Verno Labs

Offensive LLMs that red team AI agents across a large library of attack classes.

Verno Labs, formerly Pallma AI, builds custom offensive language models for testing AI agents. Its attack suite spans more than 150 original attack classes, probing how agents fail under adversarial pressure. Founded in 2025, it serves teams that need deeper agent red teaming than prompt-injection checklists provide.

Category: AI Model Security & Red Teaming
Founded: 2025
Funding: $1.6M raised (Pre-Seed, 2026); formerly Pallma AI
Website: https://vernolabs.ai

Seen at Black Hat USA 2026 (Launch Pad sponsor, booth 6103).

Visit website

Outbound clicks are tokenized for directory analytics.