Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Document a guest incident at a hospitality venue — injury, illness, foodborne complaint, altercation, or property loss — into a clear, defensible record. Use when asked to write up a guest incident, log an accident or complaint, document a slip/fall or allergic reaction, or record an incident for insurance/legal. Produces a factual incident report (who/what/when/where, witnesses, actions taken), an immediate-response checklist, notification/escalation steps, and follow-up — objective and liabili
.claude/skills/mohitagw15856-guest-incident-log/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 35% | 0% |
When a guest slips, has a reaction, or complains of illness, what you write in the next hour can protect both the guest and the business — or sink it. This skill captures the facts objectively, records the response, and routes the notifications, without the speculation or apologies-as-admissions that create legal exposure.
Given what happened, produce the full record — capture only observed facts, mark anything reported-by-guest as such, and never invent details. Flag anything that requires immediate medical, legal, or regulatory action.
Ask for (if not provided, else infer and label):
The right-now checklist: ensure safety/medical aid, secure the area/preserve evidence (the spilled item, the food, photos), don't move hazards until documented, get witness details.
| Field | Detail | |---|---| | Date / time / location | | | Incident type | injury / illness / foodborne / altercation / property | | Persons involved | guest, staff, witnesses (contacts) | | Factual account | observed facts, in sequence | | Guest statements | quoted, marked "as reported by guest" | | Actions taken | aid given, area secured, who was notified | | Evidence | photos, retained item, CCTV reference |
Who to inform and when — manager on duty, owner/GM, insurance carrier, and any mandatory reporting (a suspected foodborne illness may require the health department).
Guest follow-up (care, not admission), corrective action to prevent recurrence, and where the record is retained.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 15,004 | 11,884 | -21% | 1 | 1 | 0% | 2,748 | 2,629 | -4% | 0 | 0 | — |
case-02 | fail→pass | 13,233 | 12,660 | -4% | 1 | 1 | 0% | 2,422 | 2,737 | +13% | 0 | 0 | — |
case-03 | fail→pass | 11,097 | 9,230 | -17% | 1 | 1 | 0% | 1,939 | 2,256 | +16% | 0 | 0 | — |
case-04 | pass→pass | 9,034 | 12,408 | +37% | 1 | 1 | 0% | 1,584 | 2,770 | +75% | 0 | 0 | — |
case-05 | pass→pass | 13,027 | 12,103 | -7% | 1 | 1 | 0% | 2,257 | 2,642 | +17% | 0 | 0 | — |
case-06 | fail→pass | 13,981 | 14,805 | +6% | 1 | 1 | 0% | 2,450 | 3,013 | +23% | 0 | 0 | — |
case-07 | pass→pass | 8,876 | 10,259 | +16% | 1 | 1 | 0% | 1,560 | 2,342 | +50% | 0 | 0 | — |
case-08 | pass→pass | 14,056 | 13,843 | -2% | 1 | 1 | 0% | 2,391 | 2,864 | +20% | 0 | 0 | — |
case-09 | fail→pass | 9,782 | 11,430 | +17% | 1 | 1 | 0% | 1,861 | 2,515 | +35% | 0 | 0 | — |
case-10 | fail→pass | 6,639 | 11,763 | +77% | 1 | 1 | 0% | 1,217 | 2,540 | +109% | 0 | 0 | — |
case-11 | pass→pass | 11,294 | 12,798 | +13% | 1 | 1 | 0% | 1,965 | 2,807 | +43% | 0 | 0 | — |
case-12 | pass→pass | 11,387 | 10,329 | -9% | 1 | 1 | 0% | 1,916 | 2,568 | +34% | 0 | 0 | — |
case-13 | pass→pass | 8,274 | 11,520 | +39% | 1 | 1 | 0% | 1,551 | 2,685 | +73% | 0 | 0 | — |
case-14 | fail→fail | 5,784 | 9,192 | +59% | 1 | 1 | 0% | 1,049 | 2,152 | +105% | 0 | 0 | — |
case-15 | pass→pass | 10,542 | 10,456 | -1% | 1 | 1 | 0% | 1,941 | 2,584 | +33% | 0 | 0 | — |
case-16 | pass→pass | 6,975 | 10,727 | +54% | 1 | 1 | 0% | 1,348 | 2,561 | +90% | 0 | 0 | — |
case-17 | pass→pass | 9,315 | 11,569 | +24% | 1 | 1 | 0% | 1,881 | 2,610 | +39% | 0 | 0 | — |
case-18 | pass→pass | 9,962 | 11,508 | +16% | 1 | 1 | 0% | 1,804 | 2,527 | +40% | 0 | 0 | — |
case-19 | pass→pass | 9,485 | 9,953 | +5% | 1 | 1 | 0% | 1,716 | 2,477 | +44% | 0 | 0 | — |
case-20 | pass→pass | 12,975 | 12,011 | -7% | 1 | 1 | 0% | 2,332 | 2,546 | +9% | 0 | 0 | — |
case-21 | pass→pass | 21,793 | 16,201 | -26% | 1 | 1 | 0% | 3,662 | 3,400 | -7% | 0 | 0 | — |
case-22 | pass→pass | 14,359 | 11,991 | -16% | 1 | 1 | 0% | 2,554 | 2,619 | +3% | 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 +27 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.