{"_meta":{"billing":{"credits":0,"mode":"free","usd":0.0},"coverage":"full","endpoint":"scorecard","price_usd":0.0},"bands":[{"action":"avoid","exit_window":"none","instant_collapse_pct":59.8,"measured_rug70_pct":84.5,"measured_rug85_pct":78.3,"median_seconds_peak_to_minus70":19,"sample_size":1122,"score_max":null,"score_min":70,"score_range":"70+","verdict":"extreme"},{"action":"avoid","exit_window":"none","instant_collapse_pct":43.0,"measured_rug70_pct":68.8,"measured_rug85_pct":60.3,"median_seconds_peak_to_minus70":52,"sample_size":970,"score_max":69,"score_min":58,"score_range":"58-69","verdict":"critical"},{"action":"caution","exit_window":"seconds","instant_collapse_pct":28.8,"measured_rug70_pct":53.4,"measured_rug85_pct":43.5,"median_seconds_peak_to_minus70":203,"sample_size":2951,"score_max":57,"score_min":42,"score_range":"42-57","verdict":"elevated"},{"action":"caution","exit_window":"minutes","instant_collapse_pct":11.4,"measured_rug70_pct":59.6,"measured_rug85_pct":42.6,"median_seconds_peak_to_minus70":424,"sample_size":1978,"score_max":41,"score_min":22,"score_range":"22-41","verdict":"elevated"},{"action":"clear","exit_window":"minutes","instant_collapse_pct":7.0,"measured_rug70_pct":60.2,"measured_rug85_pct":38.6,"median_seconds_peak_to_minus70":591,"sample_size":1695,"score_max":21,"score_min":8,"score_range":"8-21","verdict":"low"},{"action":"clear","exit_window":"minutes","instant_collapse_pct":5.9,"measured_rug70_pct":57.1,"measured_rug85_pct":34.8,"median_seconds_peak_to_minus70":672,"sample_size":1349,"score_max":7,"score_min":0,"score_range":"0-7","verdict":"low"}],"base_rate_pct":45.0,"calibrated_at":"2026-07-24","says":"Share of tokens in each risk band that dropped 85% or more from peak within an hour, measured on our own labelled outcomes \u2014 and, more usefully, how fast that collapse arrives. Whether a token collapses separates weakly in this market (78% worst band against a 45% base). How fast it collapses separates about tenfold, and decides whether a position is exitable. Match a token's rug_risk.score to the row whose score_min/score_max contains it; verdict is a label, not a key.","window_days":30}