Plan eng review

ytstack:plan-eng-review

Thin wrapper around the vendored gstack plan-eng-review skill (vendor/gstack/plan-eng-review/SKILL.md). Injects ytstack slice-plan context, then inlines the vendored procedure verbatim. Vendor is single source of truth.

ytstack context (resolved at render time)

cd "${CLAUDE_PROJECT_DIR:-$PWD}" 2>/dev/null || true
_PROJECT_SLUG=$(basename "$(git rev-parse --show-toplevel 2>/dev/null)" 2>/dev/null || basename "$(pwd)")
if [ -d "$PWD/.ytstack" ]; then _YT_DIR="$PWD/.ytstack";
elif [ -d "$HOME/.ytstack/projects/$_PROJECT_SLUG" ]; then _YT_DIR="$HOME/.ytstack/projects/$_PROJECT_SLUG";
else _YT_DIR=""; fi
_CURRENT_MILESTONE=none
if [ -n "$_YT_DIR" ] && [ -f "$_YT_DIR/STATE.md" ]; then
  _CURRENT_MILESTONE=$(sed -n 's/^current_milestone: *//p' "$_YT_DIR/STATE.md" | head -1)
fi
_SLICE_PLANS=""
if [ -n "$_CURRENT_MILESTONE" ] && [ "$_CURRENT_MILESTONE" != "none" ] && [ -d "$_YT_DIR" ]; then
  _SLICE_PLANS=$(ls "$_YT_DIR"/$_CURRENT_MILESTONE-S[0-9][0-9]-PLAN.md 2>/dev/null | tr '\n' ' ')
fi
_PITCH=""
if [ -f "./OFFICE-HOURS.md" ]; then
  _PITCH="./OFFICE-HOURS.md"
elif [ -n "$_YT_DIR" ]; then
  _PITCH=$(ls "$_YT_DIR"/OFFICE-HOURS-*.md 2>/dev/null | head -1)
fi
if [ -n "$_SLICE_PLANS" ]; then
  _MODE=milestone
elif [ -n "$_PITCH" ] && [ -f "$_PITCH" ]; then
  _MODE=concept
else
  _MODE=abort
fi
echo "HAS_YTSTACK: $([ -n "$_YT_DIR" ] && echo yes || echo no)"
echo "MODE: $_MODE"
echo "CURRENT_MILESTONE: $_CURRENT_MILESTONE"
echo "SLICE_PLANS: ${_SLICE_PLANS:-none}"
echo "PITCH: ${_PITCH:-none}"

ytstack invocation notes

Two modes, auto-detected above:

Mode = milestone (slice-plans exist). The "plan" being engineering-reviewed is the current milestone's slice-plans. Read those files when the vendored procedure asks what to review. Architecture, data flow, edge cases, test coverage, performance.

Mode = concept (no slice-plans yet; pitch artifact produced by ytstack:office-hours exists). The "plan" being engineering-reviewed is the pitch at _PITCH. Focus the review on feasibility + architectural risk of the pitched approach BEFORE any infra is scaffolded: is the target tech stack viable, are there hidden integration complexities, are the perf/scale expectations realistic, are there obvious security / data-handling gaps. Output the feasibility verdict + top 3 risk items. Concept-mode eng review is OPTIONAL per the locked greenfield flow (DECISIONS 2026-04-24); office-hours + plan-ceo-review can proceed to init-project without it.

Mode = abort. If neither slice-plans nor a pitch artifact exist, abort with: "Run ytstack:office-hours first (concept-mode review) or ytstack:slice-milestone first (milestone-mode review). plan-eng-review reviews existing plan content; it does not produce it." Do NOT invoke the vendored procedure.

After the vendored procedure completes:

• Milestone mode: apply review outcomes by editing the affected slice-plans (add edge cases, adjust tests, tighten scope). Append any locked architectural decision to DECISIONS.md in append-only format.
• Concept mode: append the feasibility verdict + risk-item list as an annotation block at the top of _PITCH. If "go", recommend ytstack:init-project. If "stop / rescope", recommend revisiting office-hours on the affected question.

Vendored procedure (inlined verbatim)

cat "${CLAUDE_SKILL_DIR}/../../vendor/gstack/plan-eng-review/SKILL.md"