Refine todo list tag parsing and apply tag-first filtering

[multiplex55]

Motivation

• Separate tag tokens from text tokens in todo list queries so tag filters are applied deterministically.
• Allow explicit exclusion of tags (prefixed with !@ or !#) and preserve existing ! negation semantics for text matching.
• Ensure tag filtering runs before fuzzy text matching and that priority sorting is preserved when tag filters are used.
• Add a focused unit test to exercise tag combinations and negated text behavior.

Description

• Updated TodoPlugin::search_internal parsing so each token after todo list is split on whitespace and categorized as a requested tag, excluded tag, or text token while preserving single-! negation for text via strip_prefix('!').
• Applied tag filtering first by retaining entries that contain all requested_tags (case-insensitive) and removing any that contain any excluded_tags (case-insensitive).
• Kept the existing fuzzy text filtering (using the joined text_tokens) and preserved priority sorting when tag filters are active or when there is no text filter.
• Added a unit test list_filters_by_tags_and_text in src/plugins/todo.rs that creates sample entries and verifies todo list @testing, todo list @testing @ui, and todo list !foo @testing behaviors.

Testing

• Added the automated unit test list_filters_by_tags_and_text in src/plugins/todo.rs which asserts tag-only filtering, combined-tag (AND) behavior, and negated text after tag filtering.
• No automated test suite was executed as part of this change (tests were added but not run).

---

Codex Task