escape regex metachars in StringDetector word matching

[bereketDeneke]

What this change does

StringDetector's "word" match mode interpolated each search string directly into a regex (r"\b" + s + r"\b"). When a target string contained regex metacharacters, it was interpreted as a pattern rather than a literal — a real problem because several built-in wordlists (e.g. the Surge profanity list) contain entries like c*nt, s.o.b., and pu$sy.

This caused two classes of bugs:

• False positives — e.g. s.o.b. matching benign words like snobby (. matches any char).
• False negatives — the literal obfuscated term never matching itself, e.g. pu$sy ($ acts as an anchor) or c*nt (* makes the preceding c optional).

The fix wraps the search string in re.escape() so metacharacters are matched literally:

if re.search(r"\b" + re.escape(s) + r"\b", output_text):

Verification

• Added test_stringdetector_word_regex_metacharacters covering literal matching of s.o.b., pu$sy, and c*nt, including the false-positive/false-negative cases above.
• Run the tests and ensure they pass python -m pytest tests/detectors/test_detectors_base.py

[leondz]

please fill in the PR description. marked as draft until ready

[bereketDeneke]

please fill in the PR description. marked as draft until ready
PR description is filled-in!