RouteGuard: Internal-Signal Detection of Skill Poisoning in LLM Agents

• URL: http://arxiv.org/abs/2604.22888v1
• Date: Fri, 24 Apr 2026 09:07:05 GMT
• Title: RouteGuard: Internal-Signal Detection of Skill Poisoning in LLM Agents
• Abstract summary: We study pre-execution skill-poison detection and show that successful skill poisoning induces a structured internal effect, attention hijacking.<n>We propose RouteGuard, a frozen-backbone detector that combines response-conditioned attention and hidden-state alignment.
• Score: 26.708117345044325
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Agent skills introduce a new and more severe form of indirect injection for LLM agents: unlike traditional indirect prompt injection, attackers can hide malicious instructions inside a dense, action-oriented skill that already functions as a legitimate instruction source. We study pre-execution skill-poison detection and show that successful skill poisoning induces a structured internal effect, attention hijacking, in which response-time attention shifts from trusted context to malicious skill spans and drives harmful behavior. Motivated by this mechanism, we propose RouteGuard, a frozen-backbone detector that combines response-conditioned attention and hidden-state alignment through reliability-gated late fusion. Across both real and synthetic open-source skill benchmarks, RouteGuard is consistently the strongest or most robust detector; on the critical Skill-Inject channel slice, it reaches 0.8834 F1 and recovers 90.51% of description attacks missed by lexical screening, showing that defending against skill poisoning requires internal-signal detection rather than text-only filtering

Related papers

• SkillSecurer: Detecting and Patching Prompt-Injection Vulnerabilities in AI Agent Skills [0.6263481844384228]
  SkillSecurer is a framework for generating, detecting, localising, and remediating security risks in agent skills.<n>Its red agent generates context-compatible injections across nine threat types while recording the exact modification.<n>Its blue agent analyses complete skill packages, produces grounded evidence, and proposes patches.
  arXiv  Detail & Related papers  (2026-09-12T17:52:06Z)
• ALIBI: Adaptive Agentic Attacks on LLM-Based Vulnerability Detectors via Adversarial Code Comments [12.021872048536766]
  Large language models are increasingly deployed for security-sensitive tasks such as vulnerability detection and code review.<n>Their reliance on natural-language context embedded in source code exposes a previously underexplored attack surface: adversarial comments.<n>We present ALIBI, an automated adaptive black-box attack framework that generates and iteratively refines adversarial comments.
  arXiv  Detail & Related papers  (2026-07-27T18:13:28Z)
• POISE: Position-Aware Undetectable Skill Injection on LLM Agents [23.2593923733415]
  We introduce POISE, a position-aware attack that compresses the trigger into a single, benign-looking body instruction.<n>On Skill-Inject with codex+gpt-5.2, POISE achieves an 89.3% ASR, 28.0 points above a random-placement body baseline and 2.6 points above a YAML-only baseline.
  arXiv  Detail & Related papers  (2026-06-06T02:10:03Z)
• AdversarialCoT: Single-Document Retrieval Poisoning for LLM Reasoning [115.3243260783674]
  Retrieval-augmented generation (RAG) enhances large language model (LLM) reasoning by retrieving external documents.<n>We study knowledge-base poisoning attacks in RAG, where an attacker injects malicious content into the retrieval corpus.<n>We propose AdversarialCoT, a query-specific attack that poisons only a single document in the corpus.
  arXiv  Detail & Related papers  (2026-04-14T02:10:23Z)
• SkillAttack: Automated Red Teaming of Agent Skills through Attack Path Refinement [66.44008181092832]
  LLM-based agent systems increasingly rely on agent skills sourced from open registries to extend their capabilities.<n>We introduce SkillAttack, a framework that verifies skill vulnerability exploitability through adversarial prompting.
  arXiv  Detail & Related papers  (2026-04-05T06:25:11Z)
• Skill-Inject: Measuring Agent Vulnerability to Skill File Attacks [27.120130204872325]
  We introduce SkillInject, a benchmark evaluating the susceptibility of widely-used LLM agents to injections through skill files.<n>SkillInject contains 202 injection-task pairs with attacks ranging from obviously malicious injections to subtle, context-dependent attacks hidden in otherwise legitimate instructions.<n>Our results show that today's agents are highly vulnerable with up to 80% attack success rate with frontier models.
  arXiv  Detail & Related papers  (2026-02-23T18:59:27Z)
• SkillJect: Automating Stealthy Skill-Based Prompt Injection for Coding Agents with Trace-Driven Closed-Loop Refinement [120.52289344734415]
  We propose an automated framework for stealthy prompt injection tailored to agent skills.<n>The framework forms a closed loop with three agents: an Attack Agent that synthesizes injection skills under explicit stealth constraints, a Code Agent that executes tasks using the injected skills and an Evaluate Agent that logs action traces.<n>Our method consistently achieves high attack success rates under realistic settings.
  arXiv  Detail & Related papers  (2026-02-15T16:09:48Z)
• Malice in Agentland: Down the Rabbit Hole of Backdoors in the AI Supply Chain [82.98626829232899]
  Fine-tuning AI agents on data from their own interactions introduces a critical security vulnerability within the AI supply chain.<n>We show that adversaries can easily poison the data collection pipeline to embed hard-to-detect backdoors.
  arXiv  Detail & Related papers  (2025-10-03T12:47:21Z)
• BlindGuard: Safeguarding LLM-based Multi-Agent Systems under Unknown Attacks [58.959622170433725]
  BlindGuard is an unsupervised defense method that learns without requiring any attack-specific labels or prior knowledge of malicious behaviors.<n>We show that BlindGuard effectively detects diverse attack types (i.e., prompt injection, memory poisoning, and tool attack) across multi-agent systems.
  arXiv  Detail & Related papers  (2025-08-11T16:04:47Z)
• Exploring Backdoor Attack and Defense for LLM-empowered Recommendations [15.098844020816552]
  We propose a new attack framework termed Backdoor Injection Poisoning for RecSys (BadRec)<n>BadRec perturbs the items' titles with triggers and employs several fake users to interact with these items, effectively poisoning the training set and injecting backdoors into RecSys.<n>We propose a universal defense strategy called Poison Scanner (P-Scanner) to mitigate such a security threat.
  arXiv  Detail & Related papers  (2025-04-15T13:37:38Z)
• AgentPoison: Red-teaming LLM Agents via Poisoning Memory or Knowledge Bases [73.04652687616286]
  We propose AgentPoison, the first backdoor attack targeting generic and RAG-based LLM agents by poisoning their long-term memory or RAG knowledge base. Unlike conventional backdoor attacks, AgentPoison requires no additional model training or fine-tuning. On each agent, AgentPoison achieves an average attack success rate higher than 80% with minimal impact on benign performance.
  arXiv  Detail & Related papers  (2024-07-17T17:59:47Z)

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.