Summary

This adds focused parser coverage for the noun-first remove entrypoints introduced by the recent bucket/object command-group work.

The command-group PR added help-contract coverage and a few parser assertions around bucket cors, but the remove paths still had no direct parser tests. That left a small regression window where clap wiring for rc bucket remove or rc object remove could drift without a fast unit test catching it.

User impact

Users now have explicit regression coverage for two public noun-first removal entrypoints:

• rc bucket remove local/my-bucket
• rc object remove local/my-bucket/report.csv --dry-run

If either path stops parsing correctly in the future, the unit suite will fail immediately instead of relying on broader help output or runtime behavior to detect the breakage.

Root cause

The noun-first command groups were added in #76, but parser tests in crates/cli/src/commands/mod.rs only exercised bucket cors routing. The remove subcommands delegated into existing argument types without any direct assertion that their positional arguments and flags survived clap parsing.

Fix

This PR adds two narrow parser tests in crates/cli/src/commands/mod.rs:

• one verifies rc bucket remove reaches bucket::BucketCommands::Remove with the expected target;
• one verifies rc object remove reaches object::ObjectCommands::Remove, preserves the path list, and keeps --dry-run enabled.

The scope stays limited to the recently changed noun-first command surface and does not modify production code.

Validation

I ran the following successfully:

• cargo test -p rustfs-cli cli_accepts_bucket_remove_subcommand --lib -- --nocapture
• cargo test -p rustfs-cli cli_accepts_object_remove_subcommand --lib -- --nocapture
• cargo fmt --all --check
• cargo clippy --workspace -- -D warnings
• cargo test --workspace

I also ran make pre-commit per the automation instructions, but this checkout does not define that target and exits with:

make: *** No rule to make target `pre-commit'.  Stop.