← Knowledge

◇ Paper · How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks

Key facts · machine-readable
Official / sourcehttps://huggingface.co/papers/2608.14905
Chinese editor copy (original, collapsed)

① 这是什么

今日论文《今日论文 · How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks》（Yanlin Fei、Nazhou Liu、Xinmiao Yu、Shaolong Chen、Lei Li、Rahul Thapa），来自 HF Daily Papers 公开源。

② 值不值得用

进入 HF 每日精选说明有社区关注度；是否与你的问题相关需要读原文判断，本条目不评价研究质量。

③ 怎么开始

先读摘要，需要再读全文（https://huggingface.co/papers/2608.14905）；对照论文 ID（2608.14905）可找实现与讨论。

④ 关键证据

  • 论文 ID 2608.14905
  • 作者 Yanlin Fei、Nazhou Liu、Xinmiao Yu、Shaolong Chen、Lei Li、Rahul Thapa

⑤ 注意事项与坑

摘要来自作者原文，未做同行评议级核验；引用请以正式发表版本为准。

One-line premise

Good for: readers tracking today's AI research frontier.

Not for: reproducing full experimental details (read the paper).

Status: machine-checked (structure & source validation); functionality not individually tested.

Minimal verification (3 steps)

  1. Read abstract — How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks
  2. Check authors/date — Yanlin Fei, Nazhou Liu, Xinmiao Yu · 2026-08-14
  3. Run CLI — npx aiskillready search 14905
Quick start

Use it like this

npx aiskillready search 14905   # official CLI, verified runnable (aiskillready@0.2.0)

Note: `npx aiskillready` is this site's published CLI (npm aiskillready@0.2.0) — verified runnable; third-party commands are given per official docs and not individually tested. We don't fabricate conclusions.

Open source ↗
Trust layer: machine check / cost / confidence / schema (expand)

Entry ID: ASR-PAPER-20260818-2608-14905 ｜ Type: Paper ｜ schema v0.5

✅ machine-checked pass 💰 Enrichment cost $0.0000 🧭 Confidence medium 📌 schema 0.5 🛂 Review agent reviewed 🧪 Functionality unverified (Top-20 manual testing planned)

Related (what you should know)

Related (3) · expand
Similar

◇ Paper · Prototype-Rectified Iterative Self-supervised Manifold Denoising under Severe Acoustic Shift

Prototype-Rectified Iterative Self-supervised Manifold Denoising under Severe Acoustic Shift. by Ashish Anand Shukla, Rini Smita Thakur, Aryan Das, Vinod K. Kurmi. published 2026-08-15.

 Similar

◇ Paper · MOSS-VL Technical Report

MOSS-VL Technical Report. by Pengyu Wang, Chenkun Tan, Shaojun Zhou, Qirui Zhou, Yanxin Chen, Xingyang He, Huazheng Zeng, Jijun Cheng, Chenghao Wang,. published 2026-08-15.

 Similar

◇ Paper · WorldRover: A Scalable Synthetic Video Data Engine for World Exploration with Rich Annotations

WorldRover: A Scalable Synthetic Video Data Engine for World Exploration with Rich Annotations. by Xiaojie Xu, Zhengyuan Lin, Runyi Li, Yihao Liu, Kaipeng Zhang, Yongtao Ge. published 2026-08-16.

Feedback

Something wrong / command failed? File an issue with the entry ID; it feeds back into the next revision.

🐞 Report an issue →
AI block · machine-readable (same source of truth)
AI block · machine-readable JSON (click to expand) 
{
  "schema_version": "0.5",
  "entry_id": "ASR-PAPER-20260818-2608-14905",
  "record_type": "paper",
  "title": "今日论文 · How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks",
  "observed_at": "2026-08-18",
  "source": {
    "name": "huggingface.co/api/daily_papers",
    "url": "https://huggingface.co/api/daily_papers?limit=10",
    "snapshot_hash": "652c1f13fb67dd93f7d60fecc11565d77173419a1e4693958922f4b73738fb42",
    "fetched_at": "2026-08-18T10:51:18+00:00"
  },
  "confidence": "medium",
  "human_stream": {
    "summary": "AI has long assisted scientific research, but the rapid advance of LLMs and agentic scaffolds is reshaping the landscape; a single system can now carry whole-stage research from an initial hypothesis all the way to final published paper, which is a paradigm now referred to as AutoResearch. Existing evaluations reveal little about how these agents operate or where they break down. Tasks are narrowl",
    "key_points": [
      "论文 ID 2608.14905",
      "作者 Yanlin Fei、Nazhou Liu、Xinmiao Yu、Shaolong Chen、Lei Li、Rahul Thapa",
      "发布 2026-08-14T00:00:00.000Z"
    ],
    "note": "来源为 HuggingFace daily papers 公开 API；完整阅读请走原文链接。",
    "what_it_is": "今日论文《今日论文 · How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks》（Yanlin Fei、Nazhou Liu、Xinmiao Yu、Shaolong Chen、Lei Li、Rahul Thapa），来自 HF Daily Papers 公开源。",
    "worth_it": "进入 HF 每日精选说明有社区关注度；是否与你的问题相关需要读原文判断，本条目不评价研究质量。",
    "how_to_start": "先读摘要，需要再读全文（https://huggingface.co/papers/2608.14905）；对照论文 ID（2608.14905）可找实现与讨论。",
    "evidence": [
      "论文 ID 2608.14905",
      "作者 Yanlin Fei、Nazhou Liu、Xinmiao Yu、Shaolong Chen、Lei Li、Rahul Thapa"
    ],
    "cautions": "摘要来自作者原文，未做同行评议级核验；引用请以正式发表版本为准。"
  },
  "ai_stream": {
    "structured": {
      "paper_id": "2608.14905",
      "title": "How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks",
      "authors": [
        "Yanlin Fei",
        "Nazhou Liu",
        "Xinmiao Yu",
        "Shaolong Chen",
        "Lei Li",
        "Rahul Thapa",
        "Madalina Ciobanu",
        "Qingqing Mao",
        "Ritankar Das"
      ],
      "published_at": "2026-08-14T00:00:00.000Z",
      "url": "https://huggingface.co/papers/2608.14905",
      "source_label": "hf-daily-papers"
    },
    "raw": [
      {
        "paper_id": "2608.14905",
        "title": "How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks",
        "authors": [
          "Yanlin Fei",
          "Nazhou Liu",
          "Xinmiao Yu",
          "Shaolong Chen",
          "Lei Li",
          "Rahul Thapa",
          "Madalina Ciobanu",
          "Qingqing Mao",
          "Ritankar Das"
        ],
        "summary": "AI has long assisted scientific research, but the rapid advance of LLMs and agentic scaffolds is reshaping the landscape; a single system can now carry whole-stage research from an initial hypothesis all the way to final published paper, which is a paradigm now referred to as AutoResearch. Existing evaluations reveal little about how these agents operate or where they break down. Tasks are narrowly-scoped, evaluation measures performance but not process, and failure diagnoses lack systematic coverage or artifact-level visibility. To address this gap, we introduce AutoResearchEval, featuring 10",
        "published_at": "2026-08-14T00:00:00.000Z",
        "submitted_on_daily_at": null,
        "url": "https://huggingface.co/papers/2608.14905"
      }
    ]
  },
  "token_cost": {
    "total": 0.0,
    "currency": "USD",
    "breakdown": {
      "crawl": 0.0,
      "clean": 0.0,
      "elevate": 0.0,
      "verify": 0.0
    }
  },
  "machine_verified": true,
  "review": {
    "status": "agent_reviewed",
    "reviewer": "pipeline-validate",
    "reviewed_at": "2026-08-18T10:54:20+00:00",
    "comments": "机检通过（schema 0 error + 溯源一致 + 成本达标）"
  },
  "provenance": {
    "extracted_by": "codex",
    "extracted_at": "2026-08-18T10:53:05+00:00",
    "pipeline": "collect_hf_papers.py + build_entries.py v0.2（规则管线）",
    "access_urls": [
      "https://huggingface.co/api/daily_papers?limit=10"
    ],
    "card_generated_at": "2026-08-18T10:53:05+00:00",
    "card_pipeline": "enrich_human_stream.py v0.1（规则模板，零 Token）"
  }
}