Add AI output disclaimer to README and various export formats

[melonattacker]

This pull request adds a clear AI output disclaimer to all major user-facing outputs (Markdown, HTML, and the Web UI) to inform users that AI-generated results may contain mistakes and should be reviewed before use. The disclaimer is provided in both English and Japanese, and tests have been updated to ensure its presence in exported reports.

User-facing disclaimers

• Added a prominent AI output disclaimer (in both English and Japanese) to the Markdown export (export_md), HTML export (export_html), and the Web UI, warning users to review AI-generated results before use. [1] [2] [3] [4] [5] [6]
• Included the disclaimer in the Markdown diff export (export_diff_md).

Codebase improvements

• Centralized the disclaimer text (in English, Japanese, and Markdown-formatted) in constants.py for reuse across the codebase.
• Updated imports in relevant modules to use the new constants. [1] [2] [3]

Testing

• Updated and added tests to ensure the disclaimer is present in Markdown and HTML exports, for both empty and non-empty threat reports, and in diff reports. [1] [2] [3] [4] [5]