Escape user content in notebook HTML

[njzjz-bot]

Summary

• HTML-escape rendered JSON keys and values
• escape generated tooltip text before restoring supported formatting
• add injection regression coverage for keys, values, and docs

Fixes #120.

Validation

• python -m unittest -v tests.test_notebook with IPython display imports available
• Ruff checks on changed files

AI attribution

Coding agent: Codex
Codex version: codex-cli 0.149.0
Model: gpt-5.6-sol
Reasoning effort: xhigh

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 63e56341-0e70-4b98-a7bd-5477877eb50b

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[njzjz-bot]

Review result: one medium-severity issue found.

The user-content escaping closes the intended injection paths, but the escaping order introduces a tooltip-formatting regression. See the inline suggestion.

I reproduced the literal <br/> output with the PR head; the security-focused tests and CI otherwise pass.

Coding agent: Codex
Codex version: codex-cli 0.149.0
Model: gpt-5.6-sol
Reasoning effort: xhigh

[njzjz-bot]

[P2] Escape before inserting the trusted line-break markup

doc_head has already converted newlines into <br/> before this call, so html.escape() turns every intended tooltip line break into visible <br/> text. I reproduced output such as child: <br/> type: .... Recompute from the unformatted generated text, escape it, and only then add renderer-owned line breaks:

Suggested change

	doc_head = html.escape(doc_head, quote=False)
	doc_head = html.escape(
	self.arg.gen_doc_head().replace("| type:", "type:"),
	quote=False,
	).replace("\n", linebreak)

After applying this suggestion, the earlier doc_head assignment at lines 274–278 is redundant and should be removed.