tool-specifications

Tag

Cards List
#tool-specifications

Tool Specifications Matter: Uncovering and Mitigating Safety Risks in AI Agents

arXiv cs.AI · 2026-08-03 Cached

This paper identifies schema-formatted tool specifications as a primary source of safety degradation in AI agents, weakening LLM refusal signals. The authors propose SafeKeep, an inference-time safeguard that separates safety judgment from tool execution, increasing harmful request refusal rates from 23.8% to 70.6% and cutting prompt injection attack success from 25.6% to 2.5%.

0 favorites 0 likes
← Back to home

Submit Feedback