Programme objective: In an isolated authorised environment, you identify a finding, validate it, prioritise it, remediate it, rerun the check, and report to the owner. The work is a written assessment of a sample retrieval-augmented generation application: a risk rating, a remediation plan, and retest evidence.
Learner path: The current OWASP Top 10 for web applications, the OWASP API Security Top 10, static and dynamic testing, composition analysis, containers, infrastructure-as-code, secrets, a software bill of materials, licences, pipeline gates, threat modelling, then large language model application risks. The surrounding gates are on Complete AI-Native SDLC Engineering.
Suggested time inside the pathway is 30 hours, within a 16-week professional programme of about 160 to 200 hours. That shape is subject to batch and faculty planning. The cohort and any certificate are confirmed when you enquire. This page does not state fees, batch dates, faculty names, or placement rates.
Module 1 — OWASP Top 10 web application security
You use the current official edition. The project page names the 2025 release as the current Top 10. You study these category names, map each one to a control and a retest, and report the result to the owner:
- Broken Access Control
- Security Misconfiguration
- Software Supply Chain Failures
- Cryptographic Failures
- Injection
- Insecure Design
- Authentication Failures
- Software or Data Integrity Failures
- Security Logging and Alerting Failures
- Mishandling of Exceptional Conditions
Module 2 — OWASP API Security Top 10
This is its own module. The class uses the current official edition. The edition published on the OWASP API Security project page, and verified for this outline, is the 2023 list. You map each category name to a control, a severity, and a retest:
- Broken Object Level Authorization
- Broken Authentication
- Broken Object Property Level Authorization
- Unrestricted Resource Consumption
- Broken Function Level Authorization
- Unrestricted Access to Sensitive Business Flows
- Server Side Request Forgery
- Security Misconfiguration
- Improper Inventory Management
- Unsafe Consumption of APIs
Module 3 — Static application security testing
You review source with SonarQube, Semgrep, GitHub CodeQL, Bandit, ESLint security plugins, and Gitleaks, then you confirm a finding before you call it a defect.
Module 4 — Dynamic application security testing
You run an authorised scan with OWASP ZAP or Burp Suite in the isolated environment, covering authenticated scanning, API scanning, session checks, and a later security regression. You record what the tool reported and what a person confirmed.
Module 5 — Software composition analysis
You review dependencies with OWASP Dependency-Check and Snyk, and you triage a critical finding instead of copying the tool’s label into the report unchanged.
Module 6 — Container review
You review the container image with Trivy and you report what you validated to the owner.
Module 7 — Infrastructure-as-code review
You review the infrastructure-as-code with Checkov and KICS, and you separate a confirmed issue from a false positive.
Module 8 — Secrets review
You look for hardcoded secrets with Gitleaks. A secret in the tree is removed from the work you will keep, and you say so in the report.
Module 9 — Software bill of materials
You produce a software bill of materials with CycloneDX and SPDX so the owner can see which components the sample application includes.
Module 10 — License scanning
You record the licences the scan surfaced and you flag a licence the owner needs to read before the component is kept.
Module 11 — CI/CD security gates and signed artifacts
You place security checks on the pipeline and you require a signed artifact. A tool’s recommendation still waits for a person before a release.
Module 12 — Threat modelling with STRIDE
You walk the sample with STRIDE — spoofing, tampering, repudiation, information disclosure, denial of service, and elevation of privilege — and you write the control and the owner for each item you keep.
Module 13 — Large language model application security
You use the OWASP Top 10 for Large Language Model Applications, 2025 edition, which is the list published on the GenAI project page for this outline. Hidden context exposure and untrusted tool outputs are not separate headings in that list. You still practise tool-calling permission boundaries in the laboratory themes on this page. The category names you study are:
- Prompt injection
- Sensitive information disclosure
- Supply chain
- Data and model poisoning
- Improper output handling
- Excessive agency
- System prompt leakage
- Vector and embedding weaknesses
- Misinformation
- Unbounded consumption
For each name, you record the risk on the sample retrieval-augmented generation application, the control, and the retest. The page does not include a sample attack.
Security operations and alert review sit on AI in Cybersecurity. Ordinary expected results sit on AI-Powered Test Automation Engineering. The code you build sits on AI-Native Software Engineering and Development.
