Pre-computed context

Automation inventory: !bash "${CLAUDE_PLUGIN_ROOT}/skills/audit-automation-gaps/scripts/inventory.sh" 2>/dev/null || echo "inventory unavailable"

Purpose

Audit the repo's automation landscape and identify genuine gaps — not surface-level "you don't have X" observations, but evidence-backed gaps where no mechanism in the enforcement hierarchy covers the concern.

Design principle: most "missing automation" gaps are already covered by higher enforcement levels (compiler, analyzers, build-time checks, architecture tests, behavioral rules). This skill's job is to prove a gap exists before recommending a solution — the default verdict is REJECT, not PASS.

The enforcement hierarchy (strongest first): compiler settings → static analyzers/linters → architecture tests → unit/integration tests → git hooks → Claude Code hooks → code review → documentation/behavioral rules. A consuming repo that documents its own hierarchy in CLAUDE.md or rules files overrides this default ordering — read and use theirs.