Meta open-sources LlamaFirewall, combining PromptGuard 2 (a jailbreak and injection detector), AlignmentCheck (a chain-of-thought auditor for goal hijacking) and CodeShield (static analysis of generated code). On AgentDojo, Meta reports that the combination cut attack success from 17.63% to 1.75% while utility fell from 47.73% to 42.68%.
Why it matters
It is an open guardrail stack whose detector component was later shown to fail under adaptive attack.
Key facts
As stated in the sources, with where to find them.
- AgentDojo baseline: 17.63% ASR, 47.73% utility. PromptGuard 2 86M alone: 7.53% ASR. AlignmentCheck (Llama 4 Maverick) alone: 2.89% ASR, 43.09% utility. Combined: 1.75% ASR, 42.68% utility.Section 4.3.2
Findings that cite this record
No tracked finding cites this record yet.
Key questions this bears on
- Can prompt injection against AI agents be reliably defended?Not reliably. Bounding what untrusted input can trigger helps, but trust boundaries must survive delegation and tool execution.
Sources
Related records
Oct 27, 2026
Oct 20, 2026
Oct 16, 2026
Sep 30, 2026
Mar 24, 2025
Oct 10, 2025