AI Trends

PromptArmor Releases FrontierBreach, an Enterprise AI Agent Security Benchmark

Testing CatalogUpdated Nov 19 08:47

AI summary

PromptArmor has released FrontierBreach, which evaluates how often enterprise AI agent workflows are misled by hidden instructions. Across tests of 21 AI models, the attack success rate was 34.9%. Every frontier model was compromised, with rates of 4.3% for Gemini Argon, 15.2% for Muse Wonder, and 91.4% for Grok 4.5. Attacks hidden in agent skills had a success rate of 45.5%.

Why read this

AI score 69

The benchmark covers 21 models and enterprise agent attack chains, providing a unified reference for comparing the risk of infiltration by hidden instructions in emails, documents, and skills.

Topics

Tags

Quoted post

Shankar Krishnan

@PromptArmor

Today we’re releasing FrontierBreach, a new standard for attack success rate on agentic workflows in the enterprise. Existing benchmarks systematically underrepresent the security risks facing modern AI agents. FrontierBreach reflects attack scenarios on realistic enterprise workflows, while using sophisticated attack chains that result in cyber consequences (ransomware, data exfiltration, phishing, etc).

View original

Today we’re releasing FrontierBreach, a new standard for attack success rate on agentic workflows in the enterprise. Existing benchmarks systematically underrepresent the security risks facing modern AI agents. FrontierBreach reflects attack scenarios on realistic enterprise workflows, while using sophisticated attack chains that result in cyber consequences (ransomware, data exfiltration, phishing, etc).

View original post

Body

PromptArmor has released FrontierBreach, a benchmark of how often AI agents at work can be misled by hidden instructions.

Attackers can hide an invisible line in an email or document, and agents may try to follow it, sending data out or running the attacker's code.

> Across 21 AI models, these hidden attacks worked on 34.9% of PromptArmor's tests.

> PromptArmor breached every frontier model it tested, from Gemini Argon at 4.3% to Muse Wonder at 15.2%, with Grok 4.5 falling for 91.4% of the attacks.

> Attacks hidden inside agent skills also worked in 45.5% of the tests.

Source

Testing Catalog