{
  "added_at": "2026-11-02",
  "additional_evidence_urls": [
    "https://github.com/WPeace-HcH/WPeGPT/blob/8a7802aef9ba980dcb5d2a6a24c6ef19b06ae641/README.ZH_CN.md",
    "https://github.com/WPeace-HcH/WPeGPT/blob/8a7802aef9ba980dcb5d2a6a24c6ef19b06ae641/requirements.txt",
    "https://github.com/WPeace-HcH/WPeGPT/blob/8a7802aef9ba980dcb5d2a6a24c6ef19b06ae641/WPeGPT_Config/config.py",
    "https://github.com/WPeace-HcH/WPeGPT/blob/8a7802aef9ba980dcb5d2a6a24c6ef19b06ae641/WPeGPT.py#L443-L460",
    "https://github.com/WPeace-HcH/WPeGPT/releases/tag/v3.0"
  ],
  "category": "agent-workflows",
  "description_en": "IDA plugin, formerly WPeChatGPT and inspired by Gepetto, that sends the Hex-Rays pseudocode of a function to an OpenAI-compatible model (OpenAI, DeepSeek or another compatible endpoint) and writes the model's analysis back into IDA as comments. Interactive actions explain a function's purpose and behavior, suggest variable names, re-express small routines such as XOR decoding in Python and flag potential vulnerabilities. Version 3.0 adds WPeServer, a TCP command server inside IDA through which an external controller or the separate wpegpt-analyzer skill runs headless light, full or vulnerability-focused passes, with string classification, network-indicator extraction, function suspiciousness scoring and JSON plus Markdown reports.",
  "description_zh": "IDA 插件，前身为 WPeChatGPT，受 Gepetto 启发：把函数的 Hex-Rays 伪代码发送给兼容 OpenAI 接口的模型（OpenAI、DeepSeek 或其他兼容端点），再将模型的分析结果以注释形式写回 IDA。交互功能可说明函数的用途与行为、建议变量名、用 Python 重写 XOR 解码等小型例程，并标记潜在漏洞。3.0 版新增 WPeServer，即运行在 IDA 内的 TCP 命令服务器，外部控制器或独立的 wpegpt-analyzer 技能可借此以无界面方式执行 light、full 或侧重漏洞的分析流程，提供字符串分类、网络指标提取、函数可疑度评分，并输出 JSON 与 Markdown 报告。",
  "entry_url": "https://awesomerea.org/resources/wpegpt/",
  "entry_url_zh": "https://awesomerea.org/zh/resources/wpegpt/",
  "evidence_url": "https://github.com/WPeace-HcH/WPeGPT/blob/8a7802aef9ba980dcb5d2a6a24c6ef19b06ae641/README.md",
  "group": "agents",
  "id": "wpegpt",
  "json_url": "https://awesomerea.org/resources/wpegpt/record.json",
  "kind": "integration",
  "markdown_url": "https://awesomerea.org/resources/wpegpt/index.md",
  "markdown_url_zh": "https://awesomerea.org/zh/resources/wpegpt/index.md",
  "name": "WPeGPT",
  "name_zh": "WPeGPT IDA AI 分析插件",
  "record_type": "resource",
  "relationship": "independent-agent-integration",
  "requirements": [
    "Reviewed at tag v3.0 (2026-05-27), which is the reviewed main commit; the GitHub release has no attached files. Install by copying WPeGPT.py and the WPeGPT_Config folder (config.py and wpe_ai_controller.py) into IDA's plugins directory. IDA must run Python 3, and the plugin imports ida_hexrays, so a Hex-Rays decompiler for the target architecture is needed. Python dependencies are openai >= 0.27.0 and httpx.",
    "config.py ships with empty API_KEY, API_BASE_URL and MODEL values that must be set for an OpenAI-compatible provider, plus an optional forward proxy, worker count and default analysis mode (light); the interface defaults to Chinese (ZH_CN = True). The README estimates light, full and vuln runs at roughly 2–5, 10–30 and 5–20 minutes. WPeServer binds to 127.0.0.1 on an OS-assigned port, and headless runs start from the IDA menu or the separately maintained wpegpt-analyzer skill.",
    "Works on a binary the user opens in IDA; the repository bundles only the plugin, its configuration and screenshots. Each request sends decompiled pseudocode to the configured model provider, and automated runs send many functions, so weigh confidentiality and API cost before analyzing sensitive samples. The README calls AI results for reference only; comments and reports are model suggestions that need analyst checking.",
    "No license file is present and GitHub detects no license, so no reuse terms are granted. IDA, the Hex-Rays decompiler and model APIs are licensed separately."
  ],
  "requirements_zh": [
    "审阅版本为标签 v3.0（2026-05-27），即所审阅的 main 提交；该 GitHub 发布版未附带文件。安装方法是把 WPeGPT.py 和 WPeGPT_Config 文件夹（含 config.py 与 wpe_ai_controller.py）复制到 IDA 的 plugins 目录。IDA 必须使用 Python 3；插件导入 ida_hexrays，因此需要适用于目标架构的 Hex-Rays 反编译器。Python 依赖为 openai >= 0.27.0 和 httpx。",
    "config.py 中的 API_KEY、API_BASE_URL 和 MODEL 默认为空，需按兼容 OpenAI 接口的服务商填写；另可配置正向代理、并发数和默认分析模式（light），界面默认使用中文（ZH_CN = True）。README 估计 light、full 和 vuln 模式分别约需 2–5、10–30 和 5–20 分钟。WPeServer 绑定在 127.0.0.1 上由系统分配的端口，无界面分析可从 IDA 菜单启动，或通过另行维护的 wpegpt-analyzer 技能发起。",
    "分析对象是使用者在 IDA 中打开的二进制文件；仓库只包含插件、配置和截图。每次请求都会把反编译伪代码发送给所配置的模型服务商，自动分析会发送大量函数，因此分析敏感样本前应考虑保密性和 API 费用。README 称 AI 结果仅供参考；生成的注释和报告属于模型建议，需要分析者核实。",
    "仓库没有许可证文件，GitHub 也未识别出许可证，因此未授予任何复用条款。IDA、Hex-Rays 反编译器和模型 API 各自另行授权。"
  ],
  "targets": [
    "binary"
  ],
  "url": "https://github.com/WPeace-HcH/WPeGPT",
  "verification": "source-reviewed",
  "verified_at": "2026-10-20"
}