bug: UnicodeDecodeError on non-English Windows due to missing encoding in read_text()

[TaeBbong]

Description

On non-English Windows systems (e.g. Korean, Japanese, Chinese), Strix crashes on startup with a UnicodeDecodeError when loading XML schema files.

Error

Error loading schema file C:\...\finish_actions_schema.xml:
'cp949' codec can't decode byte 0xe2 in position 2495: illegal multibyte sequence

Root Cause

Path.read_text() without an explicit encoding parameter uses the system default encoding, which varies by OS locale:

• Korean Windows: cp949
• Japanese Windows: shift_jis
• Western Windows: cp1252
• Linux / macOS: utf-8

Since the XML and markdown files are UTF-8 encoded, this causes decoding failures on non-UTF-8 default locales.

Affected Code

• strix/tools/registry.py — _load_xml_schema()
• strix/skills/__init__.py — load_skills()

Fix

Add encoding="utf-8" to all read_text() calls that are missing it.

PR: #301