Best use case
review-writer is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
将论文笔记和对比矩阵综合为结构化学术文献综述,含 BibTeX 引用
Teams using review-writer should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/review-writer/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How review-writer Compares
| Feature / Agent | review-writer | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
将论文笔记和对比矩阵综合为结构化学术文献综述,含 BibTeX 引用
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Review Writer Skill
将分析结果综合为高质量文献综述。仅由 review-lead subagent 或 Mac_Javis 主 agent 使用。
> ⚠️ 写综述前,完整阅读本文件。遵循模板结构和质量清单。
## Runtime Router(必读)
识别当前 runtime,只读取对应 sibling,另一方休眠:
- `openclaw` → 本文件内原有指令块仍然有效(`web_fetch:` / `exec:` / `openclaw browser`)
- `claude-code` → **跳过本文件的指令块**,读 `./claude-code.md` 获取 Claude Code 原生工具调用方式
- `codex` / `cli` → **跳过本文件的指令块**,读 `./codex.md` 获取 Codex 原生工具调用方式
本节之后的章节描述 **共享知识**(源、字段契约、评分规则、故障处理)。指令块保持现状(OpenClaw 语法),Claude Code 读者请切换到 `./claude-code.md`,Codex/CLI 读者请切换到 `./codex.md`。
## 输入文件
开始写作前,必须读取所有这些文件:
```
exec: ls ~/research/[PROJECT]/notes/
read: ~/research/[PROJECT]/matrix.csv
read: ~/research/[PROJECT]/candidates.csv
read: ~/research/[PROJECT]/search_log.md
```
然后逐个读取 `notes/` 目录下的每个笔记文件。
## 综述结构模板
写入 `~/research/[PROJECT]/review.md`:
```markdown
# Literature Review: [Topic]
**Generated**: [YYYY-MM-DD]
**Papers Analyzed**: [N]
**Sources**: arXiv, Semantic Scholar, OpenAlex, ...
**Search Queries**: [list main queries used]
## 1. Introduction
[研究背景和动机。为什么这个主题重要?
本综述的范围是什么?2-3 段。]
## 2. Taxonomy
[提出一个分类体系来组织综述文献。]
| Category | Representative Papers | Core Idea |
|----------|----------------------|-----------|
| [Cat A] | [cite1, cite2] | [1-sentence] |
| [Cat B] | [cite3, cite4] | [1-sentence] |
## 3. Detailed Analysis
### 3.1 [Category A]
[该类别共有的方法特征
跨论文的方法对比(引用 matrix.csv 数据)
优缺点分析
时间线上的演进趋势]
### 3.2 [Category B]
[同上结构]
## 4. Cross-Cutting Analysis
### 4.1 Common Datasets and Benchmarks
[哪些数据集出现最频繁?是否有标准化 benchmark?]
### 4.2 Methodological Trends
[跨类别的趋势]
### 4.3 Open Challenges
[整个领域共同面对的开放问题]
## 5. Research Gaps and Future Directions
[基于分析识别:
- 覆盖不足的子主题
- 方法论空白
- 未被探索的组合
每个空白引用揭示它的论文。]
## 6. Conclusion
[2-3 段总结。]
## References
[见 references.bib]
```
## BibTeX 生成
写入 `~/research/[PROJECT]/references.bib`:
- 从所有笔记文件收集 BibTeX 条目
- 按 citekey 字母排序
- 去重(按 paper_id)
- 确保 review.md 中每个 \cite{} 在 bib 中都有对应条目
## 质量自检清单
完成综述后,逐项验证:
- [ ] 每个事实主张都引用了至少一篇论文
- [ ] notes/ 下的每篇论文都在综述中被提及或有排除理由
- [ ] 分类框架类别之间互斥且完全覆盖
- [ ] 研究空白基于证据而非猜测
- [ ] references.bib 条目与正文引用一一对应
- [ ] 无编造信息
- [ ] review.md 独立可读
## 写作规范
- 学术语气,第三人称
- 一般性论述用现在时("X demonstrates that...")
- 具体实验结果用过去时("Y achieved 95% accuracy on...")
- 精确表述:"improves by 3.2%" 而非 "significantly improves"
- 公平对待每篇论文的局限性
- 默认英文写作,除非人类明确要求中文Related Skills
verifier
Independent verification of literature review quality — citation checks, claim tracing, coverage analysis
trendr-watchdog
运行时监督器。监控 run_status/progress/log 活跃度,发现卡住后自动向 owner session 注入断点续跑指令。
research-vault
将研究成果持久化到 Obsidian vault,维护论文池索引。支持每日研究日志、论文卡片、综述归档,以及跨项目论文去重和快速检索。
platform-hotspots
Collect and summarize Zhihu, Xiaohongshu, X, Reddit, YouTube, GitHub Trending, Hacker News, and Product Hunt hotspots with strict Chrome CDP routing and reproducible extraction commands.
paper-scout
9 源学术论文搜索与筛选(arXiv, Semantic Scholar, OpenAlex, PubMed, CrossRef, DBLP, Europe PMC, bioRxiv, Papers with Code),利用本机已安装工具,零额外依赖
paper-analyzer
从学术论文中提取结构化信息,生成标准化笔记和文献对比矩阵
chrome-cdp-setup
Chrome 146+ CDP remote debugging architecture — dual-instance setup, cookie sync, profile isolation, and troubleshooting "Allow remote debugging" popups.
caveman-review
Ultra-compressed code review comments. Cuts noise from PR feedback while preserving the actionable signal. Each comment is one line: location, problem, fix. Use when user says "review this PR", "code review", "review the diff", "/review", or invokes /caveman-review. Auto-triggers when reviewing pull requests.
review-skill
Review and improve AI agent skills (SKILL.md files) against best practices from the Agent Skills specification and Anthropic's authoring guidelines. Scores skills on 10 quality dimensions, identifies specific issues, and rewrites problem areas. Use when creating, editing, auditing, or improving agent skills. Triggers on: review skill, improve skill, audit SKILL.md, check skill quality, skill best practices, optimize skill description.
script_writer
专为“AI芝士派”定制的口播脚本创作助手。基于资深科技评论员视角,生成专业、犀利且接地气的科技/AI类短视频逐字稿。
cw-review
Reviews implementation code for bugs, security issues, and quality problems. Creates FIX tasks for issues found. This skill should be used after cw-validate to catch issues before merge.
cw-review-team
Team-based concern-partitioned code review. Each reviewer sees ALL files through a specialized lens (security, correctness, spec compliance). This skill should be used after cw-validate for thorough cross-file review (requires CLAUDE_CODE_TASK_LIST_ID).