# HYD

> Public articles, projects, and profile for HYD.

## Core

- [About HYD](https://hydblog.xyz/en/about.md)
- [Projects](https://hydblog.xyz/en/projects/)
- [Answers](https://hydblog.xyz/en/answers/)
- [Profile JSON](https://hydblog.xyz/en/api/profile.json)
- [Full corpus](https://hydblog.xyz/en/llms-full.txt)
- [OpenAPI](https://hydblog.xyz/openapi.json)
- [About index](https://hydblog.xyz/.well-known/about.json)
- MCP endpoint (primary): POST https://ask.hydblog.xyz/mcp
- NLWeb v0.55-compatible restricted /ask subset: POST https://ask.hydblog.xyz/ask
- MCP ask: POST https://ask.hydblog.xyz/mcp (tool: ask, Streamable HTTP)
- Capability boundary: Supports POST /ask with conversational_search, list, summarize, buffered SSE, and MCP ask over Streamable HTTP. Does not support /await, promise responses, elicitation, chatgpt_app, arbitrary extension fields, result actions, or long-term memory.
- Protocol profile: undeclared

## Answers

- [How can AI agents read this site?](https://hydblog.xyz/en/answers/agent-ready/): The build generates llms.txt and llms-full.txt indexes, a Markdown mirror per page, JSON endpoints for articles, profile, and search, plus an OpenAPI description — so agents can cite material without parsing page layouts.
- [How can I collaborate with HYD?](https://hydblog.xyz/en/answers/collaborate/): Email HYDhyd0505@gmail.com, or open an issue or a pull request as HYDtomako on GitHub — AI agent practice, open-source collaboration, and publishing setups are all welcome.
- [How do search and asking work here?](https://hydblog.xyz/en/answers/live-ask/): A question matches pre-generated curated answers first; then it searches the published articles, notes, and answers; anything still unmatched goes to a live service that answers from retrieved public content, single-turn and source-cited.
- [What AI practices does this site document?](https://hydblog.xyz/en/answers/ai-practice/): Three layers — study notes on transformers and deep learning, source-reading notes on agent mechanics (agent loop, tools, context compression, multi-agent collaboration), and small tools built from them, such as EnCoder and QQbot.
- [What is this site?](https://hydblog.xyz/en/answers/what-is-this-site/): HYD's personal site, recording the path through AI engineering and LLMs — articles, notes, projects, and a curated FAQ, built with Astro, Starlight, and the Refined-X starter so one corpus serves readers, search engines, and AI agents.
- [Where were the earlier posts published?](https://hydblog.xyz/en/answers/where-articles-were/): On CSDN from the summer of 2025 (blog.csdn.net/2401_87876529) — Python and data analysis first, then transformers, LLMs, and agents. New writing moved here in September 2026; the old posts stay up but are no longer updated.
- [Which open source projects are here?](https://hydblog.xyz/en/answers/open-source-projects/): Six so far — EnCoder, QQbot, Minimind-notes, Simple_CNN, interview-skill, and homework-web_blog — mostly agent tooling and reproductions of deep-learning basics.
- [Who is the author?](https://hydblog.xyz/en/answers/who-is-author/): HYD is a Data Science and Big Data Technology student at Qingdao Institute of Technology (class of 2024), focused on AI agents and AI-native products — he likes pushing what he learns into small tools that actually run, then open-sourcing them.

## Recent articles

- [Three Claude Code implementations: an agent engineering notebook](https://hydblog.xyz/en/2026/08/23/agent-engineering-from-claude-code.md): Agent engineering notes drawn from three implementations. CoreCoder is a minimal implementation written by hand in Python, Claude Code is the production implementation, and learn-claude-code is the teaching implementation. Coverage spans the Agent Loop's transition condition and interrupt repair, tool definitions and file-editing strategy, context injection and three tiers of compression, the long-lived state of QueryEngine, parallel tools and sub-agents, the Hook and Task extension points, multi-agent collaboration, memory / MCP / background tasks / scheduled tasks, and the security defenses around paths and commands.
- [Transformer learning notes: from the overall architecture to multi-head attention](https://hydblog.xyz/en/2026/04/04/transformer-learning-notes.md): Transformer study notes ordered whole-first and then part-by-part: the data flow through the encoder and the decoder; the three questions to answer before learning large models (neural networks, attention, what PyTorch is made of); what Token, the word embedding matrix, and positional encoding each solve; the meaning of Q, K, and V, why the Q·K dot product stands for relevance, and why the dot product is divided by the square root of the dimension before softmax; the difference between self-attention and cross-attention, and why cross-attention needs no mask; training predicts several tokens in parallel while inference generates one token at a time, which is why a causal mask is needed to prevent peeking, implemented by setting later positions to negative infinity so that softmax turns them into 0; the definitions of the feed-forward network, multi-head attention, and layer normalization, including how layer norm and batch norm differ in which dimension they normalize over.