Remember When It Matters: Proactive Memory Agent for Long-Horizon Agents

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Yifan Wu, Lizhu Zhang, Yuhang Zhou, Mingyi Wang, Bo Peng, Serena Li, Xiangjun Fan, Zhuokai Zhao
• Venue: arXiv: cs.AI
• Published: 2026-07-09
• Collected: 2026-07-10
• URL: http://arxiv.org/abs/2607.08716v1
• PDF: https://arxiv.org/pdf/2607.08716v1
• DOI: UNKNOWN

Plain-Language Brief

Researchers propose a proactive memory agent for long-horizon tasks where decision-relevant information can get buried as trajectories grow. The agent updates a structured memory bank from the recent trajectory and decides when to inject reminders, improving performance on Terminal-Bench 2.0 and $τ^2$-Bench.

What the Document Says

• The paper introduces a proactive memory agent for long-horizon tasks where the action agent must surface relevant state information.
• The memory agent updates a structured memory bank from the recent trajectory and decides whether to inject a reminder or remain silent.
• The module is designed to be plug-and-play with existing action agents and harnesses.
• The research shows improvements in performance on Terminal-Bench 2.0 and $τ^2$-Bench.
• Ablations show that selective intervention outperforms other methods like passive bank exposure, always-on injection, advisor-only guidance, and general retrieval.
• Training Qwen3.5-27B on SETA using SFT and GRPO improves validation reward and achieves partial transfer to Terminal-Bench.

Key Concepts Mentioned

• proactive memory agent
• long-horizon tasks
• memory bank
• reminder injection

Claims Made by the Document

• The proactive memory agent improves performance on long-horizon tasks.
• Selective intervention outperforms other methods.

Evidence Used

• Improvements in pass@1 for both weaker and stronger action agents on Terminal-Bench 2.0 and $τ^2$-Bench.

Methods / Approach

• Proactive memory agent
• Plug-and-play with existing action agents

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

In long-horizon tasks, decision-relevant state is often scattered across an expanding trajectory, while the action agent must surface it and act. As trajectories grow, task requirements, environment facts, prior attempts, diagnoses, and open subgoals can be buried in the context window or pushed beyond it, failing to influence decisions when needed. We call this failure mode 'behavioral state decay'. We study memory as an active intervention mechanism rather than passive retrieval. A separate memory agent runs alongside an unmodified action agent, updating a structured memory bank from the recent trajectory and deciding whether to inject a memory-grounded reminder or remain silent. The module is plug-and-play with frontier action agents and existing agent harnesses. Across Terminal-Bench 2.0 and $τ^2$-Bench, it improves pass@1 for both weaker and stronger action agents, with gains of +8.3 pp on Terminal-Bench and +6.8 pp on $τ^2$-Bench. Ablations show that selective intervention outperforms passive bank exposure, always-on injection, advisor-only guidance, and general retrieval. As an early step toward open-weight memory policies, we train Qwen3.5-27B on SETA using SFT and GRPO, improving validation reward and achieving partial transfer to Terminal-Bench.