Do you guarantee that I will pass?
No. Certification is earned based on individual performance. We do not guarantee exam results or successful outcomes.
About the Certification
The kAIPTA certification validates junior-level, hands-on capability to security test AI-powered applications. The exam is practical and outcome-based, requiring candidates to identify weaknesses, demonstrate real impact, and submit proof of success through captured flags.
What is this exam?
kAIPTA is a practical AI penetration testing exam that measures applied skills rather than introductory knowledge. Candidates are expected to analyze how AI-enabled applications behave under adversarial interaction, uncover security weaknesses in realistic deployments, and demonstrate exploitation.
Who should take this exam?
This exam is for people with some cybersecurity or application testing experience who want to move into AI security. It’s ideal for junior pentesters, junior AppSec engineers, early-career security analysts, and developers responsible for testing or securing AI features. It also suits experienced penetration testers expanding into AI application testing. It is not intended for complete beginners with no security background.
Exam format
The exam follows a practical flag-submission style. Candidates will be given four separate AI applications to test, and each application contains intentional, realistic weaknesses. All lab applications are powered by real LLM/AI models to reflect real-world behavior and risk. The goal is to find and exploit issues to obtain flags and submit them as evidence of successful compromise and impact. The exam duration is 4 hours, with an additional 30 minutes provided for kAIPTA lab environment provisioning and access setup, giving a total available time of 4.5 hours. Scoring is based on the correctness of submitted flags.
Experience needed
Candidates should be comfortable with junior-level penetration testing workflows, including testing web or API-based applications, crafting inputs to probe system behavior, interpreting responses, and reasoning about risk and impact. Familiarity with industry baselines such as the OWASP guidance for LLM application risks is beneficial, but deep machine learning knowledge or model training experience is not required. Candidates should be ready to approach AI features with the same disciplined, methodical mindset used in modern application security testing.
Pass criteria
- Candidates must achieve a minimum overall score of 75% to pass the exam and receive their certification credentials.
- Candidates achieving an overall score of 85% or above will be awarded a Merit distinction.
Policies
All exams are conducted under strict integrity standards. Candidates must complete the exam independently—receiving or providing help, using unauthorized resources, or sharing questions or answers in any form (during or after the exam) is strictly prohibited and may result in disqualification and revocation of certification.
Retake policy
This exam includes 1 free retake as part of the voucher policy. After all included retakes are used, any additional attempt will require the purchase of a new voucher.
Certificate validity
This certification includes lifetime online verification and does not expire. Each certificate clearly indicates the exam version and the exam passing date to provide transparent context on when the assessment was completed.
As industry practices and tools evolve, we strongly recommend taking the latest exam version periodically to demonstrate that your knowledge and skills remain current.
Exam Syllabus
- Direct prompt injection
- Instruction conflict and priority manipulation
- Policy and constraint bypass attempts
- Secret exfiltration from system prompts
- Indirect prompt injection via reframing and rewording
- Context manipulation and ambiguity exploitation
- Single-turn prompt construction under stateless constraints
- Incomplete or improper filtering of sensitive data in responses
- Unintended disclosure caused by misinterpretation or processing errors
- Memorization/overfitting risk awareness and exposure conditions
- Confidential data handling failures
- Insecure Output Handling
- Unsafe output formatting/encoding handling
- Downstream trust of generated output
- Output-driven security impact pathways
- Overreliance
- Unverified acceptance of generated outputs
- Missing validation steps for high-impact outcomes
- Failure modes from assuming correctness
- Insecure Plugin/Tool Design
- Authorization and access control weaknesses
- Input/parameter handling weaknesses
- Trust boundary mistakes between the model and external capabilities
- Excessive agency, functionality, permissions & autonomy
FAQ
Are exam fees refundable?
All exam purchases are final. Once purchased, no refunds will be issued under any circumstances.
If I fail, when can I retake the exam?
You can retake the exam immediately after a failed attempt or at any time before your exam voucher expires. There is no mandatory waiting period between attempts, as long as your voucher remains valid.
Can I share my exam content or answers?
No. Sharing exam questions, answers, or any part of the assessment is strictly prohibited and may result in disqualification or revocation of certification.