pith. sign in

arxiv: 2606.12586 · v1 · pith:DLOPRYKKnew · submitted 2026-06-10 · 💻 cs.CR

Beyond Attack Success Rate: Examining Trigger Leakage in Vision-Language Agentic Systems

classification 💻 cs.CR
keywords triggerleakagetriggersattackratevisualwhetheractivation
0
0 comments X
read the original abstract

Vision-Language Agentic Systems (VLAS) connect visual perception to planning, tool use, and physical actions. This means backdoor-type triggers can propagate through both decision pipelines and their connected interfaces, thus making visual backdoors a system-level threat. Current evaluations on such backdoors focus on clean accuracy and attack success rate (ASR), metrics that capture whether a trigger works, but not whether an attack is actually "precise" -- i.e. whether it triggers hidden behaviors only when intended. In this work, we formalize the failure of trigger precision as "trigger leakage": inputs that are visually or semantically close to the intended trigger and therefore inadvertently activate the attacker-specified behavior. To quantify this leakage, we introduce Neighbor Leakage Rate (NLR). Our experiments show that at a 3% poisoning ratio, icon and text triggers remain robust to common visual transformations, but their neighboring variants leak heavily, with NLR reaching 0.996 (icon) and 0.944 (text). Using textual triggers as a controlled probe, we show that standard fine-tuning learns a broad activation region rather than an exact trigger condition, causing neighboring strings to invoke the malicious behavior even when the exact trigger is absent. Adding edit-distance-one hard-negative samples during training substantially narrows this activation region and reduces leakage, including in image-editing and embodied-manipulation workflows, where leaked triggers can propagate into executable programs and action sequences.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.