Skip to content
View zyellowl's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report zyellowl

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
zyellowl/README.md

Hi, I'm zyellowl 👋

我关注 AI Agent、开发者工具与可靠自动化。比起只做演示,我更在意源码证据、可复现实验、端到端验证,以及真正可部署的交付。

Featured work

面向真实个人经历、项目证据与技术文章的公开作品站,包含 Home、Blog、Chat 与 Thoughts 四个入口,并通过 GitHub Pages 持续部署。

Portfolio · AI engineering · Technical writing · GitHub Pages

从源码与核心调用链出发,系统拆解 Codex、Claude Code、Pi 与 DeepSeek Harness。包含 30 章内容、27 条固定源码调用链、16 个可复现实验、133 条工程不变量和 43 张架构/流程图,并用自动化测试校验内容与渲染结果。

TypeScript · Next.js · Source research · Reproducible labs

基于 LibreChat v0.8.7 的个人 AI 对话产品化项目:覆盖 macOS/iOS 客户端、ChatGPT/Codex OAuth 本地桥接、ZenMux 可选模型、上下文检索与服务端密钥隔离,并保留清晰的上游归属。

TypeScript · Swift · LibreChat · OAuth · Self-hosted

面向真实下载与交付场景的 Telegram Bot:保留原始视频尺寸,使用 ffprobe 校验流、时长与分辨率;超出 Telegram 限额时进行无损分段,并覆盖单元测试与真实端到端测试。

Python · Telegram Bot API · yt-dlp · ffmpeg · Docker

Engineering values

  • Evidence over vague claims
  • Reproducible tests over screenshots alone
  • Safe configuration and secret isolation
  • Complete delivery from implementation to deployment verification

Pinned Loading

  1. harness-atlas-deep-core harness-atlas-deep-core Public

    中文 AI Coding Agent Harness 源码教程:Codex、Claude Code、Pi 与 DeepSeek Harness

    TypeScript

  2. chatone chatone Public

    Open-source personal AI chat app built on LibreChat with ZenMux and ChatGPT subscription support.

    TypeScript

  3. telegram-youtube-downloader-bot telegram-youtube-downloader-bot Public

    Telegram bot for downloading YouTube videos with progress, original dimensions, and lossless Telegram-sized segmentation.

    Python