fix: escape bare < in generated MDX to fix Vercel doc builds

[tyler-griggs]

Summary

• The uv.lock regeneration in [megatron] support qwen3.5 models for megatron, bump mbridge + megatron-core to latest #1425 bumped griffe2md from 1.3.4 to 1.5.0, which renders docstring content slightly differently
• A <= operator at the start of a line in the FullyAsyncConfig docstring (skyrl/train/config/config.py:394) was left unescaped in the MDX output, causing the Next.js MDX parser to fail with: Unexpected character '=' (U+003D) before name
• Adds a second regex pass in sanitize_for_mdx() that escapes any bare < not followed by a valid HTML tag-start character (letter, /, !)

---

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 1 additional finding.

[gemini-code-assist]

Code Review

This pull request updates the docs/generate-api-docs.py script to escape bare < characters in docstrings, preventing MDX parsing errors for sequences like <= or <2. Feedback was provided to compile the regular expression outside the loop to improve performance and readability.

[gemini-code-assist]

For improved performance and readability, it's a best practice to compile regular expressions that are used inside a loop. This regex is recompiled on every iteration within the loop over parts of a line.

You could define the compiled regex object once, before the for line in lines: loop, and reuse it. For example:

# Before the loop
_bare_lt_re = re.compile(r"<(?![a-zA-Z/!])")

# Inside the loop
...
part = _bare_lt_re.sub(r"&lt;", part)

This advice also applies to the other regular expression substitution for invalid tags just above this change.