No public code

of the sector,
with nothing to read.

No repository with CoinGecko, none with DexScreener, and — for everything above $10M — none on the project's own website either, with any plausibly matching GitHub owner checked for a link back. This list is what survived that audit, not what the scan first guessed.

What this list is not. A private repository is a legitimate choice, and this site cannot tell one from an empty one. Nor does it follow GitLab, Codeberg or self-hosted forges. Rows marked checked by hand had their own website read for a repository link, and any plausibly matching GitHub owner checked for a link back to that website — both directions, or it did not count. The verdicts and every rejected candidate are committed to data/audit.json, the same way RealFloat commits its classification, so the calls can be argued with rather than taken on trust. Two tokens whose sites would not load are excluded from the percentages entirely rather than counted as empty.

Largest first