Scaling Adaptive Prompt Injection Attacks with Agent Swarms
cs.CR, cs.LG
Submitted: 2026-11-09
Updated: 2026-11-09
Comments: 15 pages, 4 figures
Code: https://github.com/facebookresearch/prompt-siren
License: http://creativecommons.org/licenses/by/4.0/
The gist: Prompt injections have emerged as one of the top security risks faced by AI agents today.
Terminology
Abstract
Prompt injections have emerged as one of the top security risks faced by AI agents today. Early work in prompt injection attacks looked at simple tool-call agents and static attack payloads, and were saturated since mid-2025. As of Fall 2026, human red-teaming may be nearing saturation, especially on the Gemini family of models. In this work, we argue that despite this apparent plateau, prompt injection attacks are still possible and continued effort into prompt injection defense is needed. We design a new synthetic evaluation dataset which uses random non-malicious actions as adversarial goals. Such dataset design helps us to evaluate prompt injection robustness as a model ability to respect instruction hierarchy, and disentangle this ability from model knowledge of security practices. We show that an agentic attacker of increased capabilities can reliably trick model to follow instructions in tool output in our evaluation, even for models previously considered robust on public prompt injection benchmarks such as GraySwan Arena. This shows that current frontier defenses may require proactive, automated red-teaming to remain resilient to increasingly capable adversaries.
Sources
- The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections
- RL Is a Hammer and LLMs Are Nails: A Simple Reinforcement Learning Recipe for Strong Prompt Injection
- How Vulnerable Are AI Agents to Indirect Prompt Injections? Insights from a Large-Scale Public Competition
- Universal and Transferable Adversarial Attacks on Aligned Language Models
- Meta-SecAlign: Training LLMs against Prompt Injection for Robust Agents
- May I have your Attention? Breaking Fine-Tuning based Prompt Injection Defenses using Architecture-Aware Attacks
- Assessing Automated Prompt Injection Attacks in Agentic Environments
- Learning to Inject: Automated Prompt Injection via Reinforcement Learning
- GPT-Red: Automated Red Teaming via Self-Play at Scale
- Automated Red Teaming with GOAT: the Generative Offensive Agent Tester
- Lessons from Defending Gemini Against Indirect Prompt Injections
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- Securing AI Agents with Information-Flow Control
- ExploitBench: A Capability Ladder Benchmark for LLM Cybersecurity Agents
Related papers
- From Strategy Search to Policy Learning: MCTS-Guided Adversarial Co-Evolution for LLM Safety Alignment
- Privacy Leakage from Recurrent State in LLM Inference
- Incrementally Verifiable Oracle Aided Computation
- Privilege and confidentiality in generative AI workflows
- BadMoE: Backdooring Mixture-of-Experts LLMs via Routing-Aware Triggers and Task-Underutilized Experts
- CoVault: Secure, Scalable Analytics of Personal Data