Back
advanced
Security, Safety & Risk

OWASP LLM Top 10 for AI Applications

Use the OWASP GenAI security risks as a production review checklist

26 min read· OWASP· security· LLM Top 10· risk

OWASP LLM Top 10 for AI Applications

The OWASP GenAI security project gives teams a shared vocabulary for reviewing LLM application risk.

Use this lesson as a checklist before shipping any app that sends model output to users, tools, databases, or workflows.

The risk areas

Risk areaWhat to look for
Prompt injectionuser or document instructions override app rules
Insecure output handlingmodel output is executed or rendered unsafely
Training data poisoningbad data changes model behavior
Model denial of servicehuge prompts or loops create excessive cost
Supply chain vulnerabilitiesunsafe models, datasets, packages, or plugins
Sensitive information disclosuresecrets or private data appear in prompts/logs/output
Insecure plugin/tool designtools expose too much power
Excessive agencymodel can take high-impact actions unsafely
Overreliancehumans trust wrong answers without verification
Model theftmodel, prompts, or proprietary behavior are extracted

Review questions

Ask these before launch:

  • What data can the model see?
  • What tools can it call?
  • Which actions require human approval?
  • What happens if the model is wrong?
  • What logs contain user data or secrets?
  • How do we detect prompt injection?
  • Can output become HTML, SQL, shell, or code?
  • Can a malicious document influence actions?
  • What is the rollback plan?

Mitigation patterns

RiskMitigation
Prompt injectioncontext separation, detection, least-privilege tools
Insecure outputencode output, validate schemas, avoid direct execution
Model DoStoken limits, rate limits, budgets, timeouts
Sensitive dataredaction, access control, retention limits
Excessive agencyscoped tools, confirmations, audit logs
Overreliancecitations, confidence messaging, human review

Security is a lifecycle

LLM security is not one review. It changes when you:

  • switch models
  • change prompts
  • add tools
  • add retrieval sources
  • expand user permissions
  • expose outputs in new channels

Knowledge check

Q1: Why is "excessive agency" dangerous?
Because the model may be able to take real actions beyond what is safe or intended.

Q2: What should happen before a high-impact action?
The app should require explicit permission, validation, and audit logging.