paper-assembly

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Paper Assembly

论文整合编排

Orchestrate the entire paper pipeline end-to-end with state management and checkpointing.

通过状态管理和检查点功能，端到端编排整个论文工作流。

Input

输入

```
$0
```
— Paper project directory or paper plan

```
$0
```
— 论文项目目录或论文规划文件

References

参考资料

Orchestration patterns and state management:

~/.claude/skills/paper-assembly/references/orchestration-patterns.md

编排模式与状态管理：

~/.claude/skills/paper-assembly/references/orchestration-patterns.md

Scripts

脚本

Check pipeline completeness

检查工作流完整性

bash

python ~/.claude/skills/paper-assembly/scripts/assembly_checker.py --dir paper/ --output checkpoint.json
python ~/.claude/skills/paper-assembly/scripts/assembly_checker.py --dir paper/ --verbose

Scans paper directory, checks 9 pipeline phases, reports missing artifacts, suggests next steps.

bash

python ~/.claude/skills/paper-assembly/scripts/assembly_checker.py --dir paper/ --output checkpoint.json
python ~/.claude/skills/paper-assembly/scripts/assembly_checker.py --dir paper/ --verbose

扫描论文目录，检查9个工作流阶段，报告缺失的成果物，并建议后续步骤。

Workflow

工作流程

Step 1: Assess Current State

步骤1：评估当前状态

Scan the paper directory for existing artifacts
Identify which phases are complete vs pending
Build a dependency graph of remaining work

扫描论文目录，识别已有的成果物
确定哪些阶段已完成，哪些待处理
构建剩余工作的依赖关系图

Step 2: Execute Pipeline Phases

步骤2：执行工作流阶段

Run phases in dependency order:

Phase	Skill	Input	Output
1. Literature	literature-search, literature-review	Topic	Knowledge base, BibTeX
2. Planning	research-planning	Knowledge base	Paper structure, task list
3. Code	experiment-code	Plan	Training/eval pipeline
4. Experiments	experiment-design	Code	Results JSON/CSV
5. Figures	figure-generation	Results	PNG figures
6. Tables	table-generation	Results	LaTeX tables
7. Writing	paper-writing-section	All above	main.tex sections
8. Citations	citation-management	Draft	references.bib
9. Formatting	latex-formatting	Draft	Formatted LaTeX
10. Compilation	paper-compilation	All	PDF
11. Review	self-review	PDF	Review scores

按照依赖顺序执行各阶段：

阶段	技能	输入	输出
1. 文献调研	文献检索、文献综述	研究主题	知识库、BibTeX文件
2. 规划	研究规划	知识库	论文结构、任务列表
3. 代码编写	实验代码开发	规划文件	训练/评估工作流
4. 实验	实验设计	代码	结果JSON/CSV文件
5. 图表生成	图表制作	实验结果	PNG格式图表
6. 表格制作	表格生成	实验结果	LaTeX格式表格
7. 写作	论文章节写作	以上所有成果	main.tex章节文件
8. 引用管理	引用管理	论文草稿	references.bib文件
9. 格式调整	LaTeX格式优化	论文草稿	格式化后的LaTeX文件
10. 编译	论文编译	所有成果	PDF文件
11. 评审	自评审	PDF文件	评审分数

Step 3: State Propagation

步骤3：状态传递

After each phase completes:

Save output artifacts to the paper directory
Propagate results to downstream phases
Update the progress checkpoint file

每个阶段完成后：

将输出成果物保存到论文目录
将结果传递给下游阶段
更新进度检查点文件

Step 4: Quality Gates

步骤4：质量校验关卡

Before proceeding to the next phase:

Verify all required outputs exist
Check for consistency (e.g., all cited keys in .bib)
Validate figures/tables match experimental results

进入下一阶段前：

验证所有必需的输出是否存在
检查一致性（例如，所有引用关键词都存在于.bib文件中）
验证图表与实验结果是否匹配

Step 5: Final Assembly

步骤5：最终整合

Merge all sections into main.tex
Verify all \includegraphics files exist
Verify all \cite keys exist in .bib
Compile to PDF
Run self-review for quality check

将所有章节合并到main.tex中
验证所有\includegraphics引用的文件是否存在
验证所有\cite引用的关键词是否存在于.bib文件中
编译生成PDF
执行自评审进行质量检查

Orchestration Patterns

编排模式

Sequential Pipeline (AI-Scientist)

顺序工作流（AI-Scientist）

generate_ideas → experiments → writeup → review

generate_ideas → experiments → writeup → review

Multi-Agent State Broadcasting (AgentLaboratory)

多Agent状态广播（AgentLaboratory）

python

undefined

python

undefined

Propagate results to all downstream agents

set_agent_attr("dataset_code", code) set_agent_attr("results", results_json)

undefined

set_agent_attr("dataset_code", code) set_agent_attr("results", results_json)

undefined

Copilot Mode (AgentLaboratory)

副驾驶模式（AgentLaboratory）

Human can intervene at any phase boundary for review/correction.

人类可在任意阶段节点介入进行评审和修正。

Checkpoint Format

检查点格式

json

{
  "project": "paper-name",
  "phases_completed": ["literature", "planning", "code"],
  "current_phase": "experiments",
  "artifacts": {
    "literature": "knowledge_base.json",
    "plan": "research_plan.json",
    "code": "experiments/",
    "results": null
  },
  "last_updated": "2024-01-15T10:30:00Z"
}

json

{
  "project": "paper-name",
  "phases_completed": ["literature", "planning", "code"],
  "current_phase": "experiments",
  "artifacts": {
    "literature": "knowledge_base.json",
    "plan": "research_plan.json",
    "code": "experiments/",
    "results": null
  },
  "last_updated": "2024-01-15T10:30:00Z"
}

Rules

规则

Never skip phases — each depends on previous outputs
Save checkpoints after every phase completion
Human review is recommended at phase boundaries
All numbers in the paper must trace to actual experiment logs
Re-run downstream phases if upstream changes

不得跳过任何阶段——每个阶段都依赖前序阶段的输出
每个阶段完成后保存检查点
建议在阶段节点进行人工评审
论文中的所有数据必须能追溯到实际的实验日志
若上游阶段有变更，需重新运行下游阶段

paper-assembly

Original

Translation

Paper Assembly

论文整合编排

Input

输入

References

参考资料

Scripts

脚本

Check pipeline completeness

检查工作流完整性

Workflow

工作流程

Step 1: Assess Current State

步骤1：评估当前状态

Step 2: Execute Pipeline Phases

步骤2：执行工作流阶段

Step 3: State Propagation

步骤3：状态传递

Step 4: Quality Gates

步骤4：质量校验关卡

Step 5: Final Assembly

步骤5：最终整合

Orchestration Patterns

编排模式

Sequential Pipeline (AI-Scientist)

顺序工作流（AI-Scientist）

Multi-Agent State Broadcasting (AgentLaboratory)

多Agent状态广播（AgentLaboratory）

Propagate results to all downstream agents

Propagate results to all downstream agents

Copilot Mode (AgentLaboratory)

副驾驶模式（AgentLaboratory）

Checkpoint Format

检查点格式

Rules

规则

Related Skills

相关技能