The AI Agent Risk Checklist

67 checks to catch hallucinations, prompt injection, data leaks and unsafe tool use before they reach your customers.

Open the checklist

Free · printable · no signup

Checks
67
Risk categories
10
Critical-severity
20

What's inside

10 categories, aligned with the OWASP Top 10 for LLM Applications and for Agentic AI. Every item has a severity.

  • Hallucination Detection

    7 checks

    Catch when AI makes up facts, cites non-existent sources, or confidently delivers wrong answers

  • Prompt Injection Prevention

    9 checks

    Test if users can manipulate AI to bypass rules, leak data, or behave maliciously

  • Security & Privacy Checks

    8 checks

    Critical security tests for data leakage, PII handling, and API vulnerabilities

  • Jailbreak Resistance

    6 checks

    Ensure AI can’t be tricked into generating harmful, offensive, or brand-damaging content

  • Output Validation

    5 checks

    Catch formatting errors, broken logic, missing citations, and inconsistent responses

  • Bias & Fairness Audits

    6 checks

    Test for demographic bias, stereotype reinforcement, and unfair treatment patterns

  • Content Moderation

    4 checks

    Safeguards against illegal content, brand violations, and regulated advice

  • Production Monitoring

    5 checks

    Ongoing checks to catch failures in real-time before they go viral

  • Agentic & Tool-Use Safety

    10 checks

    Critical checks for AI agents that call tools, use MCP, or take autonomous actions (OWASP Agentic AI Top 10)

  • Dark Pattern Detection

    7 checks

    Scan AI-generated UIs, copy, and e-commerce flows for deceptive design patterns that manipulate users

The failures these checks are built to catch

Four of the 70 documented incidents in our database. Each one maps to a category above.

How to use it

  1. 1

    Before deployment

    Run all 67 checks before launch. Treat any failed critical item as a release blocker.

  2. 2

    After every change

    Re-run the checklist when you change the model, the prompt or the data your agent retrieves. Small edits can undo earlier fixes.

  3. 3

    On a schedule

    Spot-check production regularly. Models drift and attack techniques change.

  4. 4

    Fit it to your agent

    Not every item applies to every agent. A support chatbot and an agent that calls tools need different emphasis; start with the critical items that apply.

Who it's for

Founders and CTOs
You are shipping an AI feature and need to know the failure modes are covered.
AI engineers
You build the agent. Use it to test edge cases and adversarial inputs before production.
QA and testing teams
Use it as an AI-specific test plan for what manual testing misses.
Security and compliance
Audit existing deployments and flag gaps for remediation.

Frequently asked questions

Is it really free?

Yes. No credit card, no email gate. Open it, print it, share it.

Do I need technical expertise?

No. Each item has a plain-English explanation and an example test. Some items, such as tool-use checks, are easier with an engineer.

How long does it take?

It depends on your agent. A simple FAQ chatbot is much quicker to check than an agent that calls tools or takes actions. Start with the critical items.

Can I share it with my team?

Yes, please do.

What format is it?

A clean, printable web page you can save as PDF from your browser.

Will it prevent every AI failure?

No checklist can. It covers the failure modes behind well-documented incidents, so you can find and fix the common ones before your customers do.

Don't ship blind

Open the checklist, print it, and work through the critical items before your next release.

Open the checklist

Not sure where to start? Take the 3-step risk quiz