# Strong but Brittle: Simple Attacks Subvert Reasoning-based Safety Guardrails

arXiv: https://arxiv.org/abs/2510.11570 · HTML: https://mlsanity.com/paper/2510.11570/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 33 of its 42 references that are in the map.
- First posted on arXiv 13 Oct 2025.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)
- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 30 of 33)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal](https://mlsanity.com/paper/2402.04249/) (2024)
- [gpt-oss-120b & gpt-oss-20b Model Card](https://mlsanity.com/paper/2508.10925/) (2025)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks](https://mlsanity.com/paper/2404.02151/) (2024)
- [Deliberative Alignment: Reasoning Enables Safer Language Models](https://mlsanity.com/paper/2412.16339/) (2024)
- [OR-Bench: An Over-Refusal Benchmark for Large Language Models](https://mlsanity.com/paper/2405.20947/) (2024)
- [Phi-4-reasoning Technical Report](https://mlsanity.com/paper/2504.21318/) (2025)
- [SafeChain: Safety of Language Models with Long Chain-of-Thought Reasoning Capabilities](https://mlsanity.com/paper/2502.12025/) (2025)
- [Language Models are Homer Simpson! Safety Re-Alignment of Fine-tuned Language Models through Task Arithmetic](https://mlsanity.com/paper/2402.11746/) (2024)
- [H-CoT: Hijacking the Chain-of-Thought Safety Reasoning Mechanism to Jailbreak Large Reasoning Models, Including OpenAI o1/o3, DeepSeek-R1, and Gemini 2.0 Flash Thinking](https://mlsanity.com/paper/2502.12893/) (2025)
- [STAIR: Improving Safety Alignment with Introspective Reasoning](https://mlsanity.com/paper/2502.02384/) (2025)
- [International AI Safety Report](https://mlsanity.com/paper/2501.17805/) (2025)
- [On the Societal Impact of Open Foundation Models](https://mlsanity.com/paper/2403.07918/) (2024)
- [Reasoning-Augmented Conversation for Multi-Turn Jailbreak Attacks on Large Language Models](https://mlsanity.com/paper/2502.11054/) (2025)
- [RealSafe-R1: Safety-Aligned DeepSeek-R1 without Compromising Reasoning Capability](https://mlsanity.com/paper/2504.10081/) (2025)
- [A Mousetrap: Fooling Large Reasoning Models for Jailbreak with Chain of Iterative Chaos](https://mlsanity.com/paper/2502.15806/) (2025)
- [Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety](https://mlsanity.com/paper/2503.05021/) (2025)
- [AdvPrefix: An Objective for Nuanced LLM Jailbreaks](https://mlsanity.com/paper/2412.10321/) (2024)
- [A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models](https://mlsanity.com/paper/2509.03871/) (2025)
- [Reasoning as an Adaptive Defense for Safety](https://mlsanity.com/paper/2507.00971/) (2025)
- [SafeKey: Amplifying Aha-Moment Insights for Safety Reasoning](https://mlsanity.com/paper/2505.16186/) (2025)
- [Safety Reasoning with Guidelines](https://mlsanity.com/paper/2502.04040/) (2025)
- [Enhancing Model Defense Against Jailbreaks with Proactive Safety Reasoning](https://mlsanity.com/paper/2501.19180/) (2025)
- [Three Minds, One Legend: Jailbreak Large Reasoning Model with Adaptive Stacked Ciphers](https://mlsanity.com/paper/2505.16241/) (2025)
- [Practical Reasoning Interruption Attacks on Reasoning Large Language Models](https://mlsanity.com/paper/2505.06643/) (2025)
- [AutoRAN: Automated Hijacking of Safety Reasoning in Large Reasoning Models](https://mlsanity.com/paper/2505.10846/) (2025)
- [Why Safeguarded Ships Run Aground? Aligned Large Language Models' Safety Mechanisms Tend to Be Anchored in The Template Region](https://mlsanity.com/paper/2502.13946/) (2025)

## Built on by (most cited first, 5 of 5)

- [Prompt Injection as Role Confusion](https://mlsanity.com/paper/2603.12277/) (2026)
- [Unreal Thinking: Chain-of-Thought Hijacking via Two-stage Backdoor](https://mlsanity.com/paper/2604.09235/) (2026)
- [Jailbreaking LLMs via Calibration](https://mlsanity.com/paper/2602.00619/) (2026)
- [Automatic Red Teaming for Implicit Vulnerabilities of Text-to-Image Models](https://mlsanity.com/paper/2609.06094/) (2026)
- [PyINE: A Framework for Scalable Elicitation and Oversight via Code Execution](https://mlsanity.com/paper/2610.04737/) (2026)

## Abstract (arXiv)

Open-weight Large Reasoning Models (LRMs) are approaching the capabilities of their frontier counterparts but pose significant safety concerns, as they are difficult to patch or monitor post-release. To prevent misuse, reasoning-based safety guardrails, where models explicitly reason on safety justifications before answering, have become a promising primary defense, achieving near-perfect refusal rates on harmful queries. We show that this strong defense is alarmingly brittle and can be subverted by embarrassingly simple attacks to elicit extremely forbidden questions, such as `How to kill a man without being caught?' Specifically, we identify one systematic vulnerability from the reasoning-then-answer mechanism: the stage-transition logic that governs when safety reasoning begins and ends can be trivially manipulated. Based on this finding, we develop four simple yet effective red-teaming methods that systematically subvert different stages of the guardrails, either by bypassing reasoning entirely or exploiting it to produce targeted harmful content. These methods achieve attack success rates up to 90% across five benchmarks on multiple LRM families. Our findings reveal that reasoning-based guardrails are necessary but not sufficient and must be paired with robust triggering mechanisms and stronger base-model alignment. Code is in https://chenxshuo.github.io/bag-of-tricks/.