Show HN: Weir - Catch credential leaks in the AI agents you build and use
Hi HN. Weir is for one problem: credentials crossing the boundaries of your AI agent. For agents you build, Weir plants fake credentials where real ones would live, runs the agent's normal flows, and checks what comes out: telemetry, logs, console output, model requests, and requests to other services. It looks for those exact values, including common encoded forms, and reports where they went without printing the credential. pip install weir-scan weir canary judge --sample That sample shows a p
阅读原始来源
来源记录
指标快照
points · 3 votes
2026-10-28 07:30:10.471Z
cumulative since publication
研究分析
项目记录
仓库：IdoGol24/weir。 来源描述：Hi HN. Weir is for one problem: credentials crossing the boundaries of your AI agent. For agents you build, Weir plants fake credentials where real ones would live, runs the agent's normal flows, and checks what comes out: telemetry, logs, console output, model requests, and requests to other services. It looks for those exact values, including common encoded forms, and reports where they went without printing the credential. pip install weir-scan weir canary judge --sample That sample shows a p
采用记录与比较对象
来源票数: 3 votes (2026年10月28日 UTC; cumulative since publication)；前次同口径记录为 3（2026年10月28日 UTC），两个观测值不等于需求增长。 可供比较的近期同类记录：“Show HN: Cerea – yet another self-hosted LLM chat front end” (Show HN, 2026年11月5日)；“Show HN: Lantern Lattice – the call graph from your compiler, for coding agents” (Show HN, 2026年11月5日)。主题相近仅用于导航，未核验其是否属于同一需求或独立证据。
采用判断
应与它要替代的工具或人工流程比较。现有记录支持发现项目，尚不能确认可靠性、运行成本优势或商业采用。
尚未核验的门槛
尚未核验：许可证权限、硬件要求、安装步骤、维护状态，以及包含失败样例的代表性任务。
下一步试验
选择一个真实任务和现有替代方案，记录 commit、环境、输入、输出、耗时与失败行为；保留结果后再给出采用建议。