{
  "added_at": "2026-10-16",
  "arxiv_id": "2605.10597",
  "arxiv_version_reviewed": "v1",
  "authors": [
    "Isaac David",
    "Arthur Gervais"
  ],
  "bibtex_url": "https://awesomerea.org/papers/citations/paper-crackmebench.bib",
  "code_url": null,
  "date_basis": "arXiv v1 submission",
  "date_basis_zh": "arXiv v1 首次提交日期。",
  "date_precision": "day",
  "description_en": "Evaluates agents on executable-checked CrackMe tasks: infer validation logic from binaries and submit an accepted input or artifact, while recording tool use, time, and scoring attempts.",
  "description_zh": "用可执行程序判定结果的 CrackMe 任务评估智能体：从二进制推断验证逻辑，提交可被接受的输入或产物，并记录工具使用、耗时与评分尝试。",
  "doi": null,
  "entry_url": "https://awesomerea.org/papers/paper-crackmebench/",
  "entry_url_zh": "https://awesomerea.org/zh/papers/paper-crackmebench/",
  "evidence_urls": [
    "https://arxiv.org/abs/2605.10597",
    "https://arxiv.org/html/2605.10597v1"
  ],
  "first_arxiv_date": "2026-05-11",
  "id": "paper-crackmebench",
  "json_url": "https://awesomerea.org/papers/paper-crackmebench/record.json",
  "kind": "research",
  "markdown_url": "https://awesomerea.org/papers/paper-crackmebench/index.md",
  "markdown_url_zh": "https://awesomerea.org/zh/papers/paper-crackmebench/index.md",
  "metadata_notes": "No conference venue is stated in the reviewed arXiv record. Tool repositories cited in its bibliography are not treated as the paper's own code.",
  "metadata_notes_zh": "所审阅 arXiv 记录未标明会议发表场所。参考文献中的工具仓库不视为论文自己的代码产物。",
  "pdf_url": "https://arxiv.org/pdf/2605.10597",
  "project_url": null,
  "publication_date": "2026-05-11",
  "publication_status": "preprint",
  "record_type": "paper",
  "runtime_tested": false,
  "scope_limitations_en": "The v0 study uses 12 generated tasks and eight public calibration tasks, Linux x86-64 ELF binaries, a shell interface, and a five-minute budget. Its compact educational tasks do not establish performance on large applications. No paper-linked public artifact URL was verified.",
  "scope_limitations_zh": "v0 研究使用 12 个生成任务和 8 个公开校准任务，限定 Linux x86-64 ELF、Shell 接口和五分钟预算。紧凑的教学任务不能证明大型应用上的表现。未确认论文直接关联的公开研究产物 URL。",
  "targets": [
    "binary"
  ],
  "title": "CrackMeBench: Binary Reverse Engineering for Agents",
  "title_zh": "CrackMeBench：面向智能体的二进制逆向工程基准",
  "url": "https://arxiv.org/abs/2605.10597",
  "venue": "arXiv",
  "verification": "source-reviewed",
  "verified_at": "2026-10-16",
  "why_fits_en": "Directly measures whether an autonomous reverse-engineering workflow produces a verifiably accepted result.",
  "why_fits_zh": "直接衡量自主逆向工程流程能否产生可验证、被目标程序接受的结果。",
  "year": 2026
}