fix(ui): make character accent color stable across process restarts

[AreteDriver]

Bug

`character_accent_color()` used Python's built-in `hash()` to map names to palette indices. `PYTHONHASHSEED` is randomized by default, so the same character renders in a different color every app launch. The intra-process `test_deterministic` test in `test_status_dock.py` never caught it because both calls happened in one process.

This is debt item #1 from the post-v3.2.0 audit.

Fix

Switch to MD5 (content-addressed). Same character → same color every launch, every machine, every Python build. The hash isn't used for security; `usedforsecurity=False` quiets bandit and FIPS environments.

```python
digest = hashlib.md5(name.encode("utf-8"), usedforsecurity=False).digest()
index = digest[0] % len(CHARACTER_ACCENT_COLORS)
```

Tests

• `test_cross_process_determinism` — runs the lookup in two subprocesses with different `PYTHONHASHSEED` values and asserts the result is identical. Would have caught the original bug.
• `test_uniform_distribution_over_palette` — 200 distinct names hit at least 6 of 8 palette entries.

Test plan

• 2 new tests, both pass
• All 5 `TestCharacterAccentColor` tests pass (3 existing + 2 new)
• Full suite: 2400 passed, 5 skipped, 0 failures
• `ruff check` + `ruff format --check` clean

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[gemini-code-assist]

Code Review

This pull request transitions the character accent color logic from hash() to hashlib.md5() to achieve cross-process determinism, addressing issues where colors changed between app launches. It also introduces unit tests to verify this determinism and ensure uniform color distribution. Feedback highlights a potential TypeError on Python versions below 3.9 due to the usedforsecurity argument and suggests refactoring another component for consistency.

[gemini-code-assist]

The usedforsecurity keyword argument was introduced in Python 3.9. Using it directly will cause a TypeError on older Python versions (such as 3.8), which are still common in many environments. Since this is a UI utility, consider adding a fallback for compatibility.

Additionally, please note that ArrangementGrid.add_character (line 430) still uses the unstable hash() function and a local color palette. It should be refactored to use this character_accent_color helper to ensure that character colors are stable and consistent across all UI components.

Suggested change

	digest = hashlib.md5(name.encode("utf-8"), usedforsecurity=False).digest()
	index = digest[0] % len(CHARACTER_ACCENT_COLORS)
	r, g, b = CHARACTER_ACCENT_COLORS[index]
	try:
	digest = hashlib.md5(name.encode("utf-8"), usedforsecurity=False).digest()
	except TypeError:
	# Fallback for Python < 3.9
	digest = hashlib.md5(name.encode("utf-8")).digest()
	index = digest[0] % len(CHARACTER_ACCENT_COLORS)
	r, g, b = CHARACTER_ACCENT_COLORS[index]