rosa-command-authoring

# ROSA Command Authoring Use this skill when: - Adding a new `rosa` command or subcommand - Changing flags or command help - Moving logic between `cmd/` and `pkg/` - Refactoring command execution flow ## Workflow 1. Read `AGENTS.md` and `guidelines/command-guidelines.md`, then inspect the nearest similar command implementation. 2. Keep Cobra command files thin and move non-Cobra logic into `pkg/`. 3. Follow the entrypoint and exit pattern already established in the nearest similar command area. 4. Many ROSA commands use `Run: run`; do not switch a command area between `Run` and `RunE`, or add/remove direct `os.Exit()` calls, unless the surrounding pattern already does so and the change keeps behavior consistent. 5. Reuse `output`, `reporter`, and `interactive` patterns already used by the surrounding command area. 6. If the command tree changes, update `cmd/rosa/structure_test/command_structure.yml`. 7. If supported flags change, update the matching `cmd/rosa/structure_test/command_args/**/command_args.yml`. 8. When command help or docs change, check whether `make generate-docs` is part of the required verification. ## Verification - `make fmt` - relevant package tests or `make test` - `make rosa` - `make generate-docs` when command docs or help output changed Follow `CONTRIBUTING.md` for the exact contributor workflow and hook expectations.