paper-analysis

name: paper_analysis description: Research assistant for extracting specific methodology and results from the author's papers. Prioritizes Author Papers (001).

Thesis Paper & Contribution Research

Extracts specific methodology, results, and justifications from the author's own published papers (001) sources). These papers are included verbatim in the thesis LaTeX, so this skill focuses on extracting precise details for cross-referencing — not for rewriting content.

Target Notebook

• Name: thesis references
• ID: 1b7df790-7858-4fc8-879c-39f41238c4ae
• Strict Rule: Exclusively use this notebook.

When to Use This Skill

Use this skill when:

• Describing the author's methodology (e.g., "How did we train the network in Paper 1?")
• Reporting the author's results (e.g., "What was the specific exclusion limit we found?")
• Explaining the author's specific choices (e.g., "Why did we choose this specific ROI?")
• Building transitions between a contextual chapter and the inserted paper

Do NOT use this skill for general theory or definitions — use literature_research instead. Do NOT use this skill for review article content — use review_analysis instead.

Key Principle: Internal References, Not External Citations

The author's papers are included as-is in the thesis LaTeX. Therefore:

• Do NOT rewrite or summarize what the paper already says
• DO extract specific facts, numbers, and methods to reference from contextual chapters
• DO use internal references: "See Chapter 6, Sec. 6.2" or "as detailed in the following paper"
• DO identify the key contributions to highlight in the introductory paragraphs before each paper

Author Papers Mapping

Prerequisites

1. Run source_registry first to obtain 001) source IDs.

MCP Tool Usage

Querying Author Papers

mcp_notebooklm_notebook_query(
    notebook_id="1b7df790-7858-4fc8-879c-39f41238c4ae",
    query="<your question>",
    source_ids=<author_001_ids>,  # from source_registry, only 001) sources
    conversation_id=<previous_id>  # for follow-up drilling
)

Targeting a Specific Paper

When you need details from a single paper, pass only that paper's source ID:

mcp_notebooklm_notebook_query(
    notebook_id="1b7df790-7858-4fc8-879c-39f41238c4ae",
    query="Detail the domain adaptation methodology used in this paper.",
    source_ids=["<paper_4_source_id>"]
)

Query Strategy

1. Methodology Extraction: Ask for step-by-step methods.

   • Prompt: "Based on '001) paper 4', detail the domain adaptation methodology used to handle dataset shift. List the specific steps and techniques."

2. Results Extraction: Ask for specific numbers and figures.

   • Prompt: "Retrieve the forecasted sensitivity to Dark Matter from '001) paper 5'. What are the exact numerical limits?"

3. Choice Justification: Ask why specific decisions were made.

   • Prompt: "Summarize the motivation for targeting unassociated sources as described in the author's papers."

4. Key Contributions: Extract the main takeaway for chapter introductions.

   • Prompt: "What is the single most important contribution of '001) paper 1'? Summarize in 2-3 sentences suitable for an introductory paragraph."

5. Comparison with Literature: Compare author results against the field.

   • Prompt: "How do the results in '001) paper 3' compare to previous constraints on the MSP luminosity function?"

Output

Internal Reference Notes

Structure output as notes for writing contextual chapters:

## Paper [N] — Key Extractions

### Methodology Summary
- [Step 1]: [brief description] → detailed in Sec X.X of the paper
- [Step 2]: [brief description]

### Key Results
- [Result 1]: [exact value/limit]
- [Result 2]: [comparison with prior work]

### Contribution Statement (for chapter intro)
"[2-3 sentence summary of what this paper achieves and why it matters]"

### Transition Points
- **From context chapter to paper**: [how the introductory chapter leads into this paper]
- **From paper to next chapter**: [how this paper's conclusions set up the next chapter]

Knowledge Saving

REQUIRED: Use the knowledge skill (save mode) to persist methodology insights to .agent/knowledge/. The knowledge skill defines the standard file format and handles deduplication.

Usage Examples

Scenario: Describing Methodology

• User: "Explain how we handled the dataset shift in Paper 4."
• Action: Query 001) paper 4 specifically.
• Query: "Based on '001) paper 4', detail the domain adaptation methodology used to handle dataset shift."

Scenario: Reporting Results

• User: "What was the final sensitivity we achieved for the CTA analysis?"
• Action: Query 001) paper 5.
• Query: "Retrieve the forecasted sensitivity to Dark Matter from '001) paper 5'."

Scenario: Building a Chapter Introduction

• User: "Write the introduction for Chapter 6, before Paper 1 is inserted."
• Action: Query 001) paper 1 for key contributions.
• Query: "What is the key contribution and motivation of '001) paper 1'? What existing gap does it fill?"