fix(jira): preserve list markers before emphasis conversion

[GuddXzy]

Description

Jira list markers are interpreted as emphasis delimiters before list conversion, corrupting bold text and inflating nesting depth. For example, ** Level two. becomes a list item indented by six spaces instead of two.

Fixes #1651.

Changes

• Run the existing list-marker conversion before emphasis conversion.
• Add eight regression cases covering bold text, literal asterisks, nested bullets, and mixed ordered/unordered markers.

Testing

• Unit tests added: all eight new cases fail before the fix and pass afterward.
• Full preprocessing unit suite: 181 passed with no skips.
• Full unit suite: 3,903 passed; 5 pre-existing environment-gated tests skipped.
• Non-live integration smoke suite: 8 passed; live-service cases remained behind their explicit suite gates.
• Ruff lint/format and mypy checks pass for both changed files.
• Generated tool documentation is current (98 tools, 25 toolsets).
• git diff --check passes.
• GitHub checks pass on Python 3.10 through 3.13.

No live Jira or Confluence testing was performed. No tool signatures or generated documentation changed.

[mcp-atlassian-maintainer]

@GuddXzy, moving list-marker conversion ahead of emphasis parsing fixes the corruption cleanly, and the 8-case regression matrix covers the tricky nested and mixed-marker cases well. Thanks for carrying this through.