paper-details

<quick_start> To get paper details:

/asta:paper-details arxiv:1706.03762

With optional expansions:

/asta:paper-details "10.1038/nature14539" --include citations,references

</quick_start>

<success_criteria>

• •Successfully resolves paper ID to full metadata
• •Returns abstract and TLDR (AI-generated summary)
• •Provides citation count and key metrics
• •Includes links to full text when available
• •Optionally expands to show citations, references, or author's other work </success_criteria>

Available Fields: abstract, authors, citations, citationCount, fieldsOfStudy, influentialCitationCount, isOpenAccess, journal, publicationDate, references, tldr, url, venue, year </context>

1. •

   Parse the input to determine ID type:

   • •DOI pattern: 10.xxxx/xxxxx
   • •ArXiv pattern: arxiv:XXXX.XXXXX or XXXX.XXXXX
   • •Semantic Scholar: CorpusId:XXXXXXX
   • •Title: Quoted string without ID pattern
2. •

   If title provided, use mcp__asta__snippet_search to find the paper first

Phase 2: Retrieve Details

1. •

   Use mcp__asta__get_paper with:

   • •paper_id: The resolved paper ID
   • •fields: Request all relevant fields
2. •

   Parse and format the response

Phase 3: Expand (if requested)

1. •If --include citations: Use mcp__asta__get_paper_citations
2. •If --include references: Use mcp__asta__get_paper_references
3. •If --include authors: Use mcp__asta__get_author_papers for each author </workflow>

<output_format> Structure the output as:

# {Paper Title}

## Quick Summary
**TLDR:** {AI-generated summary from Asta}

## Metadata

| Field | Value |
|-------|-------|
| Authors | {Full author list with affiliations} |
| Year | {Year} |
| Venue | {Conference or Journal} |
| Citations | {count} ({influential_count} influential) |
| Open Access | {Yes/No} |
| Fields | {fieldsOfStudy} |

## Abstract
{Full abstract text}

## Links
- **DOI:** [{doi}](https://doi.org/{doi})
- **ArXiv:** [{arxiv_id}](https://arxiv.org/abs/{arxiv_id})
- **PDF:** {pdf_link if available}
- **Semantic Scholar:** [View on S2]({s2_url})

## Key Citations (if requested)
{Top 5 most influential papers that cite this one}

| Paper | Year | Citations |
|-------|------|-----------|
| {Title} | {Year} | {count} |

## Key References (if requested)
{Most cited references from this paper}

## Related Work by Authors (if requested)
{Other notable papers by the same authors}

</output_format>

Quick Summary

TLDR: A new neural network architecture based solely on attention mechanisms, dispensing with recurrence and convolutions entirely, achieving state-of-the-art results on machine translation.

Metadata

Abstract

The dominant sequence transduction models are based on complex recurrent or convolutional neural networks that include an encoder and a decoder. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely...

Links

• •ArXiv: 1706.03762
• •PDF: Download
• •Semantic Scholar: View on S2 </output>

<anti_patterns>

• •Don't guess IDs - If ID format is unclear, ask for clarification
• •Don't fabricate metadata - Only return data from the API
• •Don't skip the TLDR - Include the AI-generated summary when available
• •Don't return partial data - If a field is unavailable, note it explicitly </anti_patterns>

<common_papers> For testing or quick reference:

• •Attention paper: arxiv:1706.03762
• •BERT: arxiv:1810.04805
• •GPT-3: arxiv:2005.14165
• •AlphaFold: DOI:10.1038/s41586-021-03819-2
• •ResNet: arxiv:1512.03385 </common_papers>