Paper Assistant

Paper Assistant

The central orchestrator for AI-assisted academic paper writing, managing the complete workflow from literature review to publication.

Workflow Stages

1. Literature Review → Research gap identification
2. Hypothesis Generation → Hypothesis formulation and experiment design
3. Code Implementation → Experimental code development
4. Experiment Execution → Running experiments and analyzing results
5. Paper Writing → Drafting and polishing the manuscript

Context Management

The assistant maintains context through .paper_context.json:

{
  "project_name": "my-paper",
  "current_stage": "literature_review",
  "stages": {
    "literature_review": {
      "status": "completed",
      "research_gap": "...",
      "summary": "..."
    },
    "hypothesis": { "status": "pending" },
    "code": { "status": "pending" },
    "experiment": { "status": "pending" },
    "writing": { "status": "pending" }
  },
  "user_preferences": {
    "discipline": "cs",
    "methodology": "quantitative",
    "language": "zh",
    "citation_style": "IEEE"
  }
}

Usage

Start New Project

I want to write a paper about [topic]

Continue from Stage

Continue to hypothesis generation
Skip to experiment stage

Query Status

What's the current status of my paper?
Show me the literature review summary

Stage Validation

Each stage must meet completion criteria before proceeding:

• Literature Review: Research gap identified, key papers summarized
• Hypothesis: Feasibility assessed, experiment design approved
• Code: Implementation complete, tests passing
• Experiment: Results collected, statistical significance verified
• Writing: All sections drafted, citations formatted

Integration

This coordinator delegates to specialized skills:

• paper-literature-review: Literature search and analysis
• paper-hypothesis: Hypothesis generation and validation
• paper-code: Code scaffolding and best practices
• paper-experiment: Experiment execution and analysis
• paper-writing: Manuscript drafting and polishing

Each plugin can be used independently or as part of the full workflow.

Scripts

The coordinator provides three core scripts for context management:

• scripts/context-manager.py: Load, save, and update .paper_context.json
• scripts/state-checker.py: Validate stage completion and transitions
• scripts/discipline-config.py: Manage discipline-specific configurations

Usage Examples

# Initialize new project
python scripts/context-manager.py init --project ./my-paper

# Check project status
python scripts/context-manager.py status

# Validate current stage
python scripts/state-checker.py literature_review

# List supported disciplines
python scripts/discipline-config.py list

# Get discipline info
python scripts/discipline-config.py info --discipline cs