{
  "added_at": "2026-10-20",
  "arxiv_id": "2609.01185",
  "arxiv_version_reviewed": "v1",
  "authors": [
    "Hadjer Benkraouda",
    "Hongyu Cai",
    "Berkay Celik",
    "Gang Wang"
  ],
  "bibtex_url": "https://awesomerea.org/papers/citations/paper-reveree.bib",
  "code_url": null,
  "date_basis": "First arXiv submission; no reviewed venue publication is established.",
  "date_basis_zh": "采用首次 arXiv 提交日期；未确立正式期刊或会议出版。",
  "date_precision": "day",
  "description_en": "Evaluates reverse-engineering agent traces through flag recovery, eight analysis milestones and action profiles. Deterministic checks and an outcome-blinded LLM judge help locate failures beyond a single solve-rate score.",
  "description_zh": "从旗标恢复、八个分析阶段和操作特征评估逆向智能体轨迹，通过确定性检查与不知道最终结果的模型裁判定位失败环节。",
  "doi": "10.48550/arxiv.2609.01185",
  "entry_url": "https://awesomerea.org/papers/paper-reveree/",
  "entry_url_zh": "https://awesomerea.org/zh/papers/paper-reveree/",
  "evidence_urls": [
    "https://arxiv.org/abs/2609.01185v1",
    "https://arxiv.org/html/2609.01185v1",
    "https://api.datacite.org/dois/10.48550/arxiv.2609.01185"
  ],
  "first_arxiv_date": "2026-09-01",
  "id": "paper-reveree",
  "json_url": "https://awesomerea.org/papers/paper-reveree/record.json",
  "kind": "research",
  "markdown_url": "https://awesomerea.org/papers/paper-reveree/index.md",
  "markdown_url_zh": "https://awesomerea.org/zh/papers/paper-reveree/index.md",
  "metadata_notes": "Reviewed arXiv v1 and registered DOI. The paper announces a release, but the reviewed manuscript does not identify a dedicated artifact URL.",
  "metadata_notes_zh": "已核对 arXiv v1 及注册 DOI。论文宣布发布，但所查手稿未给出专用产物链接。",
  "pdf_url": "https://arxiv.org/pdf/2609.01185v1",
  "project_url": null,
  "publication_date": "2026-09-01",
  "publication_status": "preprint",
  "record_type": "paper",
  "runtime_tested": false,
  "scope_limitations_en": "Uses 88 public CTF challenges and nine models; all four prompting strategies are compared on four models. Evaluation fixes the D-CIPHER harness and uses one attempt per setting. One expert checks 15 traces; agreement on judged comprehension stages is 53/60. Recall and surface-perturbation probes do not eliminate contamination.",
  "scope_limitations_zh": "涵盖 88 个公开 CTF 题和九种模型，仅其中四种比较全部四种提示策略；固定 D-CIPHER 框架，每配置仅一次尝试。一位专家检查 15 条轨迹，理解阶段一致为 53/60；记忆与表面扰动探测不能消除数据污染。",
  "targets": [
    "binary"
  ],
  "title": "Reveree: Diagnosing LLM Reverse-Engineering Agents",
  "title_zh": "Reveree：诊断大语言模型逆向工程智能体",
  "url": "https://arxiv.org/abs/2609.01185",
  "venue": "arXiv",
  "verification": "source-reviewed",
  "verified_at": "2026-10-20",
  "why_fits_en": "Adds process-level evaluation for binary-analysis agents.",
  "why_fits_zh": "补充二进制分析智能体的过程评测。",
  "year": 2026
}