AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents

2026-07-02

AI Summary

This work addresses the challenges of disorganized memory management and entangled memory component effects in long-horizon large language model (LLM) agents. To this end, it proposes a bounded memory contract mechanism that dynamically assembles user messages through typed retrieval for decision-making input, thereby avoiding direct concatenation of cross-turn memories. This approach controls prompt length while enabling modular ablation analysis. The mechanism establishes, for the first time, an ablatable, bounded, and typed memory access framework, offering an interpretable and reproducible memory architecture paradigm for long-horizon LLM agents. Evaluated in the Slay the Spire 2 environment, the method improves win rates from 3/10 to 6/10 when the policy-skill layer is activated. The authors also release a reproducible testbed comprising 298 annotated trajectories, memory snapshots, and analysis tools.

Abstract

Memory for a long-horizon LLM agent is a contract about what each future decision is allowed to see. The simplest contract appends past observations, tool calls, and reflections to every prompt, which makes prior context easy to access but also turns it into a jumbled mixture in which the effect of any single memory component is hard to isolate. We introduce and instrument an alternative bounded contract: every decision is made from a fresh user message assembled by typed retrieval, with no raw cross-decision transcript appended. The prompt thus stays bounded across runs of any length, and any single layer can be ablated in isolation. We instantiate the contract in Slay the Spire 2, a closed-rule stochastic deck-building game whose runs require hundreds of tactical and strategic decisions. A public online benchmark of frontier LLMs on the same game reports zero wins at the lowest difficulty across five configurations, and the developer-reported human win rate at the same difficulty is 16%; the task is hard but not saturated. Within our harness, a fixed-A0 ablation shows the largest observed difference when triggered strategic skills are enabled: the no-store baseline wins 3/10 games and adding the skill layer 6/10. At this sample size the comparison is directional rather than statistically decisive (Fisher exact p\approx0.37); a cross-backbone probe and public accumulating-context baselines are reported as operational comparisons rather than controlled tests of the contract variable itself. We release a reproducible testbed: 298 completed trajectories with condition tags, frozen memory/skill snapshots, prompt records, and analysis scripts -- an agent design and a validated, reusable methodology for studying how explicit memory layers shape long-horizon LLM-agent decisions.

Problem

• long-horizon LLM agents
• bounded memory
• memory contract
• agent decision-making
• explicit memory layers

Innovation

• bounded-memory
• long-horizon agents
• typed retrieval
• memory ablation
• LLM agent testbed