{
  "video_id": "eaedb25653d1ee27",
  "channel": "export-arxiv-org-rss-cs-ai",
  "title": "Agents' Last Exam",
  "scores": {
    "depth": 3,
    "novelty": 3,
    "aria_relevance": 2,
    "production_ready": 1
  },
  "junk_penalty": 0,
  "avg_score": 2.25,
  "effective_score": 2.25,
  "verdict": "promote",
  "one_line_reason": "Fundamental eval-framework für Real-World-Agent-Performance mit starker empirischer Grundlage (250+ Experten, 1K+ Tasks, 2.6% Pass-Rate als Reality-Check); hochrelevant für Aria's Eval-Setup und Agent-Capability-Messung, aber noch nicht direkt in Code/Workflow umsetzbar.",
  "model": "claude-haiku-4-5-20251001",
  "cost_usd": 0.0019169999999999999,
  "triaged_at": "2026-06-06T15:00:49.340357+00:00"
}