Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →brycewang-stanford/paper-parse-guide
Deep dual-mode reading of academic papers from PDF or URL sources
Passed 15 of 21 test cases with the skill, and 10 of 21 without it.
and 6 more in the full report
7 fixed9 passed either way5 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 6 test cases the skill fixed. Same model, same prompt, with and without the skill.
I need to run an open-source TEI XML extractor for high-fidelity extraction of two-column research paper structures via Docker. What service command and port setup should be executed locally?
| Checked | Without skill | With skill |
|---|---|---|
| Uses port 8070 in the Docker run command | Passed | Passed |
| Uses image lfoppiano/grobid:0.8.0 | Missed | Passed |
The instructions your model reads when the skill loads.
Perform structured, dual-mode deep reading of academic papers from PDF files or URLs. Mode A provides a rapid overview suitable for screening during literature reviews. Mode B delivers exhaustive section-by-section analysis for papers central to your research.
Reading academic papers efficiently is a core research skill, yet the density and conventions of scholarly writing make it time-consuming. A typical researcher reads dozens of papers per week during a literature review phase, requiring different levels of depth for different papers. Some need only a quick scan to determine relevance; others demand line-by-line scrutiny of methods and results.
This skill implements a dual-mode reading system. Mode A (Survey Mode) extracts key metadata, the main argument, methods summary, and key findings in under two minutes of processing time. Mode B (Deep Analysis Mode) performs exhaustive section-by-section analysis including methodology critique, statistical evaluation, figure interpretation, and connection to broader literature.
Both modes begin by parsing the paper's structure from its PDF or HTML source, extracting clean text with section boundaries, figures, tables, equations, and references. The parsing pipeline handles the common challenges of academic PDFs: two-column layouts, footnotes, headers/footers, embedded equations, and supplementary materials.
| Source | Method | Notes |
|---|---|---|
| Local PDF | Direct file path | Best quality, no network needed |
| DOI | Resolve via CrossRef/Unpaywall | Auto-fetches open access version |
| arXiv ID | https://arxiv.org/pdf/{id} | Always available |
| URL | Direct download | May require institutional access |
| OpenAlex ID | OpenAlex API + OA link | Includes metadata |
from pathlib import Path
import fitz # PyMuPDF
def parse_paper(pdf_path: str) -> dict:
doc = fitz.open(pdf_path)
sections = []
current_section = {"title": "Header", "content": []}
for page in doc:
blocks = page.get_text("dict")["blocks"]
for block in blocks:
if block["type"] == 0: # Text block
for line in block["lines"]:
test cases passed, vs 10 / 21 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.