StateFence Defends Blue-Team LLM Agents
Original title: StateFence: Task-State Protection Against Evidence-Chain Contamination in Blue-Team LLM Agents
StateFence preserves task state in SOC LLM agents against evidence-chain contamination from indirect prompt injections.
Indirect prompt injections in telemetry or memory propagate through SOC workflows like triage and IOC extraction. StateFence treats severity labels and IOC sets as protected states and routes defenses by attacked task. Controlled and public-SOC tests show reduced attack success while maintaining utility. Blue-team LLM agents need task-critical state preservation under untrusted evidence.
- Evidence-chain contamination maps to stealth multi-hop attacks
- StateFence evaluates via protected states
- Lowers failures that appear post-poisoning
- Preserves task utility
- Avoids mismatched filtering losses
- AI models
- LLM
From the original article
1 min read at arxiv.orgIndirect prompt injection creates a distinctive risk for blue-team Large Language Model (LLM) agents in Security Operations Centers (SOCs). Malicious instructions can be embedded in telemetry, tickets, retrieval results, or memory and then propagate through alert triage, incident summarization, response recommendation, and Indicator of Compromise (IOC) extraction as locally plausible intermediate decisions.
By Haowen Dai · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.