Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Backend engineering review — walks the 7 Matt Pocock forcing questions (read/write ratio + QPS, tenancy, sync vs async, data sensitivity, pattern, RPO/RTO, SLO), picks the language + pattern profile, forks into specialists (api-design-reviewer, database-designer, migration-architect, slo-architect). Invokes the cs-backend-engineer agent with context fork.
.claude/skills/alirezarezvani-csbackend-review-backend-engineering-review/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 144% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -2% | 0% |
Use the cs-backend-engineer agent (uses context: fork) to handle this inquiry:
$ARGUMENTS
Canonical source: engineering-team/skills/senior-backend/references/forcing_questions.md (7 questions, one-per-turn, recommendation + canon citation per question).
engineering-team/skills/senior-backend/references/forcing_questions.md. One per turn. Recommend with cited canon. Track in /tmp/backend-grill-<date>.md.bash python engineering-team/skills/senior-backend/scripts/backend_decision_engine.py \ --team-size <N> --qps-p99 <N> --read-write-ratio <ratio> \ --tenancy <single-tenant|shared-multi-tenant|isolated-multi-tenant> \ --data-sensitivity <public|pii|phi|pci> \ --pattern <monolith|modular-monolith|domain-bounded-services|microservices|serverless> \ --language-preference <typescript|python|go|rust|java|kotlin|dotnet>
slo-architect FIRST — no SLO, no designapi-design-reviewer — API contractdatabase-designer + database-schema-designer — schema + ERDmigration-architect — only if changing existing schemaobservability-designer — golden signals + alertsci-cd-pipeline-builder — pipeline matching cadence targetsenior-security + adversarial-reviewer — before public launchra-qm-team/* — if data sensitivity is PHI / PCI / regulatedcs-karpathy-reviewer — before any commitapi-design-reviewer.Profiles live at engineering-team/skills/senior-backend/profiles/. Four built-in: node-express, fastapi-python, django-monolith, go-or-rust-microservice. Copy one to <your-org>.json and adjust constraints / SLO floor / approver chain.
/cs:fullstack-review — full-stack lens (parent)/cs:frontend-review — for API consumer side/cs:engineer-grill — cross-role 21-question grill/slo-design — explicit SLO design via slo-architect/karpathy-check — Karpathy 4-principle review| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 11,858 | 4,085 | -66% | 1 | 1 | 0% | 2,396 | 1,688 | -30% | 0 | 0 | — |
case-02 | fail→pass | 10,312 | 6,285 | -39% | 1 | 1 | 0% | 1,613 | 1,725 | +7% | 0 | 0 | — |
case-03 | fail→pass | 5,214 | 5,163 | -1% | 1 | 1 | 0% | 779 | 1,904 | +144% | 0 | 0 | — |
case-04 | pass→pass | 7,140 | 5,843 | -18% | 1 | 1 | 0% | 1,404 | 2,099 | +50% | 0 | 0 | — |
case-05 | pass→pass | 9,545 | 7,210 | -24% | 1 | 1 | 0% | 1,440 | 2,274 | +58% | 0 | 0 | — |
case-06 | pass→pass | 9,980 | 12,277 | +23% | 1 | 1 | 0% | 1,920 | 3,353 | +75% | 0 | 0 | — |
case-07 | pass→pass | 19,178 | 9,733 | -49% | 1 | 1 | 0% | 2,269 | 2,681 | +18% | 0 | 0 | — |
case-08 | pass→pass | 18,810 | 10,009 | -47% | 1 | 1 | 0% | 2,335 | 2,390 | +2% | 0 | 0 | — |
case-09 | fail→pass | 15,777 | 8,985 | -43% | 1 | 1 | 0% | 2,127 | 2,515 | +18% | 0 | 0 | — |
case-10 | fail→fail | 9,910 | 5,879 | -41% | 1 | 1 | 0% | 1,580 | 1,953 | +24% | 0 | 0 | — |
case-11 | fail→pass | 12,687 | 6,029 | -52% | 1 | 1 | 0% | 2,072 | 2,029 | -2% | 0 | 0 | — |
case-12 | fail→pass | 12,461 | 11,314 | -9% | 1 | 1 | 0% | 1,882 | 2,570 | +37% | 0 | 0 | — |
case-13 | fail→pass | 9,982 | 5,970 | -40% | 1 | 1 | 0% | 1,809 | 2,047 | +13% | 0 | 0 | — |
case-14 | fail→pass | 8,803 | 4,232 | -52% | 1 | 1 | 0% | 1,525 | 1,685 | +10% | 0 | 0 | — |
case-15 | fail→fail | 8,250 | 4,538 | -45% | 1 | 1 | 0% | 1,677 | 1,998 | +19% | 0 | 0 | — |
case-16 | fail→pass | 9,430 | 4,296 | -54% | 1 | 1 | 0% | 1,311 | 1,677 | +28% | 0 | 0 | — |
case-17 | fail→pass | 8,061 | 6,857 | -15% | 1 | 1 | 0% | 1,361 | 2,458 | +81% | 0 | 0 | — |
case-18 | fail→pass | 8,762 | 2,396 | -73% | 1 | 1 | 0% | 1,256 | 1,391 | +11% | 0 | 0 | — |
case-19 | fail→pass | 6,391 | 4,406 | -31% | 1 | 1 | 0% | 1,088 | 1,490 | +37% | 0 | 0 | — |
case-20 | pass→pass | 3,959 | 2,120 | -46% | 1 | 1 | 0% | 639 | 1,349 | +111% | 0 | 0 | — |
case-21 | fail→pass | 15,544 | 2,402 | -85% | 1 | 1 | 0% | 1,226 | 1,347 | +10% | 0 | 0 | — |
case-22 | fail→pass | 12,387 | 2,223 | -82% | 1 | 1 | 0% | 2,116 | 1,313 | -38% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +64 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.