feat(phase-2): add output format selection and auto mode

[overtrue]

Summary

• add global --format auto|human|json selection while keeping the legacy --json flag
• introduce terminal-aware output resolution that disables colors and progress automatically in non-TTY contexts
• make the new noun-first bucket and object command groups default to auto output selection while legacy commands keep human-readable defaults

Validation

• cargo fmt --all
• cargo clippy --workspace -- -D warnings
• cargo test --workspace

Notes

• bucket and object now default to structured output when stdout is not a TTY
• legacy flat commands continue to default to human-readable output unless --json or --format is provided
• raw content commands such as cat and head keep their existing stdout semantics

Next phase

• enrich machine-readable error output with typed fields and repair suggestions
• add example-first help for the highest-value mutating commands
• extend dry-run and idempotent/upsert-style flows for bucket mutations

[Copilot]

Pull request overview

Adds a global output-format selection mechanism to the Rust CLI, introducing an auto mode that adapts output behavior to TTY/non-TTY environments while preserving legacy --json usage and maintaining existing semantics for raw-content commands.

Changes:

• Introduces global --format auto|human|json and centralizes output resolution (TTY-aware color/progress + JSON selection).
• Updates command execution wiring so noun-first bucket/object default to auto/structured behavior, while legacy flat commands retain human defaults.
• Extends the CLI help contract tests to require discoverability of the new global option.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
crates/cli/src/commands/mod.rs	Adds OutputFormat + TTY-aware resolution and applies per-command default output behavior during dispatch.
crates/cli/tests/help_contract.rs	Ensures --format appears as a discoverable global option across --help outputs.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

--json and --format can both be provided, and --json silently wins (selected_format forces Json when self.json is true). This is ambiguous UX and can mask mistakes in scripts/wrappers. Consider making them mutually exclusive via clap (conflicts_with = "format" / overrides_with = "format"), and/or update the --json help text to explicitly state it is a legacy alias that overrides --format.

Suggested change

	/// Output format: human-readable or JSON
	#[arg(long, global = true, default_value = "false")]
	/// Legacy alias for `--format json` (mutually exclusive with `--format`)
	#[arg(long, global = true, default_value = "false", conflicts_with = "format")]

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 08f8d42405

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Avoid auto-enabling JSON for non-JSON object dry-runs

Routing all object subcommands except show/head to StructuredDefault makes json=true whenever stdout is non-TTY, but several of those commands still print human dry-run text (for example, cp/mv/rm dry-run paths use formatter.println("Would ...") and return). In piped/CI usage, rc object copy|move|remove --dry-run now emits non-JSON output while auto mode has selected JSON, which breaks machine parsing and contradicts the new output-format contract.

Useful? React with 👍 / 👎.