Pith.

Artificial Intelligence · arXiv:2609.04869 v1 · submitted 4 Sep 2026 · 24 references

From Interaction Traces to Persistent Skills: Online Evolution for Computer-Use Agents

Longtao Hu, Xiao Liang, Linchao Zhu

The paper says

“Computer-use agents can execute increasingly complex tasks in graphical interfaces, but their interaction experience is typically transient: procedural knowledge acquired from one rollout is not systematically retained, refined, and reused in later tasks.”

more of the abstract

Existing skill libraries provide external procedural knowledge, yet their incremental value over the same agent operating without skills, as well as their longitudinal dynamics under repeated interaction, remain insufficiently characterized. We present an online skill-evolution framework that converts interaction trajectories and evaluator feedback into a persistent, versioned library of reusable procedures. Each iteration executes against a frozen library snapshot, and evidence-guided skill updates become available in subsequent iterations without changing model parameters. We compare the full evolving-library system with a configuration-matched empty-library control across four OSWorld application domains under the same fixed action-generation and GUI-grounding stack, task sets, and iteration horizons. Following a five-iteration empty-library warm-up, Full attains a higher post-warm-up mean evaluator score in all four observed domain runs, with mean differences ranging from 5.7 to 18.6 percentage points and domain-dependent temporal stability. In GIMP, provenance-aware analysis reveals retrieval across task-of-origin boundaries and revision churn, where repeated accepted edits fail to recover the originating task. These findings characterize evolving skill libraries as auditable, shared procedural memory that can improve a fixed computer-use stack, while showing that their benefits are conditional and repeated revision does not guarantee recovery. Code is released at https://github.com/LongtaoHu/Skill-Evo4GUI.

from the abstract

Pith says

Machine review · deepseek-v4-flash · 8 Sep 2026 · what was read was not recorded · not checked by a person

It can be wrong. Challenge this review or reply to it.

Holds up, with conditions. Honest, well-scoped system paper whose real value is the versioned skill-library framework and the GIMP provenance findings; the four-domain performance claim is descriptive and under-powered, exactly as the authors admit. This line is wrong

3 major objections · 5 minor

How often objections from reviews like this one are right: not yet measured

Are you an author? What you can do.

Not checked: figures, code, what the 24 references say.

The paper says, at §4.2, Table 2

The review did not quote this passage.

Pith says: Objection 1 of 3 · major

The Writer domain is the clearest example of why the post-warm-up differences cannot be read as evidence of skill value: the Full-minus-Empty difference during the empty-library warm-up is already +9.1 percentage points, while the post-warm-up difference is +11.8, so 77% of the post-warm-up gap is present before any skill is available.

more of objection 1

The paper acknowledges this in §4.2 and §5, but the abstract and Contribution 2 still assert that the system 'can add measurable value' and 'demonstrate the value' of the evolving library; the Writer observation does not support that inference without an adjustment for the warm-up offset, such as a difference-in-differences or change-score analysis.

The paper says, at §4.2, Figure 2

The review did not quote this passage.

Pith says: Objection 2 of 3 · major

With one independently executed longitudinal run per condition–domain pair, there is no within-domain variance estimate, no confidence interval, and no significance test. The statement that Full has a higher post-warm-up mean 'in all four observed domain runs' is descriptively accurate, but the abstract's stronger claim that the library 'can improve a fixed computer-use stack' and Contribution 2's claim to 'demonstrate the value' require replication.

more of objection 2

At minimum, the authors should report multiple independent seeds per condition–domain pair, or a bootstrap over task-level scores within a run, and a paired or permutation test on the post-warm-up means or on the change from the warm-up period.

The paper says, at §4.3

The review did not quote this passage.

Pith says: Objection 3 of 3 · major

The cross-origin retrieval share (43.3% of recorded skill calls) supports only the modest claim that skills are consulted beyond their originating tasks; it does not measure whether the Executor followed the returned procedure or whether the retrieval improved the outcome.

more of objection 3

The authors explicitly disclaim causal attribution for telemetry, and the revision-churn analysis is likewise observational. The abstract and Conclusion should therefore not present these results as evidence that the library constitutes 'shared procedural memory that can improve' the stack without the same conditional language used in §4.3 and §5.

5 minor objections

The paper says, at Equation (2)

Pith says: minor objection 1 of 5

The notation S^Full_t = ∅ for t = 0,...,4 is clear, but the equation does not define how the library evolves for t ≥ 5; a compact definition of S^Full_{t+1} in terms of accepted mutations would make the protocol easier to follow.

The paper says, at Table 2

Pith says: minor objection 2 of 5

The column heading 'Pre diff. (pp)' could be misread as a performance score; consider renaming it 'Warm-up diff. (pp), t = 0–4' so it is clear that this is the Full-minus-Empty difference during the period when both libraries are empty.

show 3 more

The paper says, at §4.3

Pith says: minor objection 3 of 5

The statistic '82.4% of valid rollouts invoke get_skill at least once' needs a definition of 'valid rollout' (e.g., rollouts with complete telemetry) before the denominator is interpretable.

The paper says, at Abstract and §1

Pith says: minor objection 4 of 5

The abstract gives the code URL as 'https://github.com/LongtaoHu/Skill-Evo4GUI', while the full text gives 'https://github.com/LongtaoHu/Skill-Evo4GUI.git'; please harmonize.

The paper says, at §3.4

Pith says: minor objection 5 of 5

The serial decision mode is described as order-dependent, which is correct and acknowledged, but it would help to state explicitly in Table 2 or the protocol description that the reported results use the serial mode only, since a parallel mode could change the revision dynamics and hence the outcome.

What would settle it

Compute a difference‐in‐differences (DiD) estimator per domain: DiD = (Full_post_mean − Empty_post_mean) − (Full_pre_mean − Empty_pre_mean). Perform a one‐sided permutation test: for each of 10,000 permutations, independently flip the condition labels (Full vs. Empty) within each domain, recompute the average DiD across the four domains, and record the null distribution. If the observed average DiD falls above the 95th percentile, the skill contribution is statistically significant beyond pre‐existing variation. If not, the post‐warm‐up advantage is consistent with run‐specific noise.

from the skeptic's reading

How the review went

4 readings by deepseek-v4-flash · 8 Sep 2026

  1. Reader holds up, with conditions
  2. Skeptic agrees 29 s · 10,136 tokens read, 3,055 written
  3. Editor wrote the verdict above 51 s · 11,099 tokens read, 5,723 written
  4. Referee major revision: 3 major objections, 5 minor 37 s · 10,660 tokens read, 3,744 written

Not run on this paper: the simulated authors' defence.

Pith did not check this review's objections against the paper.

The review in full

The review in brief, machine-written

Two things to know. First, the real value here is the framework and the provenance analysis, not the headline numbers. Second, the headline claim — Full beats the empty-library control in all four domains — is a descriptive comparison the authors themselves decline to call a significance test or a causal estimate. On the evidence as presented, that restraint is correct.

The framework is cleanly designed. Tasks execute against a frozen library snapshot; accepted mutations are committed to a versioned git-backed library and only reach the next snapshot; retrieval is progressive; all model parameters stay fixed. The empty-library control is configuration-matched, including the catalog prompt and get_skill interface, which is the right way to isolate the library's contribution. Code is released.

The genuinely new material is the GIMP provenance analysis. 43.3% of retrievals target skills created by a different task — real evidence that the library operates as shared procedural memory, not per-task scratch storage. The revision-churn case is a useful negative result: gimp-add-alpha-channel receives repeated accepted edits while its originating task succeeds in only 2 of 35 iterations. That's the kind of observation the skill-library literature mostly does not print.

The soft spot is the performance comparison, and it is soft in the way the stress test says. Writer already differs by +9.1 pp during the empty-library warm-up against a +11.8 pp post-warm-up difference, so most of Writer's margin is pre-existing. With one run per condition–domain pair, stochastic persistence cannot be ruled out.

But the stress test overstates the damage. Across the other three domains the pattern is friendlier to the paper than the Writer cell alone suggests: GIMP and Thunderbird reverse from negative to positive exactly when the library activates (−4.0→+5.7, −6.7→+12.4), and VLC's gap quadruples (+4.7→+18.6). A difference-in-differences reading supports the skill-availability hypothesis in three of four domains; the paper itself flags Writer and declines to claim causality. So the central claim is plausible but under-supported, not refuted.

This paper deserves a serious referee. The path forward is replicated runs with significance testing, or at minimum a difference-in-differences adjustment for the warm-up imbalance, plus held-out tasks. The framework and the qualitative findings stand regardless of how replications land. My recommendation: accept for peer review with a major-revision ask focused on the statistics, not the system, and engage with it — it is an honest, reproducible artifact with a finding the community should see.

The strongest objection

The central empirical claim—that the evolving-library system outperforms the empty‐library control—rests on the post‐warm‐up mean difference across four domains. However, the Full and Empty runs already diverge during the five‐iteration empty‐library warm‐up: the pre‐period difference ranges from −6.7 to +9.1 percentage points (Table 2). This indicates that stochastic factors (e.g., randomness in model outputs, environment initialization, or task order) can produce substantial performance gaps even when no skill library is available. Without multiple independent runs, the post‐warm‐up advantage cannot be uniquely attributed to skill availability rather than to the persistence or amplification of these pre‐existing differences. The Writer domain is particularly problematic: the pre‐warm‐up difference of +9.1 pp accounts for 77% of the post‐warm‐up difference of +11.8 pp, so the incremental effect of skills could be as small as 2.7 pp. The paper acknowledges this limitation but does not provide any quantitative adjustment. Therefore, the claim that the skill library adds measurable value is not adequately supported by the evidence as presented.

The reader correctly identified the weakest assumption as the inability to attribute post-warm-up differences to skill availability rather than pre-existing stochastic differences. This is the single most load-bearing concern because the paper’s central claim hinges on the comparison between Full and Empty. Without controlling for pre-existing differences, the post-warm-up advantage could be an artifact of run-to-run variation, especially in Writer where the pre-difference is large. A concrete difference-in-differences analysis with a permutation test would directly address this concern by isolating the incremental effect of skills relative to the warm-up period. The paper’s own limitation paragraph acknowledges this issue, and the reader’s CONDITIONAL verdict appropriately flags it. My stress test does not uncover a new concern beyond what the reader already identified; therefore, the verdict should remain unchanged. The non-finding here reinforces the need for multiple runs or a DiD adjustment, which the paper itself lists as future work.

The referee's report

If the findings are confirmed with replications, the paper would make a useful contribution to external memory for GUI agents: it separates execution from evolution, versiones skill changes, and provides provenance machinery that goes beyond aggregate endpoint scores. The frozen-snapshot design and the telemetry-based provenance tracking are sensible methodological choices, and the release of code is a concrete asset. The paper is also admirably candid about the observational nature of its comparisons. However, the empirical core rests on four single-run comparisons in which pre-existing warm-up imbalances are not controlled for, so the headline claim that the evolving library 'can improve a fixed computer-use stack' is currently supported only as a descriptive observation, not as an established causal effect.

Pith's reading of the paper (a machine's summary, not the authors')

This paper aims to establish that a computer-use agent can improve its performance by converting its own interaction trajectories and evaluator feedback into a persistent, versioned library of reusable skills, without modifying the underlying action-generation or GUI-grounding models. The authors compare two conditions across four OSWorld application domains: a Full system that accumulates and retrieves skills, and an Empty control that retains no skills. They find that Full achieves higher post-warm-up mean evaluator scores in all four domains, with differences ranging from 5.7 to 18.6 percentage points. However, these results are descriptive: each condition–domain pair uses only one independently executed run, and warm-up imbalances (most notably in Writer) limit causal attribution. In GIMP, provenance analysis reveals that 43.3% of skill retrievals are cross-origin, meaning skills are reused beyond the tasks that created them, but also that repeated accepted edits can fail to recover the originating task—a pattern the paper terms revision churn. The core contribution is a longitudinal, provenance-rich characterization of skill evolution as a shared, auditable adaptation layer whose value is domain-dependent and not guaranteed.

What carries the argument. The central mechanism is an iterative pipeline: (1) iteration-frozen execution under a read-only library snapshot $S_t$, (2) trace abstraction converting rollouts into structured facts, (3) evidence-guided proposal generation (create/edit/delete/no-op), (4) versioned commit of accepted mutations to form $S_{t+1}$, and (5) progressive retrieval at runtime via a lightweight catalog and on-demand skill body retrieval. The comparison uses the Full condition (catalog populated) against an Empty condition (identical interface, empty catalog), with all other components (Executor, Grounding model, prompt templates, task sets) held fixed.

Core claim. The paper presents a framework where a fixed computer-use agent (Executor and Grounding models unchanged) repeatedly executes tasks under a frozen skill snapshot, then converts trajectories and evaluator feedback into structured facts that guide skill creation, revision, or preservation. Accepted mutations form the next snapshot, separating execution from evolution. Through a matched longitudinal comparison across four OSWorld domains, the Full system (with evolving library) achieves a higher mean evaluator score than the Empty control in all observed runs after the warm-up period. In VLC the difference reaches 18.6 percentage points; in Writer a 9.1-point pre-existing warm-up offset accounts for most of the 11.8-point post-warm-up gap. The paper also provides provenance evidence from GIMP that skills are retrieved across task boundaries (56.7% same-task, 43.3% cross-origin) and that revision churn can persist—the background-transparency task succeeds only 2 of 35 times despite repeated edits.

The premise it rests on. The post-warm-up performance differences are attributed to skill availability, but the separately executed Full and Empty conditions already show non-trivial warm-up differences (e.g., Writer +9.1), so not all of the later gap can be uniquely credited to the library.

If the paper is right (Pith's inferences, not the authors')

Computer-use agents can accumulate procedural knowledge without updating model parameters, enabling continual adaptation through external memory alone.

Skill libraries can serve as auditable shared memory: version history and provenance allow tracing a skill's origin, edits, and retrieval contexts.

Cross-origin retrieval (43% in GIMP) suggests that libraries facilitate transfer of procedures across related tasks, offering a path toward generalizable skill reuse.

Revision churn (repeated accepted edits without task recovery) indicates that current evidence-guided evolution policies can be counterproductive; smarter policies (e.g., validation or rollback) may be needed.

The framework's separation of execution from evolution allows independent replication and ablation of the library component, making it a test bed for studying procedural memory dynamics.

The single-run design means the observed performance advantages could be stochastic; replication with multiple independent runs is essential to assess robustness and quantify effect sizes.

The warm-up imbalance in Writer (Full already +9.1 points before any skills exist) suggests that factors other than skill availability (e.g., random success on earlier tasks) can account for much of the later gap, weakening causal claims about skill effectiveness in that domain.

Revision churn in GIMP implies that the Proposer's failure-driven edits may introduce noise rather than convergence; this points to a need for reinforcement learning over library-mutation policies, as the paper itself suggests.

The framework's reliance on a fixed Executor–Grounding stack means that gains are bottlenecked by those components; combining skill evolution with improvements in the base agent might yield larger and more consistent benefits.

Check this review

Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.

How often reviews like this one are right

Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.

How often objections from reviews like this one are right: not yet measured

Every value above, in the review's public record (JSON)

Reviews by people

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review

How Pith checks papers · The full audit of this paper

The paper

Abstract

Computer-use agents can execute increasingly complex tasks in graphical interfaces, but their interaction experience is typically transient: procedural knowledge acquired from one rollout is not systematically retained, refined, and reused in later tasks. Existing skill libraries provide external procedural knowledge, yet their incremental value over the same agent operating without skills, as well as their longitudinal dynamics under repeated interaction, remain insufficiently characterized. We present an online skill-evolution framework that converts interaction trajectories and evaluator feedback into a persistent, versioned library of reusable procedures. Each iteration executes against a frozen library snapshot, and evidence-guided skill updates become available in subsequent iterations without changing model parameters. We compare the full evolving-library system with a configuration-matched empty-library control across four OSWorld application domains under the same fixed action-generation and GUI-grounding stack, task sets, and iteration horizons. Following a five-iteration empty-library warm-up, Full attains a higher post-warm-up mean evaluator score in all four observed domain runs, with mean differences ranging from 5.7 to 18.6 percentage points and domain-dependent temporal stability. In GIMP, provenance-aware analysis reveals retrieval across task-of-origin boundaries and revision churn, where repeated accepted edits fail to recover the originating task. These findings characterize evolving skill libraries as auditable, shared procedural memory that can improve a fixed computer-use stack, while showing that their benefits are conditional and repeated revision does not guarantee recovery. Code is released at https://github.com/LongtaoHu/Skill-Evo4GUI.

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.

abstract compiled · 2026-09-23

The authors' main claim, as the classifier picked it out 1.00

“These findings characterize evolving skill libraries as auditable, shared procedural memory that can improve a fixed computer-use stack, while showing that their benefits are conditional and repeated revision does not guarantee recovery.”

How the main claim is stated: asserted 0.98

What the abstract itself states, as probabilities: claims first 0.07 · states limitation 0.95 · human participants 0.09 · controlled comparison 0.99 · randomized assignment 0.03 · states quantitative result 0.99 · compares with prior results 0.93 · mentions code or data availability 0.98

The abstract, sentence by sentence

This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.

SentenceRoleThe authors' own workStates the main conclusionStrength
Computer-use agents can execute increasingly complex tasks in graphical interfaces, but their interaction experience is typically transient: procedural knowledge acquired from one rollout is not systematically retained, refined, and reused in later tasks. background 0.81 0.07 0.04 asserted 0.98
Existing skill libraries provide external procedural knowledge, yet their incremental value over the same agent operating without skills, as well as their longitudinal dynamics under repeated interaction, remain insufficiently characterized. gap 0.99 0.04 0.03 asserted 0.95
We present an online skill-evolution framework that converts interaction trajectories and evaluator feedback into a persistent, versioned library of reusable procedures. contribution 0.91 0.93 0.09 asserted 1.00
Each iteration executes against a frozen library snapshot, and evidence-guided skill updates become available in subsequent iterations without changing model parameters. method 0.99 0.78 0.06 asserted 1.00
We compare the full evolving-library system with a configuration-matched empty-library control across four OSWorld application domains under the same fixed action-generation and GUI-grounding stack, task sets, and iteration horizons. method 0.97 0.93 0.03 asserted 0.91
Following a five-iteration empty-library warm-up, Full attains a higher post-warm-up mean evaluator score in all four observed domain runs, with mean differences ranging from 5.7 to 18.6 percentage points and domain-dependent temporal stability. result 1.00 0.97 0.27 asserted 1.00
In GIMP, provenance-aware analysis reveals retrieval across task-of-origin boundaries and revision churn, where repeated accepted edits fail to recover the originating task. result 0.98 0.92 0.07 asserted 1.00
These findings characterize evolving skill libraries as auditable, shared procedural memory that can improve a fixed computer-use stack, while showing that their benefits are conditional and repeated revision does not guarantee recovery. interpretation 1.00 0.91 0.87 asserted 0.98
Code is released at https://github.com/LongtaoHu/Skill-Evo4GUI. contribution 0.72 0.48 0.02 asserted 0.79

What kind of paper the classifier reports this is: method or tool 0.96 · evidence from observation or experiment 0.99

What was examined, and what was not

Sources read

  • arXiv listing abstract v1

Not examined

  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers

Levels run

  • P abstract

Not run

  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

How the labels were made

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 7,050 input tokens.

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON

Figures (2)
Figure 1 of the paper
Figure 1. Online skill evolution. Execution under St produces trajectories that are abstracted and diagnosed to create, edit, delete, or preserve skills. Accepted changes form St+1. The Executor sees only a lightweight catalog unless it explicitly retrieves a complete skill body. External memory offers a way to retain and reuse interaction experience without updating model parameters. Prior agents store verbal reflections or distilled experiential insights (Shinn et al., 2023; Zhao et al., 2023), while subsequent systems represent reusable behavior as workflows, executable programs, or…
Figure 2 of the paper
Figure 2. Longitudinal performance across four GUI application domains. Purple curves show the Full

Figures are the authors', reproduced from the paper.

References (24)
  1. [1] Salaheddin Alzubi, Noah Provenzano, Jaydon Bingham, Weiyuan Chen, and Tu Vu. EvoSkill: Automated skill discovery for multi-agent systems.arXiv preprint arXiv:2603.02766, on PithPith's reviewarXiv
  2. [3] doi: 10.48550/arXiv.2606.01993. Tianyi Chen, Yinheng Li, Michael Solodko, et al. CUA-Skill: Develop skills for computer using agent.arXiv preprint arXiv:2601.21123, on Pithdoi:10.48550/arxiv.2606.01993
  3. [4] doi: 10.48550/arXiv.2601.21123. Kanzhi Cheng, Qiushi Sun, Yougang Chu, Fangzhi Xu, Yantao Li, Jianbing Zhang, and Zhiyong Wu. SeeClick: Harnessing GUI grounding for advanced visual GUI agents.arXiv preprint arXiv:2401.10935, on Pithdoi:10.48550/arxiv.2601.21123
  4. [6] doi: 10.48550/arXiv.2604.10923. Jinyi Han, Yuanjian Xu, Ying Liao, et al. Skill-Use: Can LLMs actually use skills in agentic harnesses? arXiv preprint arXiv:2608.04828, on PithPith's reviewarXivdoi:10.48550/arxiv.2604.10923
  5. [7] doi: 10.48550/arXiv.2608.04828. Guanyu Jiang, Zhaochen Su, Xiaoye Qu, and Yi R. Fung. XSkill: Continual learning from experience and skills in multimodal agents.arXiv preprint arXiv:2603.12056, on PithPith's reviewarXivdoi:10.48550/arxiv.2608.04828
  6. [8] doi: 10.48550/arXiv.2603.12056. Kimi Team. Kimi K2.5: Visual agentic intelligence.arXiv preprint arXiv:2602.02276, on PithPith's reviewarXivdoi:10.48550/arxiv.2603.12056
  7. [9] doi: 10.48550/ arXiv.2602.02276. Xiangyi Li, Yimin Liu, Wenbo Chen, et al. SkillsBench: Benchmarking how well agent skills work across diverse tasks.arXiv preprint arXiv:2602.12670, on PithPith's reviewarXivdoi:10.48550/arxiv.2602.02276
  8. [10] doi: 10.48550/arXiv.2602.12670. 9 Huawei Lin, Peng Li, Jie Song, Fuxin Jiang, and Tieying Zhang. MUSE-Autoskill: Self-evolving agents via skill creation, memory, management, and evaluation.arXiv preprint arXiv:2605.27366, on PithPith's reviewarXivdoi:10.48550/arxiv.2602.12670
  9. [11] doi: 10.48550/arXiv.2605.27366. Siru Ouyang, Jun Yan, Yanfei Chen, Rujun Han, et al. SkillOS: Learning skill curation for self-evolving agents.arXiv preprint arXiv:2605.06614, on PithPith's reviewarXivdoi:10.48550/arxiv.2605.27366
  10. [12] doi: 10.48550/arXiv.2605.06614. Noah Shinn, Federico Cassano, Edward Berman, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. Reflexion: Language agents with verbal reinforcement learning.arXiv preprint arXiv:2303.11366, on PithPith's reviewarXivdoi:10.48550/arxiv.2605.06614
  11. [14] doi: 10.48550/arXiv.2403.03186. Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anandkumar. Voyager: An open-ended embodied agent with large language models.arXiv preprint arXiv:2305.16291, on PithPith's reviewarXivdoi:10.48550/arxiv.2403.03186
  12. [15] doi: 10.48550/arXiv.2305.16291. Jiongxiao Wang, Qiaojing Yan, Yawei Wang, Yijun Tian, et al. Reinforcement learning for self-improving agent with skill library.arXiv preprint arXiv:2512.17102, on PithPith's reviewarXivdoi:10.48550/arxiv.2305.16291
  13. [17] doi: 10.48550/arXiv.2409.07429. Tianbao Xie, Danyang Zhang, Jixuan Chen, et al. OSWorld: Benchmarking multimodal agents for open- ended tasks in real computer environments.arXiv preprint arXiv:2404.07972, on PithPith's reviewarXivdoi:10.48550/arxiv.2409.07429
  14. [18] doi: 10.48550/arXiv. 2404.07972. Taofeng Xue, Chong Peng, Mianqiu Huang, et al. EvoCUA: Evolving computer use agents via learning from scalable synthetic experience.arXiv preprint arXiv:2601.15876, on PithPith's reviewarXivdoi:10.48550/arxiv.2404.07972
  15. [19] doi: 10.48550/arXiv.2601.15876. An Yang, Anfeng Li, Baosong Yang, et al. Qwen3 technical report.arXiv preprint arXiv:2505.09388, on Pithdoi:10.48550/arxiv.2601.15876
  16. [20] doi: 10.48550/arXiv.2505.09388. Kangning Zhang, Shuai Shao, Qingyao Li, et al. MMSkills: Towards multimodal skills for general visual agents.arXiv preprint arXiv:2605.13527, on PithPith's reviewarXivdoi:10.48550/arxiv.2505.09388
  17. [21] doi: 10.48550/arXiv.2605.13527. Yunzhu Zhang, Zeyu Pan, Zhengwen Zeng, Shuheng Shen, Changhua Meng, and Linchao Zhu. MVP: Multiple view prediction improves GUI grounding.arXiv preprint arXiv:2512.08529, on PithPith's reviewarXivdoi:10.48550/arxiv.2605.13527
  18. [22] doi: 10.48550/ arXiv.2512.08529. Andrew Zhao, Daniel Huang, Quentin Xu, Matthieu Lin, Yong-Jin Liu, and Gao Huang. ExpeL: LLM agents are experiential learners.arXiv preprint arXiv:2308.10144, on Pithdoi:10.48550/arxiv.2512.08529
  19. [23] doi: 10.48550/arXiv.2308.10144. Shuyan Zhou, Frank F. Xu, Hao Zhu, et al. WebArena: A realistic web environment for building autonomous agents.arXiv preprint arXiv:2307.13854, on PithPith's reviewarXivdoi:10.48550/arxiv.2308.10144
  20. [24] doi: 10.48550/arXiv.2307.13854. 10 on PithPith's reviewarXivdoi:10.48550/arxiv.2307.13854
show 4 more
  1. [2023] doi: 10.48550/arXiv.2303.11366. Weihao Tan, Wentao Zhang, Xinrun Xu, et al. Cradle: Empowering foundation agents towards general computer control.arXiv preprint arXiv:2403.03186, on PithPith's reviewarXivdoi:10.48550/arxiv.2303.11366
  2. [2024] doi: 10.48550/arXiv.2401.10935. Zihao Cheng, Zeming Liu, Yingyu Shan, Xinyi Wang, et al. Mem2Evolve: Towards self-evolving agents via co-evolutionary capability expansion and experience distillation.arXiv preprint arXiv:2604.10923, on PithPith's reviewarXivdoi:10.48550/arxiv.2401.10935
  3. [2025] doi: 10.48550/arXiv.2512.17102. Zora Zhiruo Wang, Jiayuan Mao, Daniel Fried, and Graham Neubig. Agent workflow memory.arXiv preprint arXiv:2409.07429, on PithPith's reviewarXivdoi:10.48550/arxiv.2512.17102
  4. [2026] doi: 10.48550/arXiv.2603. 02766. Xinyu Che, Junqi Xiong, Yunfei Ge, et al. MMG2Skill: Can agents distill in-the-wild guides into self-evolving skills?arXiv preprint arXiv:2606.01993, on Pithdoi:10.48550/arxiv.2603
Cited by 4 Pith papers

Challenges and discussion

Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.

No comments yet.

Sign in to comment