Agent Safety Is Action Alignment

• URL: http://arxiv.org/abs/2606.28739v1
• Date: Sat, 27 Jun 2026 05:26:43 GMT
• Title: Agent Safety Is Action Alignment
• Abstract summary: We argue that action safety cannot be installed in weights.<n>It must be expressed as emphleast privilege, enforced emphoutside the model at the action boundary.
• Score: 5.990318568221089
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models increasingly act as agents: they call tools, move money, delete records, and send messages on a user's behalf. To keep them safe, practitioners imported the chatbot-era recipe (train the model to refuse unsafe inputs) into the agentic setting, and treat the resulting capability loss as a manageable ``alignment tax.'' We argue this is a \emph{category error}. Refusal is a primitive for \emph{content safety}, where the harm is in the model's output and is therefore a learnable function of it. Agentic harm is different in kind: it lies not in any output but in the relation between the authority an action exercises and the authority the user granted, which is absent from the text the model sees. Importing content-safety methods into this regime does not trade capability for safety; it pays capability and buys negative security. We support this with three lines of evidence spanning the autonomy spectrum: defense-trained models learn surface patterns rather than intent; the same training collapses multi-step agents before any threat appears while leaving them exploitable; and even undefended frontier models exceed granted authority under ordinary use. We conclude that action safety cannot be installed in weights. It must be expressed as \emph{least privilege}, enforced \emph{outside} the model at the action boundary, and evaluated as \emph{action alignment} (a relational, deployment-conditioned property) rather than a refusal score.

Related papers

• Safe responses matter: Output-aware safety guardrail mitigate over-refusal in MLLMs [13.624659633750861]
  Existing safety mechanisms for large language models (MLLMs) face a fundamental trade-off between safety and utility.<n>We propose a paradigm shift toward output-aware safety guardrails.<n>Our method operates within the model's hidden state space to predict whether the forthcoming generation will be unsafe.
  arXiv  Detail & Related papers  (2026-06-20T12:09:13Z)
• ClayBuddy: A Framework, Evaluation, & Mitigation of Coding Agent Failures [1.0742675209112622]
  We study failure modes as stemming from three mechanisms: underspecification, capability errors, and agent harness errors.<n>We propose ClayBuddy, a harness that molds to user preferences and can be modified by the model in-session to mitigate these errors.
  arXiv  Detail & Related papers  (2026-06-13T18:29:53Z)
• Jailbreaking Frontier Foundation Models Through Intention Deception [6.119674554651102]
  Large (vision-) models exhibit remarkable capability but remain highly susceptible to jailbreaking.<n>We introduce a novel multi-turn jailbreaking method that exploits this vulnerability.<n>Our approach also uncovered an additional class of model vulnerability that we call para-jailbreaking.
  arXiv  Detail & Related papers  (2026-04-27T06:12:43Z)
• Compiling Activation Steering into Weights via Null-Space Constraints for Stealthy Backdoors [48.881343993730844]
  Safety-aligned large language models (LLMs) are increasingly deployed in real-world pipelines.<n> adversaries can distribute backdoored checkpoints that behave normally under standard evaluation but jailbreak when a hidden trigger is present.<n>Recent post-hoc weight-editing methods offer an efficient approach to injecting such backdoors.
  arXiv  Detail & Related papers  (2026-04-14T06:48:33Z)
• Sleeper Cell: Injecting Latent Malice Temporal Backdoors into Tool-Using LLMs [0.0]
  Openweight Large Language Models (LLMs) have democratized agentic AI, yet finetuned weights are frequently shared and adopted with limited scrutiny beyond leaderboard performance.<n>This creates a risk where third-party models are incorporated without strong behavioral guarantees.<n>We show that poisoned models maintain state-of-the-art performance on benign tasks, incentivizing their adoption.
  arXiv  Detail & Related papers  (2026-03-02T22:01:08Z)
• Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning? [68.82210578851442]
  We investigate why safety alignment fails in reasoning models through a mechanistic interpretability lens.<n>Using a linear probing approach to trace refusal intentions across token positions, we discover a phenomenon termed as textbfrefusal cliff<n>We propose textbfCliff-as-a-Judge, a novel data selection method that identifies training examples exhibiting the largest refusal cliff to efficiently repair reasoning models' safety alignment.
  arXiv  Detail & Related papers  (2025-10-07T15:32:59Z)
• Can AI Perceive Physical Danger and Intervene? [16.825608691806988]
  New safety challenges emerge when AI interacts with the physical world.<n>How well do state-of-the-art foundation models understand common-sense facts about physical safety?
  arXiv  Detail & Related papers  (2025-09-25T22:09:17Z)
• Oyster-I: Beyond Refusal -- Constructive Safety Alignment for Responsible Language Models [93.5740266114488]
  Constructive Safety Alignment (CSA) protects against malicious misuse while actively guiding vulnerable users toward safe and helpful results.<n>Oy1 achieves state-of-the-art safety among open models while retaining high general capabilities.<n>We release Oy1, code, and the benchmark to support responsible, user-centered AI.
  arXiv  Detail & Related papers  (2025-09-02T03:04:27Z)
• Turning Logic Against Itself : Probing Model Defenses Through Contrastive Questions [50.40122190627256]
  We introduce POATE, a novel jailbreak technique that harnesses contrastive reasoning to provoke unethical responses.<n>PoATE crafts semantically opposing intents and integrates them with adversarial templates, steering models toward harmful outputs with remarkable subtlety.<n>To counter this, we propose Intent-Aware CoT and Reverse Thinking CoT, which decompose queries to detect malicious intent and reason in reverse to evaluate and reject harmful responses.
  arXiv  Detail & Related papers  (2025-01-03T15:40:03Z)
• What Makes and Breaks Safety Fine-tuning? A Mechanistic Study [64.9691741899956]
  Safety fine-tuning helps align Large Language Models (LLMs) with human preferences for their safe deployment. We design a synthetic data generation framework that captures salient aspects of an unsafe input. Using this, we investigate three well-known safety fine-tuning methods.
  arXiv  Detail & Related papers  (2024-07-14T16:12:57Z)
• Who's asking? User personas and the mechanics of latent misalignment [12.92431783194089]
  misaligned capabilities remain latent in safety-tuned models. We show that even when model generations are safe, harmful content can persist in hidden representations. We investigate why certain personas break model safeguards and find that they enable the model to form more charitable interpretations.
  arXiv  Detail & Related papers  (2024-06-17T21:15:12Z)
• Navigating the OverKill in Large Language Models [84.62340510027042]
  We investigate the factors for overkill by exploring how models handle and determine the safety of queries. Our findings reveal the presence of shortcuts within models, leading to an over-attention of harmful words like 'kill' and prompts emphasizing safety will exacerbate overkill. We introduce Self-Contrastive Decoding (Self-CD), a training-free and model-agnostic strategy, to alleviate this phenomenon.
  arXiv  Detail & Related papers  (2024-01-31T07:26:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.