Inside COASP:Offensive Security for Modern AI Systems

Agenda for the Masterclass

Why AI systems need offensive security testing
Overview of COASP certification and its relevance
Setting expectations for demos and learning outcomes
AI Attack Surface & Threat Landscape
Differences between traditional pentesting and AI pentesting
Key risks: prompt injection, model poisoning, data leakage
Mapping threats to OWASP Top 10 for LLMs
Hands‑On Demo: Offensive AI Exploits
Prompt Injection → bypassing guardrails
Sensitive Information Disclosure → unintended data leakage
Model Manipulation → adversarial inputs and poisoning
Walkthrough of exploitation methodology
Pentesting Methodology for AI Systems
Reconnaissance: OSINT on AI integrations
Exploitation: adversarial prompts, embedding manipulation
Post‑exploitation: persistence in agent workflows
Safe lab setup for AI pentesting
Defensive Countermeasures & Blue Team Perspective
Guardrails and input validation
Monitoring and anomaly detection
Case study: spotting prompt injection attempts
Human‑in‑the‑loop strategies

Why Attend This Masterclass

Get CPE Certificate

FREE Career Guidance & Mentorship

Learn from Industry Experts

Ashish Dhyani

10+ Years of Experience

Upcoming Events for You

FREE Training

Think Like a Security Architect: Building Architecture Patterns in Practice

28 July (Tue) , 08:00 – 09:00 PM (IST)

309People have registered

Paid Masterclass

Practical AI Audit Bootcamp

05 – 06 Aug 2026 (Wed – Thu) , 07:00 – 10:00 PM (IST)

FREE Training

Crack AAISM: Practice Questions & Exam Strategy

07 Aug (Fri) , 08:00 – 09:00 PM (IST)

217People have registered

Paid Masterclass

Practical Cloud Audit Bootcamp

11 – 13 Aug 2026 (Tue – Thu) , 07:00 – 10:00 PM (IST)

Save a seat before the spots fill out!

<ul>
<li>Why AI systems need offensive security testing</li>
<li>Overview of COASP certification and its relevance</li>
<li>Setting expectations for demos and learning outcomes</li>
<li>AI Attack Surface & Threat Landscape</li>
<li>Differences between traditional pentesting and AI pentesting</li>
<li>Key risks: prompt injection, model poisoning, data leakage</li>
<li>Mapping threats to OWASP Top 10 for LLMs</li>
<li>Hands‑On Demo: Offensive AI Exploits</li>
<li>Prompt Injection → bypassing guardrails</li>
<li>Sensitive Information Disclosure → unintended data leakage</li>
<li>Model Manipulation → adversarial inputs and poisoning</li>
<li>Walkthrough of exploitation methodology</li>
<li>Pentesting Methodology for AI Systems</li>
<li>Reconnaissance: OSINT on AI integrations</li>
<li>Exploitation: adversarial prompts, embedding manipulation</li>
<li>Post‑exploitation: persistence in agent workflows</li>
<li>Safe lab setup for AI pentesting</li>
<li>Defensive Countermeasures & Blue Team Perspective</li>
<li>Guardrails and input validation</li>
<li>Monitoring and anomaly detection</li>
<li>Case study: spotting prompt injection attempts</li>
<li>Human‑in‑the‑loop strategies</li>
</ul>

By submitting your contact details, you acknowledge and agree to our Terms of Use and Privacy Policy.

I agree to receive your marketing and promotional emails

201 People have registered