Add centralized argument parsing and usage responses for bright and todo commands

[multiplex55]

Motivation

• Provide a centralized, non-panicking argument parsing helper that returns either parsed values or a usage string so command plugins can present helpful guidance on empty/invalid input.
• Ensure commands expose a safe zero-arg path (open dialog/view) while also surfacing usage text for partial or malformed queries.

Description

• Add src/common/command.rs with parse_args and ParseArgsResult to return Parsed(T) or Usage(String) and expose it via src/common/mod.rs as command.
• Update the brightness plugin to use parse_args and return a usage Action (label starts with Usage: bright) for partial/invalid inputs while still returning the dialog action for empty bright input.
• Update the todo plugin to use parse_args for add, pset, tag, rm and to return usage Actions (labels like Usage: todo ...) for empty/invalid input while retaining zero-arg dialog/view actions.
• Add/adjust unit tests in tests/brightness_plugin.rs and tests/todo_plugin.rs to validate usage responses and fuzz-ish partial queries such as "br", "bright", and "todo !" do not panic and surface helpful usage output.

Testing

• Ran cargo test --test brightness_plugin --test todo_plugin to execute the modified tests, but the build failed with a dependency build error caused by a missing system library (alsa/alsa-sys), so the test run did not complete successfully.
• cargo fmt was run as part of the rollout to format changes (no test assertion impact).

---

Codex Task