Agent Eng News

Wednesday, August 12, 2026 · Past edition · 16 stories

Retrospective edition reconstructed from later captures. Paper versions may include later updates; coverage labels describe the evidence used.

1. Safety, alignment & governance

   Safety, alignment & governance

   ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents

   ToolHazard addresses a gap in testing tool-using LLM agents: indirect prompt injections can be hidden in environmental states, and attack timing and placement affect effectiveness.

   Why it matters: Generated environments let evaluators vary attack timing and placement across complex workflows. The reported alignment results also suggest this approach can improve security on two benchmarks while preserving benign task utility.

   Original · arxiv (new tab)Read