CuraSec

tag: Ai-Ml · 2 items

2026-08-10 · arXiv cs.CR · source ↗ #llm-security#supply-chain#ai-ml
  • Engineer — Learn: Identifies a real supply-chain risk for teams consuming third-party LoRA adapters: a backdoored adapter can alter model output on hidden triggers without modifying base model weights. LoRAScan’s inference-time monitoring approach is worth evaluating if your ML pipelines pull adapters from untrusted registries or Hugging Face.
  • SOC/IR — Learn: No active exploitation, IOCs, or ATT&CK-mappable TTPs to act on; this is foundational research on a threat class. Worth filing as context if your org is building detections around AI/ML pipeline integrity, but no hunt or rule work warranted today.
  • Leader — Learn: Surfaces an emerging supply-chain risk category for AI workloads—untrusted fine-tuned adapters as a malware vector—useful background for shaping AI vendor-risk policy before it becomes a control requirement.
2026-08-03 · The Hacker News · source ↗ #supply-chain#ai-ml#rce
  • Engineer — Plan: If your pipelines load Hugging Face Diffusers models, audit which model repos are consumed and pin to reviewed/trusted sources; check whether you are on the patched Diffusers version once fixes land, as these flaws bypass the trust_remote_code safeguard.
  • SOC/IR — Learn: No active exploitation or IOCs reported; understand that model-loading in ML pipelines can be a code-execution vector and begin thinking about detection coverage for anomalous process spawning from Python ML workloads.
  • Leader — Learn: Illustrates that AI/ML supply chain risk is not theoretical — if your teams consume external model repositories, ask whether a policy governing approved model sources exists before a control is needed.