efficiency-reviewlisted

# Efficiency Review A specialist independent reviewer focused exclusively on performance and resource cost. Spawned by `/ship-feature` when the efficiency gate is `required` in `gates.manifest.json`. The default `/independent-review` skill checks "no N+1, no unbounded ops" as one bullet. That's not enough. This skill compares the implementation against the **Efficiency budget** the plan declared, demands measurement evidence, and flags missing baselines. ## When to use - `/ship-feature` invokes this when the efficiency gate triggers (per `ai/gates.config.mjs`). - Standalone: when investigating "the dashboard got slower this week" — point this at the suspect PR. - Before merging any change to a known hot path. ## What the reviewer gets - `git diff origin/main...HEAD` — the changes - `ai/runs/<run>/plan.md` — declared Efficiency budget - `ai/runs/<run>/efficiency-evidence.md` if it exists — implementer's measurements - `ai/runs/<run>/worklog.md` — measurements may be embedded here instead - `ai/STANDARDS.md` - `ai/checklists/efficiency.md` ## What the reviewer does NOT get - `review.md` — implementer's conclusions - Conversation history (Note: the worklog IS given for this reviewer because measurements often live there. The implementer's *interpretive* narrative is filtered out by only consuming the measurement-relevant sections.) ## Step 1 — Gather inputs ```bash RUN_DIR="${RUN_DIR:-$(ls -dt ai/runs/*/ | head -1)}" DIFF=$(git diff origin/main...HEAD) PLAN=$(cat "$RU