完整覆盖为 3 个外部 PDF 图 + 1 个文本框
figure；未发现弃用图。原始名称与纳入说明见工作笔记。
Other Discussion Roles / 其他讨论角色
七个 MIT MAS.S60 角色逐字 prompt 与基于公开证据的中文回应。
Scientific Peer Reviewer
The paper has not been published yet and is currently submitted to a
top conference where you’ve been assigned as a peer reviewer.
Complete a full review of the paper answering all prompts of the
official review form of the top venue in this research area (e.g.,
NeurIPS). This includes recommending whether to accept or reject the
paper.
This paper was found buried under ground in the desert. You’re an
archeologist who must determine where this paper sits in the context
of previous and subsequent work. Find and report on one older paper
cited within the current paper that substantially influenced the
current paper and one newer paper that cites this current paper.
较早且有实质影响：ReAct。论文 references 引用 Yao
et al. (2022),
ReAct: Synergizing Reasoning and Acting in Language Models（arXiv:2210.03629）。该工作让 LLM 以交错 reasoning/action 轨迹与环境交互，并已覆盖
ALFWorld 等决策任务；SkillOpt 继承“从 agent
实际轨迹获取反馈”的思想，进一步把执行规程外化为可优化
skill，目标不是只生成一次动作序列。其摘要已通过 arXiv
公开页检索核验。
较新引用本文的工作：截至
2026-10-23，本次可直接核实的公开证据中，未找到一篇可确认发表时间晚于本文且明确引用
arXiv:2605.23904 的后续论文。本文 v2 于 2026-05-25 上传；arXiv
官方页能核对版本与摘要，但未提供可用的完整引用后继清单。Semantic
Scholar API 曾返回 429，export.arxiv.org DNS
也不可用；故不虚构“后继论文”。此处结论是“当前检索边界内未核实”，并非断言不存在引用者。
You’re a researcher who is working on a new project in this area.
Propose an imaginary follow-up project not just based on the current
but only possible due to the existence and success of the current
paper.
You work at a company or organization developing an application or
product of your choice (that has not already been suggested in a
prior session). Bring a convincing pitch for why you should be paid
to implement the method in the paper, and discuss at least one
positive and negative impact of this application.
You’re a hacker who needs a demo of this paper ASAP. Implement a
small part or simplified version of the paper on a small dataset or
toy problem. Prepare to share the core code of the algorithm to the
class and demo your implementation. Do not simply download and run
an existing implementation – though you are welcome to use (and give
credit to) an existing implementation for “backbone” code.
Toy demo（已执行，极简教学实现，非论文复现）：使用合成字段题，初始 skill 读取 DRAFT 字段，任务答案由 FINAL
字段决定。自写的 rule-based proposer 从训练失败提出一个
patch（优先读取 FINAL）；在单条固定 selection
题上严格提升才接受。执行输出为 test accuracy 0%→100%（1
条合成测试题），无拒绝编辑。这个微型数字只证明代码路径可运行，不具统计意义。
def optimize(skill, train, selection, propose, score):
current = score(skill, selection)
best = skill
rejected = []
for batch in batches(train):
edits = propose(skill, batch, rejected)
candidate = apply_bounded_patch(skill, edits, budget=1)
candidate_score = score(candidate, selection)
if candidate_score > current:
skill, current = candidate, candidate_score
best = skill
else:
rejected.append((edits, candidate_score - current))
return best, rejected
You are a detective who needs to run a background check on one of
the paper’s authors. Where have they worked? What did they study?
What previous projects might have led to working on this one? What
motivated them to work on this project? Feel free to contact the
authors, but remember to be courteous, polite, and on-topic.
本文作者共 14 位；arXiv 官方页列出 Yifan Yang 等人及投稿时间。仅凭
arXiv author list
不足以可靠地归属特定个人的雇主、学历、历史项目或个人动机。本次不将同名搜索结果拼接为作者履历，也不推测个人动机；没有足够可核验的公开来源时应留白，而不是用论文主题反推生平。
Identify how this paper self-assesses its (likely positive) impact
on the world. Have any additional positive social impacts left out?
What are possible negative social impacts that were overlooked or
omitted?
