2 papers
cs.AI2026
TRACER: Per-Tool Context Retention for LLM Agents via Consequence-Attributed Reinforcement Learning
Ziqi Lin, Ye Wu, Mengying Yang +4
Enterprise data agents answer business queries by chaining many tool calls over multiple reasoning steps, routinely accumulating hundreds of thousands of context tokens per session…
cs.CR2026
Toward Metacognitive One-Shot Indirect Prompt Injection: Strategy Abstraction Via Outcome-Conditioned Reflection
Sihan Hou, Xinmeng Hou, Zhijun Zhang +5
Tool-using large language model (LLM) agents are vulnerable to indirect prompt injection (IPI), in which malicious instructions embedded in external observations manipulate subsequ…