FINDING #5198 · UNIT ID 1322509558
AMAP-ML/LongHorizon-Harness
The long-horizon computer-use harness. Run AI agents across desktop apps and the CLI for extended periods while preserving task state and making reliable progress on complex workflows. Features fresh-context execution, durable verified state, independent auditing, recoverable progress, and native Claude Code / Codex / OpenClaw integration.
SURPRISE SCORE
0.00
Score Breakdown
SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE
SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)
2% OF STARS IN ARCHIVE
Growth Telemetry
VELOCITY /D
5.43
ACCEL
-0.79
RETENTION
38.2%
PEAK 2026-10-12 · FORK-RETENTION 0.0% · 38 STARS / WINDOW
Author Audience
AUDIENCE
3,101
FOLLOWERS
385
OWNER ★
11,657
Engagement Signals
FORKS
189
ISSUE AUTH
0
PR AUTH
0
UNIQUE STARGAZERS 38 / 38 (DIVERSITY 1.00)
Why This Is A Finding
AMAP-ML/LongHorizon-Harness собрал 38 звёзд за окно, тогда как у автора всего 385 подписчиков — эффективная аудитория ≈ 3,101. Это даёт surprise-индекс 0.00173 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 0.0% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация отрицательная — внимание остывает после пика.
Related Findings
RANKS ABOVE 46% OF 9651 FINDINGS
METRICS IN CONTEXT
MEDIAN ACROSS ALL 9651 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW
METRICVALUEMEDIANΔ MEDPERCENTILE
SCORE0.000.00-0.00ABOVE 46%
VELOCITY5.433.14+2.29ABOVE 68%
RETENTION38.2%16.7%+21.5 PPABOVE 73%
FORKS189277-88ABOVE 43%
SURPRISE0.000.00-0.00ABOVE 39%