跳到正文

正在自动登录…

← 返回项目列表

Blazenumenhir/scam-story-checker

信息核查证据保存线索追踪

本地运行的免费工具,用于在交易、求职或租房前快速核查对方说辞与公开记录是否一致,并保存证据。适合个人在付款、签约前自行核实对方身份。

AI 中文解读 · 根据仓库描述与 README 生成 · 2026/10/31 16:47:02 · 查看原文

值得看JavaScript应用与服务
累计 Stars
163
当前热度
未达当前门槛
仓库创建时间
打开 GitHub ↗

选用指南

AI 根据 README 整理 · 2026/10/31 16:47:02

适用场景

适用于交易前核查卖家、招聘方或房东的说辞,也可自查公开信息。

查看 README 依据
<img src="assets/alibi.gif" alt="Picking 'A seller', typing a nickname, pasting what the seller said, and getting the verdict: this seller's story doesn't hold up" width="880"> </p> <h3 align="center">Spot a scammer like a pro investigator. 30 seconds, $0.</h3> ## They have a story. Does it hold up? The seller with a boxed console and a "bank transfer first" rule. The recruiter paying £42 an hour for data entry. The landlord who's somehow never free for a viewing.
> [!NOTE] > `Main.exe` is code-signed, so Windows starts it without a warning. Keep `alibi` next to it: they ship together. ## Does it hold up? Start by saying who they are: **a seller**, **a recruiter or job**, **a landlord**, **a website or shop**, or **yourself**. Alibi runs the checks that matter for that kind of story and words the verdict to fit. | Verdict | What it means | |---|---| | **Holds up** | The public trail matches what they told you. | | **Doesn't add up** | Some of it checks out and some of it doesn't. Alibi says which part. |
使用方式

Windows 10/11 便携版,下载后无需安装,选择核查对象并粘贴已知信息即可使用。

查看 README 依据
<td><b>Timeline.</b> Real people leave a long trail. Fake ones start last week.</td> </tr> </table> ## Get it in 3 steps **1. Download** the file from the [latest release](https://github.com/Blazenumenhir/scam-story-checker/releases/latest): | System | File | |---|---| | Windows 10 and 11 | [`scam-story-checker.zip`](releases/latest) | **2. Open it.** No installer, no admin rights, no sign-up. **3. Pick who you're checking, paste what you know, press Enter.** A nickname, an email, a phone number or a link is enough. Paste what they told you too, and Alibi checks that as well.
部署要求

需要 Windows 10 或 11;可选安装 Ollama 以使用本地 AI 摘要。

查看 README 依据
<td><b>Timeline.</b> Real people leave a long trail. Fake ones start last week.</td> </tr> </table> ## Get it in 3 steps **1. Download** the file from the [latest release](https://github.com/Blazenumenhir/scam-story-checker/releases/latest): | System | File | |---|---| | Windows 10 and 11 | [`scam-story-checker.zip`](releases/latest) | **2. Open it.** No installer, no admin rights, no sign-up. **3. Pick who you're checking, paste what you know, press Enter.** A nickname, an email, a phone number or a link is enough. Paste what they told you too, and Alibi checks that as well.
- **Check yourself.** See what anyone can find about you, then clear it up: opt-out links for people-search sites, old accounts to delete, breaches to deal with. - **Watch.** Get told when something new shows up. It works for yourself, your own domains and companies, never for a private person, and every watch has an end date. - **Local AI.** Summaries run on your PC through [Ollama](https://ollama.com), or you can switch AI off. Nothing leaves the PC unless you pick an online model yourself. - **One-click PDF.** Exhibits, hashes and the case log in one file. <p align="center">
已知限制

不查家庭住址、不进行人脸搜索、不登录任何网站、不显示实时位置、不探测账号是否存在、不频繁请求网站。

查看 README 依据
<img src="assets/yourself.png" alt="Check yourself: people-search opt-outs, old accounts and breaches as a checklist" width="880"> </p> ## What Alibi won't do - Look up home addresses. - Search for a stranger's face. Photo matching compares image hashes, never faces. - Log in anywhere, or read anything behind a login. - Show anyone's live location. - Probe password-reset or sign-up pages to see if an account exists. - Hammer websites. Requests are polite and rate-limited. Every case keeps a log of what was checked and when. ## Use it right

README

正在加载 README…

收录信息

仓库创建时间
2026/10/31 06:02:39
首次达标收录
2026/10/31 16:42:16
统计更新时间
2026/11/1 20:31:26

按创建时长与累计 Stars 判断等级。首次达标收录后,即使未达当前门槛仍保留;累计 Stars 和仓库状态会定期更新。确认只有 README 及辅助文件的仓库会停止公开。

模型评估

值得看加权总分 4.1 / 5

新颖度
3.7
实用价值
4.9
成熟度
4.7
实质含量
3.4

由 Jev 模型根据仓库描述与 README 评估,2026/10/31 16:42:16 生成。评分仅供参考,不是代码、安全或许可证审查;热度等级另按创建时长与累计 Stars 计算。

仓库信息

最近推送
2026/10/31 06:04:37
许可证标识
MIT
Fork 数
11
当前状态
未归档

收录记录

按创建时长与累计 Stars 首次收录:小火
首次发现项目