Searches Semantic Scholar and arXiv via the bundled papers.py CLI: metadata, citation/reference graphs, arXiv PDF download, and PyMuPDF text extract. Use when finding papers, analyzing citations, building reading lists, or summarizing arXiv preprints. Not for paywalled publisher PDFs, OCR of scanned pages, citation-alert watches, or general web due diligence (end-to-end-research).
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
Searches Semantic Scholar and arXiv via the bundled papers.py CLI: metadata, citation/reference graphs, arXiv PDF download, and PyMuPDF text extract. Use when finding papers, analyzing citations, building reading lists, or summarizing arXiv preprints. Not for paywalled publisher PDFs, OCR of scanned pages, citation-alert watches, or general web due diligence (end-to-end-research).
Papers Skill turns a coding agent into a literature-research assistant. It orchestrates a bundled Python CLI (scripts/papers.py) that hits the free Semantic Scholar and arXiv APIs, downloads arXiv PDFs, and extracts text with PyMuPDF. The agent decides which subcommand to invoke and how to combine results into a literature scan, a deep read of one paper, an impact analysis, or a reading list.
This skill is the Skill-mode port of the papers-mcp MCP server by the same author. Both projects share the same feature set; this one ships as a plugin so it can be installed with a single command and needs no long-running MCP process.
When to Use
Use when the user asks to search academic papers by topic, author, or venue.
Use when the user names a specific paper (by DOI, arXiv ID, or title) and wants metadata, the abstract, the TL;DR, or its reference list.
Use when the user wants to find work that cites a known paper (impact analysis, follow-up tracking).
Use when the user wants to download an arXiv PDF and have it summarized.
Use when the user asks to build a reading list around a topic.
Do NOT use this skill when:
The user wants paywalled non-arXiv full text. This skill cannot bypass publisher paywalls; it can only fetch arXiv PDFs and metadata everywhere.
The user wants OCR over scanned PDFs. PyMuPDF extracts embedded text only; scanned image-PDFs return the fallback message and need a separate OCR step.
The user wants real-time citation alerts or RSS-style watching. This skill is request-driven.
Prerequisites
Three Python packages are required: httpx, arxiv, and PyMuPDF. Check once per session using the same interpreter for both the import-check and the install so targets stay in sync.
No API keys or credentials are needed. The CLI performs outbound HTTPS only to api.semanticscholar.org and arxiv.org. No authentication tokens are sent.
Procedure
Step 1 — Locate the bundled CLI
The script lives at ${CLAUDE_PLUGIN_ROOT}/skills/papers-skill/scripts/papers.py and is bundled with this skill (no separate install needed). Always quote the path so it survives spaces.
Cluster the citing papers by year/theme and highlight the most-cited follow-ups.
Step 4 — Format output
Always include the paper ID alongside every title so the user can re-query precisely.
Cite as [FirstAuthor et al., Year] *Title* (cites: N).
For PDFs you download, always report the absolute save path.
Confirm the save path with the user before downloading to an unexpected location.
Hard rules
✅ Always call detail before download to confirm the paper matches user intent. Skipping this leads to wrong PDFs being fetched.
❌ Don't crawl. The script auto-retries 429s with exponential backoff (3×); don't pile on parallel queries.
❌ Don't raise --max-pages to 100+ without warning the user — it can consume a large amount of context.
read opens a local PDF file with PyMuPDF — make sure the path the user supplies is one they trust.
download writes a PDF to the directory the user specifies (default: current working directory).
Examples
Example: Build a reading list on "RLHF"
# 1. Search Semantic Scholar for breadth
python "$env:CLAUDE_PLUGIN_ROOT/skills/papers-skill/scripts/papers.py" search "reinforcement learning from human feedback" --limit 10
# 2. Cross-check arXiv for recent preprints
python "$env:CLAUDE_PLUGIN_ROOT/skills/papers-skill/scripts/papers.py" arxiv "RLHF" --max-results 5
# 3. For each top hit, pull detail + citations
python "$env:CLAUDE_PLUGIN_ROOT/skills/papers-skill/scripts/papers.py" detail 2204.05862
python "$env:CLAUDE_PLUGIN_ROOT/skills/papers-skill/scripts/papers.py" citations 2204.05862 --limit 10
Present a curated list grouped by sub-topic with citation counts and one-line summaries.
Pitfalls
Missing optional dependency message. The script returns 需要安装 arxiv: pip install arxiv or 需要安装 PyMuPDF: pip install PyMuPDF instead of crashing when an optional dependency is missing. Offer to run the install command: python -m pip install arxiv PyMuPDF.
Semantic Scholar rate limiting.搜索失败: rate limit, retries exhausted from search, detail, or citations. Semantic Scholar's anonymous tier rate-limits aggressively; the script retries 3× with exponential backoff. Fix: Wait ~10 seconds and retry once. For repeated runs, fall back to the arxiv subcommand for arXiv-indexed work.
Download fails with 找不到 arXiv ID: …. The user gave a non-arXiv ID (likely a DOI for a non-arXiv paper). Use detail to inspect; only papers with an externalIds.ArXiv field can be downloaded.
Scanned PDF returns fallback text.PDF无法提取文本（可能是扫描件） — PyMuPDF extracts embedded text only. Offer the user an alternative version or note that OCR is required (separate tool).
Garbled Chinese output on Windows. The script already forces UTF-8 stdout. If the host terminal is still misconfigured, set the environment variable before running:
$env:PYTHONIOENCODING = "utf-8"
Paywalled full text. The skill cannot fetch full text from paywalled publishers (Elsevier, Springer, Wiley, etc.). It can only read open arXiv PDFs. Metadata and abstracts are available everywhere via Semantic Scholar.
Verification
After running any subcommand, verify the output is valid:
Check 1 — Search returns results:
python "$env:CLAUDE_PLUGIN_ROOT/skills/papers-skill/scripts/papers.py" search "attention is all you need" --limit 3
Expected: JSON or formatted output with at least 1 result containing title, year, citationCount, and paperId fields.
