Collect arXiv papers, create per-paper summaries, and render a hierarchical batch report.
Documents
ArXiv Paper Processor
Downloads ArXiv artifacts and produces a full-text-based summary.md for each paper.
What it does
Downloads ArXiv source or PDF artifacts for one paper or a batch, then guides model-side full-text reading and synthesis. It writes one summary.md per paper in the selected language and fixed section format, with traceable paper-specific method, evaluation, results, and uncertainty details. Existing artifacts and completed summaries can be reused or skipped.
When to use it
- Summarizing one ArXiv paper from full text
- Predownloading artifacts for a paper batch
- Resuming an interrupted paper-processing run
- Producing Chinese or English per-paper summaries
The skill document
ArXiv Paper Processor
Use this skill for per-paper manual summarization, with optional batch artifact download.
- Single-paper mode: process one paper directory (e.g.
//). - Batch predownload mode: process many paper directories under one run dir before writing summaries.
Language Parameter
- Use a workflow language parameter (for example
EnglishorChinese) and apply it manually. - The per-paper
summary.mdmust be written in the selected language. - If download scripts are called directly, pass
--languagefor traceability.
Core Principle
Scripts only fetch artifacts. The model performs reading and writing.
Non-negotiable Constraint
- Do not generate
summary.mdby script-based snippet extraction, regex harvesting, or template autofill. - Do not use Python/shell scripts to auto-compose section text from abstract/introduction fragments.
- Scripts in this skill are only for artifact download (
source/pdf) and trace logs. - The final
summary.mdmust come from model-side reading and synthesis of the paper content.
Optional Batch Artifact Download (Many Papers)
Use this first when Stage B has many papers:
python3 scripts/download_papers_batch.py \
--run-dir /path/to/run \
--artifact source_then_pdf \
--max-workers 3 \
--min-interval-sec 5 \
--language English
Key behavior:
- Supports
--artifact source,--artifact pdf, or--artifact source_then_pdf(default). - Supports concurrency (
--max-workers) and safe throttling/retry (--min-interval-sec, retry args). - Uses run-local throttle state by default (
/.runtime/arxiv_download_state.json) to reduce 429 risk. - Skips papers that already have usable
source/source_extract/*.texor existingsource/paper.pdf(unless--force). - Resume-friendly: if a paper already has a completed
summary.md, you can skip that paper's summary-writing step. - Writes batch log to
/download_batch_log.jsonby default.
Step 1: Download Source (Preferred)
python3 scripts/download_arxiv_source.py \
--paper-dir /path/to/run/2602.00528 \
--language English
This writes:
source/source_bundle.binsource/source_extract/source/download_source_log.json
If usable source already exists and --force is not set, the script reuses local artifacts.
Step 2: If Needed, Download PDF
python3 scripts/download_arxiv_pdf.py \
--paper-dir /path/to/run/2602.00528 \
--language English
This writes:
source/paper.pdfsource/download_pdf_log.json
If PDF already exists and --force is not set, the script reuses local artifacts.
Step 3: Model Reads and Summarizes
- If
summary.mdalready exists and follows the required format, skip this paper and mark it complete. - Read
metadata.mdfirst. - If
source/source_extract/already exists with readable.texfiles, use it directly. - Otherwise, if
source/paper.pdfalready exists, use PDF directly. - If neither exists, run download scripts (single-paper scripts or batch script) first.
- Manually write
summary.mdin the same paper directory, in the selected language.
Do not rely on rule-based auto summarization. Do not rely on auto-extracted snippets as the primary writing basis.
Quality Requirement
- Every section should include paper-specific details that are traceable to full-text reading.
- Section 4/5/10 should reflect concrete method and evaluation details, not generic wording.
- If key details are unclear in the source, explicitly note uncertainty instead of guessing.
- Match the detail level shown in
references/summary-example-en.mdandreferences/summary-example-zh.md. - If your draft is clearly shorter or less specific than the examples, expand it before finishing.
Required Output
/summary.mdin fixed section format.- Pay special attention to section
## 10. Brief Conclusion: write a 3-4 sentence mini-conclusion that covers contribution, method, evaluation setup, and results with paper-specific details. - In section
## 1. Paper Snapshot, use exact keys:ArXiv ID,Title,Authors,Publish date,Primary category,Reading basis. - Do not use key variants such as
Reading source,Author list,Published on, or lowercase key names.
See references/summary-format.md for exact section requirements.
Related Skills
This skill is a sub-skill of arxiv-summarizer-orchestrator.
Pipeline position:
- Step 1 (upstream):
arxiv-search-collectorproduces the selected paper directories and metadata. - Step 2 (this skill):
arxiv-paper-processordownloads artifacts and writes onesummary.mdper paper. - Step 3 (downstream):
arxiv-batch-reporteruses these per-paper summaries to generate the final collection report.
Use this skill together with Step 1 and Step 3 for full end-to-end execution.
Questions people ask
- Does it generate summaries automatically from abstracts or extracted snippets?
- No. Scripts are limited to artifact downloads and trace logs; summary.md must be written through model-side reading and synthesis of the paper content.
- Can it process multiple ArXiv papers in one run?
- Yes. Batch predownload supports source, PDF, or source-then-PDF artifacts with configurable concurrency, throttling, retries, and a JSON batch log. Summary writing still happens per paper.
- What files and language options does it support?
- It prefers readable extracted TeX source and falls back to an existing or downloaded PDF. The selected workflow language, such as English or Chinese, is used for each paper's fixed-format summary.md.
Related skills
Compress documents, transcripts, threads, data, and code changes while preserving claims, caveats, and actions.
Convert YouTube URLs into inline chaptered summaries, timestamped transcripts, or transcript JSON.
Turn prompts and source details into designed PDF documents or editable DOCX files.
Draft legal documents, policies, memos, and compliance checklists from detailed natural-language requirements.
Research recent discussions across Reddit, X, YouTube, and the web, then synthesize cited findings.