Escape command name in _remove_command regex

[Nas01010101]

Problem

_remove_command interpolates the user-supplied command directly into a regex base_pattern:

base_pattern = (
    r'\\'
    + command
    + r'(?:\[(?:.*?)\])*\{((?:[^{}]+|\{(?1)\})*)\}(?:\[(?:.*?)\])*'
)

So a command name containing regex metacharacters is matched as a pattern rather than literally. For example, with commands_to_delete=['todo.']:

• . matches any character, so \todoX{...} is wrongly removed alongside \todo.{...} (over-removal).
• Quantifiers like */+ in a name make the pattern match the wrong span (or fail to match), so the targeted command may not be removed at all (under-removal).

Fix

Wrap the name with regex.escape() so it is matched literally:

    + regex.escape(command)

This is consistent with how the rest of this module already builds patterns from
dynamic strings (e.g. the filename handling in _replace_includegraphics, which
uses regex.escape in several places).

commands_to_delete / commands_only_to_delete are documented as literal command
names (CLI --help and cleaner_config.yaml), so literal matching is the intended
behavior; this does not change handling of normal alphabetic command names.

Tests

Adds a parameterized regression test (test_remove_command_escapes_regex_metacharacters_in_name)
covering ., *, and + in command names. Each case asserts the look-alike command is
preserved and the targeted command is removed. The cases fail before this change and pass after;
the full unit-test suite remains green.