StealthyPoison: Manipulating LLM Agent Behavior through Procedural Poisoning of Third-Party Skills
cs.CR
Submitted: 2026-11-09
Updated: 2026-11-09
Code: https://github.com/cisco-ai-defense/skill-scanner
Terminology
Sources
- gpt-oss-120b & gpt-oss-20b Model Card
- Overcoming the Retrieval Barrier: Indirect Prompt Injection in the Wild for LLM Systems
- M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation
- Defeating Prompt Injections by Design
- A Survey on LLM-as-a-Judge
- MalTool: Malicious Tool Attacks on LLM Agents
- SkillJect: Effectively Automating Skill-Based Prompt Injection for Skill-Enabled Agents
- SPLADE-v3: New baselines for SPLADE
- Llama2Vec: Unsupervised Adaptation of Large Language Models for Dense Retrieval
- SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks
- Under the Hood of SKILL.md: Semantic Supply-chain Attacks on AI Agent Skill Registry
- Skill-Inject: Measuring Agent Vulnerability to Skill File Attacks
- Skill Retrieval Augmentation for Agentic AI
- The Llama 3 Herd of Models
- Qwen3 Technical Report
- The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions
- GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning
- C-Pack: Packed Resources For General Chinese Embeddings
- TRUSTDESC: Preventing Tool Poisoning in LLM Applications via Trusted Description Generation
- SkillRouter: Skill Routing for LLM Agents at Scale
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