TRACER: Trace-Regularized Training Against Indirect Prompt Injection

Summary

The TRACER paper shows that indirect prompt injection succeeds far more often when the malicious text sits in retrieved context (53%) than when it arrives as a follow-up message (18%). The authors propose TRACER, a channel-aware adversarial training method that combines an answer-anchored, evidence-gated reward with ATRC. On GSM8K with Qwen2.5-3B-Instruct, it cuts RAG-framed attack success from 65.5% to 6.1% while keeping 78-82% clean accuracy.

Mitigation

The source proposes TRACER, a channel-aware adversarial training method, as the defense. It does not describe a patch, fixed version, configuration change or workaround for a specific product.

Topics

• Prompt injection and jailbreaks

Related items

• HighCVE-2026-76066: IntentKit authorization bypass lets callers execute owner-restricted toolsSimilar attack · NVD/CVE Database
• LowAgents at the Edge: Extending AI Security Governance to Physical AI with DefenseClawSimilar attack · Cisco AI Blog
• InfoMMA-SafetyBench: A Benchmark for Multimodal Agent Safety EvaluationSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• InfoPolicy-Verified Self-Training: Learning Agent Compliance from Verifiable Policy RewardsSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• MediumVisual Studio Copilot and SSMS Remote Code Execution (CVE-2026-103572)Similar attack · Embrace The Red