claude-code/support-automation

support automation Shipped

What this lens looks for

Skills and agents should expose capabilities through scriptable interfaces — not just interactive use; design operations as discrete, composable commands that can be invoked programmatically; provide non-interactive entry points (shell scripts, CLI flags, batch modes) so workflows can drive the extension without human intervention

What its verifier checks

Key operations callable non-interactively (no required interactive prompts for automation paths); shell scripts wrap deterministic operations for direct invocation; skills can be chained without requiring human confirmation at each step; outputs are machine-parseable (not just human-readable)