fix(fetchers): avoid utf-8 panic in hn html stripping

[chaliy]

Motivation

• Prevent a panic when stripping HTML from untrusted Hacker News item/comment text by avoiding slicing at non-char UTF-8 boundaries.

Description

• Iterate the input with html.char_indices() and compute the lookahead slice using idx + c.len_utf8() instead of result.len(), and add a regression assertion strip_html_tags("ab<é>xy<") == "abxy".

Testing

• Added a focused unit assertion in test_strip_html_tags and ran cargo test -p fetchkit hackernews::tests::test_strip_html_tags, which passed.

---

Codex Task