leo-lilinxiao/codex-autoresearch — GitHub trend

FINDING #1259 · UNIT ID 1184866064

leo-lilinxiao/codex-autoresearch

Codex Autoresearch Skill — A self-directed iterative system for Codex that continuously cycles through: modify, verify, retain or discard, and repeat indefinitely. Inspired by Karpathy’s autoresearch concept.

[ PYTHON ][ GITHUB ↗ ]

SURPRISE SCORE

0.00

Score Breakdown

SURPRISE0.0384

ENGAGEMENT0.24

FRESHNESS1.27

SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE

SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)

14% OF STARS IN ARCHIVE

Growth Telemetry

VELOCITY /D

12.00

ACCEL

+0.32

RETENTION

51.0%

PEAK 2026-09-08 · FORK-RETENTION 11.1% · 360 STARS / WINDOW

Author Audience

AUDIENCE

272

FOLLOWERS

OWNER ★

2,506

Engagement Signals

FORKS

130

ISSUE AUTH

PR AUTH

UNIQUE STARGAZERS 360 / 360 (DIVERSITY 1.00)

Why This Is A Finding

leo-lilinxiao/codex-autoresearch собрал 360 звёзд за окно, тогда как у автора всего 23 подписчиков — эффективная аудитория ≈ 272. Это даёт surprise-индекс 0.0384 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 11.1% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация положительная — рост ещё не выдохся.

Related Findings

RANKS ABOVE 87% OF 9679 FINDINGS

#13Kizzuwatnaa/DLSS5-Autopilot[ PYTHON ]0.33 #16MerZlin/dsh-pet-indesktop[ PYTHON ]0.30 #21LBH-123-AI/Comfyui_Minimax_h3_latent_Upscaler[ PYTHON ]0.28 #23User0332/rewards-farmer[ PYTHON ]0.27 #242akouwu/reverify[ PYTHON ]0.27

METRICS IN CONTEXT

MEDIAN ACROSS ALL 9679 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW

METRICVALUEMEDIANΔ MEDPERCENTILE

SCORE0.020.00+0.02ABOVE 87%

VELOCITY12.003.80+8.20ABOVE 82%

RETENTION51.0%25.0%+26.0 PPABOVE 91%

FORKS130112+18ABOVE 54%

SURPRISE0.040.01+0.03ABOVE 86%