feat(cli): add detach option to orx up and add orx down command

[jyuan-dev]

Summary

This PR adds two complementary capabilities for managing the orx up dashboard server:

1. Detached mode (-d, --detach) on orx up: Starts the dashboard server in the background as a detached process (using setsid), redirecting stdio to ~/.local/share/openresearch/orx-up.log, verifying health, and exiting immediately. This allows users to disconnect their terminal or SSH session while keeping the server running.
2. orx down command: Stops running orx up dashboard instances regardless of PID. It dynamically detects instances via the local orx-up.json pidfile, loopback health endpoint probe, or process table scan, shutting them down gracefully through /api/down and escalating to SIGTERM/SIGKILL if needed.

Changes

• src/main.rs:
  • Added -d, --detach option to UpArgs.
  • Added Command::Down(DownArgs) subcommand with --port, --remote, -f/--force, and --all flags.
• src/commands/up.rs:
  • Handled args.detach by spawning a detached process with libc::setsid() and checking /api/health.
  • Added UpPidFile guard to record {pid, port} on startup.
  • Added routes POST /api/down and POST /api/shutdown for graceful shutdown.
  • Added "pid": std::process::id() to the /api/health response.
• src/commands/down.rs:
  • Implemented dynamic target discovery (pidfile, health probing, /proc / ps process table inspection).
  • Implemented graceful shutdown with escalation.
  • Added remote delegation support via --remote <HOST>.
• src/commands/up_remote.rs:
  • Exposed parse_remote_target for orx down --remote.

Test Plan

• Unit tests: cargo test --bin orx cli_tests && cargo test --bin orx down::tests (all passed).
• Integration test:
  • orx up --port 5899 -d starts the dashboard in the background and exits 0 immediately.
  • Verified curl http://127.0.0.1:5899/api/health returns ok.
  • orx down --port 5899 terminates the background server cleanly.
• Verified cargo fmt --all --check and cargo clippy --all-targets -- -D warnings.