# AutoRefine: Compiling Trajectories into Validated Typed Agent Artifacts

arXiv: https://arxiv.org/abs/2601.22758 · HTML: https://mlsanity.com/paper/2601.22758/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use.
- Builds on 9 of its 24 references that are in the map.
- First posted on arXiv 30 Jan 2026.

## Placed in

- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 9 of 9)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [ALFWorld: Aligning Text and Embodied Environments for Interactive Learning](https://mlsanity.com/paper/2010.03768/) (2020)
- [Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models](https://mlsanity.com/paper/2510.04618/) (2025)
- [TravelPlanner: A Benchmark for Real-World Planning with Language Agents](https://mlsanity.com/paper/2402.01622/) (2024)
- [ScienceWorld: Is your Agent Smarter than a 5th Grader?](https://mlsanity.com/paper/2203.07540/) (2022)
- [EvolveR: Self-Evolving LLM Agents through an Experience-Driven Lifecycle](https://mlsanity.com/paper/2510.16079/) (2025)
- [SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulation](https://mlsanity.com/paper/2406.14991/) (2024)
- [ATLAS: Constraints-Aware Multi-Agent Collaboration for Real-World Travel Planning](https://mlsanity.com/paper/2509.25586/) (2025)

## Built on by (most cited first, 5 of 5)

- [Dynamic Agent Skills: A Lifecycle Survey and Taxonomy of Evolving Skill Libraries](https://mlsanity.com/paper/2607.10113/) (2026)
- [Hierarchical Experimentalist Agents](https://mlsanity.com/paper/2606.29315/) (2026)
- [Rethinking Self-Evolution: A Constrained Exploration-Exploitation Process for Mitigating Skill Overfitting](https://mlsanity.com/paper/2607.26643/) (2026)
- [SkillSpec: Consensus-Gated Agent Skill Evolution via Representation Specialization](https://mlsanity.com/paper/2610.00704/) (2026)
- [TeleTune: Evolving Agent Skills From Offline Telemetry](https://mlsanity.com/paper/2610.05437/) (2026)

## Abstract (arXiv)

Large language model agents repeatedly encounter related tasks, yet systems that learn from trajectories commit every lesson to one predefined artifact form. A local constraint, a reusable procedure, and a delegated objective require different amounts of runtime ownership, so one form either under-specifies the correction or wraps it in execution machinery it does not need. We present AutoRefine, which treats trajectory learning as typed artifact compilation. It contrasts failed against successful trajectories to derive a type-neutral, evidence-linked intervention specification, then compiles that specification into the first Rule, Skill, or bounded Subagent that closes it under a runtime-relative ownership order: the selected schema must own every specified observation, state variable, dependent decision, and completion condition. Validation is stated in the same terms: a type-specific contract gate tests whether the generated object realizes its declared boundary, and a replay gate admits it only when it improves the correction cases linked to its source failures without regression on preservation cases. With GPT-5.6-terra as the shared backbone, AutoRefine records the highest success on ALFWorld, ScienceWorld, TravelPlanner, and SpreadsheetBench, and ties the best result on SkillCraft; on TravelPlanner it reaches 80.56% success against 50.0% for the strongest baseline. Removing boundary closure or replay validation costs 15.00 and 16.11 percentage points, the two largest losses among our construction and admission policies. In a longitudinal TravelPlanner stream, the repository holds 89--91% held-out success after 60 learning tasks with no net loss on previously solved tasks, and frozen repositories improve all 25 evaluated source--target pairs, more within a domain (14.20 points on average) than across domains (6.99).