yet. Join the waitlist and we will write the moment a date is set.
Support Our Work
Subscribe, like, and share our videos.
No Spam. Only updates.