Library
WineChord Books
A compact shelf for source reading, agent runtime, coding practice, system design, and interview preparation.
Start Here
Good first paths
Featured path Prompt Cache and Context Budget Start from optional context, prompt assembly, typed fragments, and the budget discipline behind long-running agent work. Source track Codex From Source Read the full runtime map when you want tools, sandboxing, app-server, plugins, and multi-agent work tied back to code. Systems track System Design Interview Book Use the ChatGPT case study when you want product, backend, inference, and safety design in one interview loop.
Library
All current books
01 Runtime Context, Compaction, Replay Codex Context Management Codex 上下文管理 A source-pinned technical book about how Codex governs context across turns, tools, compaction, replay, and clients. 02 OpenAI Codex CLI Codex From Source Codex 源码剖析 A source-equivalent architecture book about OpenAI Codex CLI. 03 LLM Product, Serving, Agents System Design Interview Book 系统设计面试核心书 A single-page system design interview book starting with designing ChatGPT end to end. 04 Context, Compaction, Multi-Agent Claude Code Interview Book Claude Code 面试核心书 A single-page interactive book about Claude Code context management, memory, compaction, subagents, and interview answers. 05 Logic, Probability, Construction Classic Brain Teasers 经典智力题高频题库 A single-page interactive book of classic logic, probability, coding, scheduling, and construction brain teasers. 06 Coding Practice LeetCode Practice Workbook 力扣练习题册 A filterable coding-practice workbook with Hot 100, company buckets, previews, local marks, and submit helper. 07 Suffix Order, Rank Classes, Doubling Suffix Array Primer 后缀数组入门 A compact interactive suffix-array book focused only on suffix order, rank classes, and doubling construction. 08 Adjacent Suffixes, Kasai, Reused Height LCP Array Primer LCP 数组入门 A compact interactive LCP book focused only on adjacent suffix matches and the Kasai scan.
Articles
Long notes outside the shelf
Research map AI Agent Readlist A curated route through context management, multi-agent systems, sandboxing, self-evolution, harness engineering, and evaluation. Model systems DeepSeek V4 Interview Notes A systems-oriented interview note for million-token context, hybrid attention, MoE, evaluation, serving, and cache tradeoffs.