OWASP LLM Top 10 + Promptfoo Red-Team Harness

Live AI Agent & System-Prompt Vulnerability Grader

Test your AI agent system prompt and tool permissions against 6 empirical exploit vectors powered by promptfoo/promptfoo (26,080 ), confident-ai/deepeval (18,921 ), and NVIDIA garak. Get an instant 0-100 Resilience Score, auto-patch exposed vectors, and export a ready-to-run promptfooconfig.yaml.

25,999promptfoo/promptfoo

18,867confident-ai/deepeval

9,575NVIDIA/garak

3,217protectai/llm-guard

OWASP LLM Red-Team Grader & 9-File Stack Architecture Visual

1. Select Agent Architecture Preset (or Paste Custom Prompt) Agent Identifier Enabled Agent Tools (Attack Surface) System Prompt / Agent Directives

Auto-Hardened System Prompt Patch (1-Click Fix)

Running initial red-team probe...

Empirical Red-Team Resilience

Evaluating...

--/100

OWASP LLM & Promptfoo Exploit Vector Breakdown

Generated promptfooconfig.yaml Harness

Free $0 Functional Transport — Email & Unlock Copy / Download

Enter your work email to copy or download the free promptfooconfig.yaml ($0), or grab the $4.99 Config Pack below.

Share Score on X

Enterprise Security Procurement Anchor

Skip the $8,500 Manual Pentest Delay • Certify Your Repo Today

Save 98% vs Pentest Firms

Enterprise teams block unverified AI tools at security review. Start with the $0 Free CLAUDE.md, unlock the $4.99 Config Pack or $9.99 9-File Hardening .ZIP, or claim a $49 Builder Launch Kit (Tier-1 Verified Shield + HUD Scorecard).

$0 Free CLAUDE.md → $4.99 Config Pack → $9.99 9-File .ZIP → Submit / Boost Tool ($49) → 1-Click $149 Pro OWASP Suite & 30d Pin →