Rust CLI and Daemon Work

Use this when the program is a process boundary first: command-line tool, worker, daemon, long-running job, or supervisor-managed service.

Working stance

• Keep parsing, config loading, and reporting at the edge.
• Make process lifetime explicit: startup, steady state, shutdown, exit code.
• Write machine-readable output to stdout and diagnostics to stderr.
• Long-running jobs need supervision, cancellation, and cleanup contracts.
• Pair this skill with rust-errors; add rust-async-and-concurrency only if lifecycle or parallelism is the real pressure.

Domain pressure

• Config precedence should be stable and unsurprising.
• Exit codes are part of the interface.
• Background workers must decide who owns retries, backoff, and shutdown.
• Progress reporting should never corrupt data output.

Red flags

• main contains business logic and retry loops,
• shutdown depends on ad hoc signal handling spread across modules,
• human output and machine output share the same channel,
• daemon state is shared globally because ownership was never assigned.