CLI Design for Low-Error Operations

Goal

I design CLI tools for predictable behavior, not clever interfaces.

Principles

1. Every mutating command has a dry-run mode.
2. Output is structured and stable enough for scripts.
3. Validation happens before side effects.
4. Errors are explicit and mapped to actionable fixes.

Example Pattern

For bookkeeping-style operations, a command should confirm what it parsed and what it will write before writing anything.

Outcome

The tool becomes safer for daily use and easier to integrate into CI or scheduled tasks.