Show HN: What research is your code standing on?

A new Show HN project called Standing on analyzes a pasted GitHub repository's Python code — its imports, calls, loaded models, and papers cited in comments and README — and matches each piece to the research paper behind it using Lacuna, giving each paper a one-line key idea. Two AI models, Claude and Codex, select the papers and the links between them, with each pick checked against Lacuna; solid lines mark links both models agree on and dashed lines mark one model's pick. Results export as BibTeX, CSV or Markdown, and users can dig back from any paper or open the Tree to trace an idea to earlier work, sometimes a century back.