paperAugust 2026Unreviewed

ElasticBack: Stealthy Conditional Backdoor in LLM-Agent Skills via Coupled Trigger-Rule Optimization

Hao Sui, Simeng Qin, Jie Liao, Xiao-Jun Jia, Bing Chen, Yang Liu

Abstract

Agent skills, bundles of instructions and resources that an LLM agent loads on demand, form an emerging supply chain where a single poisoned skill can persistently compromise every agent that installs it. However, existing skill attacks either fire on every request or rely on fine-tuned weights or multiple skills, leaving a conditional and low-cost backdoor unexplored. In this work, we present ElasticBack, an effective conditional single-skill backdoor that plants a rule R in the skill document

Framework mappings

OWASP Top 10 for LLM Applications

LLM03Supply Chain
LLM04Data and Model Poisoning

MITRE ATLAS

AML.T0010AI Supply Chain Compromise
AML.T0020Poison Training Data

Suggested from the entry's categories.

Cite

@misc{sui2026elasticback,
  title = {{ElasticBack: Stealthy Conditional Backdoor in LLM-Agent Skills via Coupled Trigger-Rule Optimization}},
  author = {Hao Sui and Simeng Qin and Jie Liao and Xiao-Jun Jia and Bing Chen and Yang Liu},
  year = {2026},
  month = aug,
  eprint = {2608.09577},
  archivePrefix = {arXiv},
  url = {https://www.semanticscholar.org/paper/406fad7e7519f35b9fae4ce4268a40833cf9e403}
}