mirror of
https://github.com/magnus919/agent-skills.git
synced 2026-09-16 05:56:30 +03:00
765 B
765 B
Agent-Readiness Checklist
- No interactive prompts (
read,select,dialog) - All inputs arrive through flags or environment variables
- Every subcommand has
--helpexamples --jsonoutput parses as JSON- Every destructive operation supports
--dry-run --forceor--yesskips confirmation- Repeating an operation produces a no-op rather than an error
- Commands use a consistent
resource verbstructure - Errors go to stderr
- JSON mode emits no non-JSON stdout
--helpand--dry-runwork without credentials- Exit codes distinguish success, usage errors, and runtime failures
- Each read endpoint has a live-server verification
- Each chained API path has a dry-run verification