Sort the Security Risks
Classify common open-model application risks using OWASP LLM risk categories.
Sort each scenario into the best OWASP LLM risk bucket. Prompt injection Supply chain vulnerability Excessive agency Sensitive information disclosure Overreliance A retrieved web page contains hidden instructions telling the model to ignore the user's task. The team downloads weights from an unofficial mirror with no checksum verification. The model can directly update customer account status through a connected tool. Full prompts containing customer identifiers are stored in broad-access debug logs. Support agents accept generated summaries without checking source snippets during peak queues. A fine-tuning dataset includes poisoned examples from a public issue tracker.
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in