EC-COUNCIL

EC-Council Certified Offensive AI Security Professional

COASP

The EC-Council Certified Offensive AI Security Professional (COASP) certification validates your ability to red-team AI systems, exploit vulnerabilities in LLMs and agents, and build defenses that survive real-world attacks. It is designed for offensive security professionals who need to master prompt injection, data poisoning, and model manipulation. Earning COASP proves you can simulate attacks, find vulnerabilities, and harden AI systems before attackers exploit them.

Practice Free Questions

7Domains

28Objectives

252Practice pages

1198Free questions

COASP free practice questions

Choose any objective and open any question page — no sign-in required.

AI and machine learning fundamentals from an offensive security perspective

41 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–41

AI attack surfaces, threat landscapes, and adversary techniques (MITRE ATLAS-aligned)

47 questions · 10 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–47

AI system hacking methodologies, frameworks, and risk implications

42 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–42

AI attack taxonomies and classification models

39 questions · 8 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–39

OWASP LLM and ML Top 10 (2025) mapping to AI threat and governance

48 questions · 10 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–48

AI Reconnaissance and Attack Surface Mapping

45 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45

API Reconnaissance and model fingerprinting

38 questions · 8 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–38

AI Vulnerability Scanning and Fuzzing

54 questions · 11 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–50 51–54

Prompt injection attacks

37 questions · 8 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–37

LLM jailbreaking

45 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45

RAG poisoning attacks

44 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–44

Cross-LLM attacks

37 questions · 8 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–37

Adversarial machine learning techniques

38 questions · 8 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–38

FGSM and PGD attacks on image classifiers

35 questions · 7 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35

Model extraction and theft

41 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–41

Model privacy attacks

42 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–42

Data and training pipeline attacks

48 questions · 10 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–48

Training data poisoning

39 questions · 8 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–39

Agentic AI and model-to-model attacks

41 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–41

Agent hijacking

56 questions · 12 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–50 51–55 56–56

Authentication weaknesses in AI model APIs

48 questions · 10 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–48

Parameter manipulation

41 questions · 9 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–41

SSRF through AI tool calls

28 questions · 6 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–28

Misconfigured cloud AI services (AWS Bedrock, Azure OpenAI, Google Vertex AI)

34 questions · 7 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–34

AI Security Testing, Evaluation, and Hardening

48 questions · 10 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–48

AI Incident Response and Forensics

56 questions · 12 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–50 51–55 56–56

Scoping an AI red team exercise and threat modeling

37 questions · 8 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–37

Professional AI red team reporting

49 questions · 10 pages

Questions 1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–49

Want a smarter study session?Try adaptive questions aimed at your weak spots, the illustrated book, and readiness you can trust.

Continue Free Practice

Percentages reflect share of the current practice bank, not official exam weightings. Every page above is a live, crawlable practice page (28 objectives · 252 pages).

EC-Council Certified Offensive AI Security Professional

COASP free practice questions

Domain 1 · Offensive AI Foundations and Hacking Methodology18% of practice bank · 5 objectives · 217 questions18%

Domain 2 · AI Reconnaissance and Vulnerability Discovery11% of practice bank · 3 objectives · 137 questions11%

Domain 3 · Prompt Injection and LLM Application Attacks14% of practice bank · 4 objectives · 163 questions14%

Domain 4 · Adversarial Machine Learning and Model Privacy Attacks13% of practice bank · 4 objectives · 156 questions13%

Domain 5 · Data Pipeline and Agentic AI Attacks15% of practice bank · 4 objectives · 184 questions15%

Domain 6 · AI Infrastructure and Supply Chain Attacks13% of practice bank · 4 objectives · 151 questions13%

Domain 7 · Security Testing, Hardening, and Incident Response16% of practice bank · 4 objectives · 190 questions16%