[Bug] extract-content.ts crashes on Windows: fs.cpSync aborts tsx with 0xC0000409

Problem

npm run dev in web/ never starts on Windows. The predev hook runs npm run extract, which uses tsx scripts/extract-content.ts. That script aborts partway through:

> web@0.1.0 extract
> tsx scripts/extract-content.ts

Extracting course content...
  Repo root: <repo root>
  Source: root chapter folders (17)

...and then the process dies. No stack trace, no error message — just the shell prompt coming back with a non-zero exit code.

Root cause

copyChapterAssets() calls fs.cpSync(imagesDir, outDir, { recursive: true }):

https://github.com/shareAI-lab/learn-claude-code/blob/main/web/scripts/extract-content.ts#L218

On Windows with Node 22, fs.cpSync terminates the process at the native layer:

Because the failure is native, there is no stack trace to follow. This is why it looks like the script "hangs after Source: root chapter folders (17)".

Minimal reproduction

Reproducible outside this repo — copying a directory that contains a single one-line text file is enough:

const fs = require("fs");
fs.mkdirSync("src", { recursive: true });
fs.writeFileSync("src/a.txt", "x");
fs.cpSync("src", "dst", { recursive: true });  // process dies, exit 0xC0000409

$ node repro.js
$ echo $?
3221226505

Environment

Impact

The extraction produces no web/public/course-assets/ output and no generated course data, so the web/ app cannot be built or run locally on Windows.

Possible fix

fs.cpSync is only used in this one place. Replacing it with a small recursive copy helper (fs.readdirSync + fs.mkdirSync + fs.copyFileSync) avoids the crashing native code path and keeps the behaviour identical:

before: exit 0xC0000409, extraction aborted, no course-assets output
after:  exit 0, "17 versions / 16 diffs / 51 docs", 74 asset files written

I have a patch ready and will open a PR that references this issue.

Note: opened with AI assistance for research/reproduction; I ran the reproduction and the fix on my own machine and stand behind the report.

> web@0.1.0 extract
> tsx scripts/extract-content.ts

Extracting course content...
  Repo root: <仓库根目录>
  Source: root chapter folders (17)

const fs = require("fs");
fs.mkdirSync("src", { recursive: true });
fs.writeFileSync("src/a.txt", "x");
fs.cpSync("src", "dst", { recursive: true });  // 进程死亡，退出码 0xC0000409

$ node repro.js
$ echo $?
3221226505

修复前：退出码 0xC0000409，提取中断，无 course-assets 产物
修复后：退出码 0，17 versions / 16 diffs / 51 docs，生成 74 个 asset 文件

Start in web/scripts/extract-content.ts at copyChapterAssets(), then run npm run extract from web on Windows with Node 22. Replace the failing recursive copy path while preserving the existing extraction behavior. Done means the command exits successfully and produces the reported course data and 74 asset files, allowing npm run dev to start.