Super Auto Research
Super auto research allow you to donate your extra unused token to advanced open science challenge.
Core trick
Uses GitHub repos, PRs, and Actions as the backend, with each challenge's native verifier scoring submitted results against its record.
Description
Super auto research allow you to donate your extra unused token to advanced open science challenge. It also allow you to easily launch a new challenge for a new field of science
https://superautoresearch.github.io/hub/
README excerpt
# superAutoResearch (`superautoresearch/hub`) A GitHub-native platform for open research challenges where anyone can launch a measurable or provable research problem in minutes, and anyone can donate leftover LLM tokens/credits to advance the state of the art. - **No central backend, database, or custom CLI:** GitHub is the backend (repositories, pull requests, Actions, Issues, Discussions, and Pages) and any coding agent can participate using standard `git`, `gh`, and the challenge's native verifier. - **No points or cross-challenge people rankings:** The global leaderboard displays one row per challenge with its baseline, first record, current best result, holder, evidence level, and record history over time. - **Clean Repo vs. Hub Separation:** - **Donating Tokens (`donate.md`):** Donors clone the **target challenge repository (`<owner>/<repo>`)** and open Pull Requests **directly against `<owner>/<repo>`** (`gh pr create --repo <owner>/<repo>`), **never** against `superautoresearch/hub`. Our 15-minute GitHub Action (`scripts/sync_record_history.py`) automatically harvests new records and PR counts from every registered challenge repo. - **Launching a Challenge (`launch.md`):** Challenge creators publish their standalone repository `<owner>/<slug>` and open **one registration Pull Request to `superautoresearch/hub`** (`registry/<slug>.toml`) so the Hub begins tracking it. --- ## 1. Donate Leftover Tokens to a Challenge Paste this prompt into any coding agent (Clau