← ClaudeAtlas

download-paperslisted

Download IEEE Xplore publisher PDFs by document ID through institutional or subscriber access, especially when inherited proxies or browser-only entitlement cookies defeat direct downloads. Not for general literature search.
PRO-2684/skills · ★ 0 · Data & Documents · score 58
Install: claude install-skill PRO-2684/skills
# Download Papers IEEE only, for now. Resolve the helper relative to this `SKILL.md` and inspect its CLI before use: ```bash python3 scripts/download_ieee.py --help ``` Pass one or more numeric IEEE document IDs and an output directory: ```bash python3 scripts/download_ieee.py 9757872 9879360 --output-dir papers/ ``` The helper owns the workflow: it creates a fresh proxy-free `agent-browser` session, confirms and reports `Access provided by`, exports browser cookies into private temporary storage, waits a randomized interval before each PDF request, downloads through `wget`, `curl`, or `aria2c`, validates each PDF, writes `<DOCUMENT_ID>.pdf`, prints JSON metadata to stdout, and removes the browser session and credentials. Use `--downloader` only to select a transport when `auto` is unsuitable. Use `--force` only when replacing existing `<DOCUMENT_ID>.pdf` files is intended. Rename files afterward from the emitted metadata when useful. ## Failure Meaning - IEEE 418: stop; do not add retries or reduce the randomized delays. - Repeated `tag=1` redirects or 502: entitlement cookies were not accepted. - HTML instead of `%PDF-`: access denial or an IEEE error page, not a PDF. Do not bypass a paywall. This carries access already granted to the user's direct institutional, open-access, or personal IEEE session.