# Agent-Controlled Forgetting for Tool-Using Agents: Reversible Context Curation in Practice

arXiv: https://arxiv.org/abs/2610.10590 · HTML: https://mlsanity.com/paper/2610.10590/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 15 references that are in the map.
- First posted on arXiv 6 Oct 2026.

## Builds on (most cited first, 11 of 11)

- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models](https://mlsanity.com/paper/2510.04618/) (2025)
- [Scaling Long-Horizon LLM Agent via Context-Folding](https://mlsanity.com/paper/2510.11967/) (2025)
- [ACON: Optimizing Context Compression for Long-horizon LLM Agents](https://mlsanity.com/paper/2510.00615/) (2025)
- [AgentFold: Long-Horizon Web Agents with Proactive Context Management](https://mlsanity.com/paper/2510.24699/) (2025)
- [Learning Agent-Compatible Context Management for Long-Horizon Tasks](https://mlsanity.com/paper/2605.30785/) (2026)
- [FadeMem: Biologically-Inspired Forgetting for Efficient Agent Memory](https://mlsanity.com/paper/2601.18642/) (2026)
- [ACM: Agentic Context Management for Long Horizon Tasks](https://mlsanity.com/paper/2607.23809/) (2026)
- [PRISM: Agentic Retrieval with LLMs for Multi-Hop Question Answering](https://mlsanity.com/paper/2510.14278/) (2025)
- [Selective Forgetting: A Graph-Based Memory Framework for Long-Term LLM Agents](https://mlsanity.com/paper/2608.28978/) (2026)
- [Context Language Models](https://mlsanity.com/paper/2609.37725/) (2026)

## Abstract (arXiv)

Tool-using agents repeatedly carry observations whose useful content can be much smaller than their original payload. We study agent-controlled forgetting: the acting model selects previously observed tool results, replaces each with a short note at its original position, and retains the exact original in a recoverable archive. A Python harness exposes batch archival and explicit recovery without task-specific model training, while protecting user instructions and assistant messages from these operations. In an exploratory OpenTelemetry debugging case followed by an unrelated implementation task, the method ended with 231,951 provider-reported prompt tokens versus 912,492 under retained history, used 50% fewer cumulative input tokens, and had an estimated API cost of USD 1.28-1.44 versus approximately USD 4.38. Both arms passed the two-case primary behavioral oracle; neither fully satisfied the follow-up evaluation. The method made more requests and took 17% longer. A contrasting application-development pair produced no context or cost saving, and an earlier continuation exhibited lower manually assessed quality despite reduced context. These observations demonstrate substantial resource savings in noisy tool-use trajectories and identify workload dependence as a central consideration for reversible context management.