The paper introduces a curriculum reinforcement learning approach to overcome the cold‑start problem in prompt‑injection red‑teaming of frontier large language models. By training an attacker LLM sequentially against increasingly robust target models and ensuring partial success at each stage, the method achieves high attack success rates (93.8% against GPT‑5.6‑Luna and 45.0% against GPT‑5.6‑Terra) where prior RL methods fail. The attacker LLM also transfers its effectiveness to other frontier models it was not explicitly trained on.
arXiv:2603. 13026v2 Announce Type: replace Abstract: Prompt injection poses serious security risks to real-world LLM applications, particularly autonomous agents.
arXiv:2607. 26115v1 Announce Type: cross Abstract: We introduce \textbf{GPT-Red}, an automated red-teaming agent that is trained to discover novel prompt injection attacks against frontier LLMs.
arXiv:2608. 04317v1 Announce Type: cross Abstract: Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost exclusively against static, heuristic red agents, leaving their robustness against adaptive threats critically understudied.
The paper introduces CoRL, a co-evolutionary reinforcement learning framework designed to defend against adaptive indirect prompt-injection attacks on tool-augmented language agents. CoRL operates in three stages—attacker fine‑tuning, bilateral Co‑PPO training, and defender fine‑tuning—using verifier‑grounded repairs to adapt to changing attack strategies. Experiments on 1,514 executions show that CoRL reduces attack success rates to 0% while improving task utility, demonstrating its effectiveness against adaptive adversaries.
arXiv:2610.00590v1 Announce Type: cross Abstract: An autonomous cyber defender trained with reinforcement learning (RL) is typically tied to the network on which it was trained, limiting its ability...
arXiv:2608. 15372v1 Announce Type: new Abstract: We study generating game-theoretically optimized Courses of Action (COAs) for a Blue UAS swarm against an adaptive Red adversary in a communication-degraded environment, motivated by (but not derived from) a public U.
Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost exclusively against static, heuristic red agents, leaving their robustness against adaptive threats critically understudied. Meanwhile, recent advances in Reinforcement Learning with Verifiable Rewards (RLVR) have improved LLM reasoning, but their integration into cybersecurity remains elusive due to the absence of suitable benchmark environments and interaction datasets.
arXiv:2607. 18063v1 Announce Type: cross Abstract: LLM-based agents process external content, exposing them to prompt injection and multi-turn manipulation.
EvoFlint is an evolutionary atlas that maps multi‑turn vulnerabilities in large language models by treating attack discovery as a search problem rather than a generation task. It uses evolutionary quality‑diversity search to evolve phased conversation plans, employing Pareto fitness for success rate and severity, novelty search for diversity, and a generation‑level memory to incorporate model insights. The resulting risk‑indexed archive, tested on HarmBench, shows high attack success rates across models such as Claude Sonnet, GPT‑5, and Qwen3, revealing which harm categories each model’s safety training covers or misses.
The paper proposes CANOPY, a minimalist reinforcement learning protocol that addresses two common pitfalls—signal starvation and policy drift—in outcome‑only RL for long‑horizon interactive tasks. By scaling same‑task exploration, keeping updates on‑policy, and anchoring updates with KL divergence, CANOPY enables a Qwen3‑14B agent to achieve top leaderboard results on the AppWorld coding benchmark without auxiliary supervision or elaborate scaffolding. The approach also improves performance on SWE‑bench for a Qwen3.5‑9B model.
arXiv:2506. 07468v4 Announce Type: replace Abstract: Conventional large language model (LLM) safety alignment relies on a reactive, disjoint loop: attackers exploit a static model, then defenders patch exposed vulnerabilities.
The paper introduces the SAST-IR framework to evaluate large language models’ robustness against persuasion attacks in a memory‑less setting, revealing a flaw called "Refusal Inertia" that masks true vulnerability. Using the CP‑Agent and a custom CounterFact‑Strict dataset, the authors demonstrate that simple, diverse attack strategies achieve a 96% success rate, while complex attacks often trigger defensive compliance. The study highlights severe brittleness in current state‑of‑the‑art models when deprived of conversation history.