Imported from tiangong-ai/agent-skills (figshare-data-download/SKILL.md). Install upstream with npx skills add tiangong-ai/agent-skills --skill figshare-data-download. Copyright stays with the author.

Figshare Data Download

Use this skill to fetch Figshare dataset files behind anti-bot checks.

Workflow

1. Normalize inputs:

• item_or_file_url: DOI URL, item page URL, or ndownloader URL
• output_path: final local path
• expected_name: expected browser filename (optional)

2. Resolve the file download link in browser:

• Open the DOI/item page with browser tooling.
• Find and click the dataset "Download" / "Download file" element.
• If already given an ndownloader URL, open it directly in browser.

3. Complete any browser challenge/verification and trigger the real file download.

4. Copy the downloaded file from ~/Downloads to output_path.

5. Verify output:

• Ensure file size is non-zero.
• Ensure extension/type matches expectation (.xlsx, .csv, etc.).

Decision Rules

• Do not use web_fetch for this skill.
• Do not run curl probing as part of this skill workflow.
• Always use browser path first and keep it end-to-end.

Output Contract

• Always produce one final file at output_path.
• Keep the downloaded source file in ~/Downloads unchanged.
• Report:
  • final path
  • file size
  • that the transfer used browser-only workflow

Script

• scripts/figshare_data_download.py (browser-only helper)
  • supports semi-automatic flow: --open-browser + wait/poll in ~/Downloads + copy/verify
  • does not perform curl/web_fetch probing