Show HN: Conduit – 基于 SHA-256 哈希链的无头浏览器 - Ed25519 审计追踪
2 分•作者: TaxFix•5 个月前
我一直在构建 AI 智能体工具，并反复遇到同样的问题：智能体浏览网页、执行操作、填写表格、抓取数据——但对于实际发生了什么，却没有任何证据。截图可以伪造。日志可以编辑。如果出现问题，你只能对着一个黑盒子指手画脚。
因此，我构建了 Conduit。它是一个无头浏览器（底层使用 Playwright），将每个操作记录到 SHA-256 哈希链中，并使用 Ed25519 对结果进行签名。每个操作都会与前一个哈希值进行哈希运算，形成一个防篡改链。在会话结束时，你会得到一个“证明包”——一个 JSON 文件，其中包含完整的操作日志、哈希链、签名和公钥。任何人都可以独立验证该包，而无需信任生成它的当事方。
我主要针对以下用例：
- *AI 智能体审计*——你给智能体一个浏览器。稍后你需要证明它做了什么。Conduit 为你提供加密收据。
- *合规性自动化*——SOC 2、GDPR 数据主体访问工作流程，任何你需要证明流程正确运行的情况。
- *网页抓取溯源*——证明你收集的数据确实来自你所说的位置，并在你所说的时间收集。
- *诉讼支持*——使用可验证的监管链捕获网络内容。
它还作为 MCP（模型上下文协议）服务器提供，因此 Claude、GPT 和其他基于 LLM 的智能体可以通过工具调用本地使用该浏览器。智能体可以进行浏览、点击、填写、截图，并且证明包会在后台自动构建。
免费、MIT 许可、纯 Python。无需帐户、无需 API 密钥、无需遥测。
GitHub：https://github.com/bkauto3/Conduit
安装：`pip install conduit-browser`
欢迎对证明包格式和 MCP 集成提供反馈。很乐意回答有关加密设计的问题。
查看原文
I've been building AI agent tooling and kept running into the same problem: agents browse the web, take actions, fill out forms, scrape data -- and there's zero proof of what actually happened. Screenshots can be faked. Logs can be edited. If something goes wrong, you're left pointing fingers at a black box.<p>So I built Conduit. It's a headless browser (Playwright under the hood) that records every action into a SHA-256 hash chain and signs the result with Ed25519. Each action gets hashed with the previous hash, forming a tamper-evident chain. At the end of a session, you get a "proof bundle" -- a JSON file containing the full action log, the hash chain, the signature, and the public key. Anyone can independently verify the bundle without trusting the party that produced it.<p>The main use cases I'm targeting:<p>- *AI agent auditing* -- You hand an agent a browser. Later you need to prove what it did. Conduit gives you cryptographic receipts.
- *Compliance automation* -- SOC 2, GDPR data subject access workflows, anything where you need evidence that a process ran correctly.
- *Web scraping provenance* -- Prove that the data you collected actually came from where you say it did, at the time you say it did.
- *Litigation support* -- Capture web content with a verifiable chain of custody.<p>It also ships as an MCP (Model Context Protocol) server, so Claude, GPT, and other LLM-based agents can use the browser natively through tool calls. The agent gets browse, click, fill, screenshot, and the proof bundle builds itself in the background.<p>Free, MIT-licensed, pure Python. No accounts, no API keys, no telemetry.<p>GitHub: https://github.com/bkauto3/Conduit<p>Install: `pip install conduit-browser`<p>Would love feedback on the proof bundle format and the MCP integration. Happy to answer questions about the cryptographic design.