Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build an onboarding and training guide for restaurant front-of-house staff (servers, hosts, bartenders). Use when asked to train a new server, create FOH onboarding, write service standards, or build a restaurant training program. Produces a phased training plan (shadow → hands-on → solo with support), the service-sequence standards, menu and allergen knowledge checks, POS and side-work basics, and a sign-off checklist that says when someone's ready to work a section alone.
.claude/skills/mohitagw15856-server-training-guide/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 7% | 0% |
Most restaurants "train" by throwing a new hire on the floor with a tray and hoping. The result is inconsistent service and quick turnover. This skill builds a real ramp: a new server learns the steps of service, the menu, and the systems in a sequence, with a clear bar for when they're ready to run a section solo.
Given the restaurant type, produce the full guide — infer the service style (fine dining, casual, fast-casual, bar) and scale the depth and standards to it.
Ask for (if not provided, else infer and label):
A staged ramp with what happens and who owns it:
The steps of service for this restaurant (greet time, drink/order timing, check-backs, pre-bussing, the close), with the specific standards (e.g. "greet within 2 minutes").
Menu and allergen quiz points, upsell/pairing basics, and the "86 / substitution / allergy" protocol every server must know cold.
POS order flow, payment and comps/voids policy, opening/closing side-work, and section/station setup.
The observable competencies a trainer checks before solo — greeting, order accuracy, timing, allergen handling, POS, closing.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 60,477 | 44,146 | -27% | 1 | 1 | 0% | 8,285 | 5,847 | -29% | 0 | 0 | — |
case-02 | fail→pass | 58,233 | 42,334 | -27% | 1 | 1 | 0% | 7,539 | 5,814 | -23% | 0 | 0 | — |
case-03 | pass→pass | 55,578 | 51,864 | -7% | 1 | 1 | 0% | 8,267 | 6,614 | -20% | 0 | 0 | — |
case-04 | fail→pass | 40,827 | 31,546 | -23% | 1 | 1 | 0% | 5,548 | 4,501 | -19% | 0 | 0 | — |
case-05 | fail→pass | 54,619 | 37,306 | -32% | 1 | 1 | 0% | 6,335 | 5,637 | -11% | 0 | 0 | — |
case-06 | fail→pass | 43,589 | 43,791 | +0% | 1 | 1 | 0% | 5,831 | 6,264 | +7% | 0 | 0 | — |
case-07 | fail→pass | 58,180 | 49,087 | -16% | 1 | 1 | 0% | 8,254 | 7,412 | -10% | 0 | 0 | — |
case-08 | fail→pass | 61,864 | 59,223 | -4% | 1 | 1 | 0% | 8,253 | 5,688 | -31% | 0 | 0 | — |
case-09 | fail→pass | 55,152 | 46,245 | -16% | 1 | 1 | 0% | 6,633 | 5,961 | -10% | 0 | 0 | — |
case-10 | pass→pass | 61,709 | 42,961 | -30% | 1 | 1 | 0% | 7,319 | 6,012 | -18% | 0 | 0 | — |
case-11 | pass→pass | 27,545 | 36,251 | +32% | 1 | 1 | 0% | 3,434 | 5,021 | +46% | 0 | 0 | — |
case-12 | pass→pass | 50,647 | 37,983 | -25% | 1 | 1 | 0% | 6,550 | 5,425 | -17% | 0 | 0 | — |
case-13 | fail→pass | 49,821 | 51,719 | +4% | 1 | 1 | 0% | 5,415 | 6,309 | +17% | 0 | 0 | — |
case-14 | pass→pass | 51,587 | 44,997 | -13% | 1 | 1 | 0% | 5,554 | 5,454 | -2% | 0 | 0 | — |
case-15 | pass→pass | 57,609 | 36,604 | -36% | 1 | 1 | 0% | 6,619 | 5,997 | -9% | 0 | 0 | — |
case-16 | pass→pass | 45,921 | 40,005 | -13% | 1 | 1 | 0% | 5,999 | 5,622 | -6% | 0 | 0 | — |
case-17 | pass→pass | 44,789 | 37,201 | -17% | 1 | 1 | 0% | 4,884 | 5,236 | +7% | 0 | 0 | — |
case-18 | pass→pass | 45,995 | 38,249 | -17% | 1 | 1 | 0% | 6,095 | 4,778 | -22% | 0 | 0 | — |
case-19 | pass→pass | 53,973 | 49,471 | -8% | 1 | 1 | 0% | 8,182 | 6,403 | -22% | 0 | 0 | — |
case-20 | pass→pass | 30,951 | 45,574 | +47% | 1 | 1 | 0% | 4,053 | 6,204 | +53% | 0 | 0 | — |
case-21 | pass→pass | 27,429 | 34,806 | +27% | 1 | 1 | 0% | 4,584 | 5,958 | +30% | 0 | 0 | — |
case-22 | pass→pass | 24,480 | 31,622 | +29% | 1 | 1 | 0% | 3,261 | 4,933 | +51% | 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 +41 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.