cac-apply

Apply Suggestion — Multi-Agent Refactoring Orchestrator

Orchestrates the cac-executor and cac-qa agents to reliably apply pending suggestions. Reads suggestions, spawns executor for changes, spawns QA for independent verification, iterates on failure (max 3 cycles), rolls back on persistent failure.

Arguments

• No argument: list all pending items, ask user to select
• {timestamp}: apply specific item by timestamp ID (e.g., /cac-apply 20260323-143052)
• all: ask user "몇 개를 적용하시겠습니까?" then apply that many sequentially

---

Phase 0: Safety Pre-checks

Before doing ANYTHING, verify the environment is safe for refactoring.

0a. Dirty tree guard

DIRTY=$(git status --porcelain 2>/dev/null | head -1)

If output is non-empty:

• Report: "uncommitted changes가 있습니다. 리팩토링 전에 커밋 또는 stash 하세요."
• Ask user: A) git stash 후 진행 / B) 현재 상태에서 진행 (위험) / C) 취소
• If B: warn "테스트 실패 시 롤백이 다른 변경까지 되돌릴 수 있습니다" and proceed
• If C: stop

0b. Test runner detection

# Detect test runner from any supported manifest
HAS_TESTS="NO_TESTS"
[ -f package.json ] && grep -q '"test"' package.json && HAS_TESTS="HAS_TESTS"
[ -f Makefile ] && grep -q '^test:' Makefile && HAS_TESTS="HAS_TESTS"
[ -f Cargo.toml ] && HAS_TESTS="HAS_TESTS"
[ -f go.mod ] && HAS_TESTS="HAS_TESTS"
[ -f pyproject.toml ] && HAS_TESTS="HAS_TESTS"
[ -f setup.py ] && HAS_TESTS="HAS_TESTS"
[ -f Gemfile ] && HAS_TESTS="HAS_TESTS"
echo "$HAS_TESTS"

Record HAS_TESTS or NO_TESTS. If tests exist, verify they pass BEFORE any changes:

# Run pre-change tests using detected manifest
if [ -f package.json ] && grep -q '"test"' package.json; then npm test 2>&1
elif [ -f Cargo.toml ]; then cargo test 2>&1
elif [ -f go.mod ]; then go test ./... 2>&1
elif [ -f pyproject.toml ]; then pytest 2>&1
elif [ -f Makefile ] && grep -q '^test:' Makefile; then make test 2>&1
elif [ -f Gemfile ]; then bundle exec rake test 2>&1
fi

If tests already fail: report "기존 테스트가 이미 실패합니다." Ask user whether to proceed. Record failing test names as PRE_EXISTING_FAILURES.