Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Set the shared-space norms that end the small daily frictions — meeting room booking discipline (the ghost-booking cure), hybrid-call room behavior, the shared kitchen/space contracts, and the enforcement that works without a hall monitor. Use when asked set office space norms, rooms are always booked but empty, hybrid meetings are terrible for remote people, or write the office etiquette guide. Produces the norms card by space type, the ghost-booking fix, the hybrid-room checklist, and the no-h
.claude/skills/mohitagw15856-meeting-room-etiquette/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 40% | 0% |
Office friction lives in the commons: rooms booked-and-empty while teams huddle in hallways, the hybrid call where the room forgets the remote half exists, the kitchen that becomes a passive-aggressive noticeboard. These are working-agreements problems at building scale — solved the same way: a few observable norms derived from the actual frictions, made visible where the friction happens (the room's door, the kitchen's wall), and enforced by structure (release timers, checklists) rather than by anyone having to be the hall monitor — because norms that require a cop get one, and everyone hates the cop more than the problem.
Ask for these if not provided:
Rooms: 3–5 rules incl. the release rule] · Hybrid rooms: the checklist, on the table] · Kitchen/commons: the 3 that matter]
Auto-release configured (or the door-card fallback) · 25/50 defaults org-wide · the quarterly ghost-hold audit]
Faces on screen · mic tested · remote advocate named · no side-talk · share via call]
Structure-first mapping: norm → its default/timer/checklist · the chronic-case path (manager, private) · no hall monitors, by design]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 21,040 | 22,578 | +7% | 1 | 1 | 0% | 3,688 | 5,268 | +43% | 0 | 0 | — |
case-02 | fail→pass | 28,062 | 16,376 | -42% | 1 | 1 | 0% | 4,359 | 4,031 | -8% | 0 | 0 | — |
case-03 | fail→pass | 22,268 | 19,852 | -11% | 1 | 1 | 0% | 3,512 | 4,839 | +38% | 0 | 0 | — |
case-04 | fail→pass | 14,015 | 12,333 | -12% | 1 | 1 | 0% | 2,283 | 3,216 | +41% | 0 | 0 | — |
case-05 | pass→pass | 13,487 | 13,162 | -2% | 1 | 1 | 0% | 2,258 | 3,607 | +60% | 0 | 0 | — |
case-06 | fail→pass | 21,861 | 10,749 | -51% | 1 | 1 | 0% | 2,123 | 2,972 | +40% | 0 | 0 | — |
case-07 | fail→pass | 21,702 | 15,878 | -27% | 1 | 1 | 0% | 1,462 | 2,834 | +94% | 0 | 0 | — |
case-08 | pass→pass | 10,143 | 8,429 | -17% | 1 | 1 | 0% | 1,480 | 2,287 | +55% | 0 | 0 | — |
case-09 | pass→pass | 26,430 | 22,306 | -16% | 1 | 1 | 0% | 1,131 | 2,164 | +91% | 0 | 0 | — |
case-10 | pass→pass | 12,056 | 24,711 | +105% | 1 | 1 | 0% | 1,459 | 2,336 | +60% | 0 | 0 | — |
case-11 | pass→pass | 6,124 | 28,096 | +359% | 1 | 1 | 0% | 871 | 2,114 | +143% | 0 | 0 | — |
case-12 | pass→pass | 12,952 | 11,499 | -11% | 1 | 1 | 0% | 2,188 | 3,057 | +40% | 0 | 0 | — |
case-13 | pass→pass | 11,028 | 12,595 | +14% | 1 | 1 | 0% | 1,882 | 3,405 | +81% | 0 | 0 | — |
case-14 | pass→pass | 15,356 | 9,277 | -40% | 1 | 1 | 0% | 2,467 | 2,869 | +16% | 0 | 0 | — |
case-15 | fail→fail | 17,786 | 16,364 | -8% | 1 | 1 | 0% | 2,916 | 3,794 | +30% | 0 | 0 | — |
case-16 | fail→fail | 21,284 | 17,006 | -20% | 1 | 1 | 0% | 4,071 | 4,537 | +11% | 0 | 0 | — |
case-17 | fail→fail | 19,546 | 17,864 | -9% | 1 | 1 | 0% | 3,354 | 4,696 | +40% | 0 | 0 | — |
case-18 | pass→pass | 11,405 | 6,896 | -40% | 1 | 1 | 0% | 1,870 | 2,490 | +33% | 0 | 0 | — |
case-19 | pass→pass | 13,005 | 5,632 | -57% | 1 | 1 | 0% | 1,933 | 2,263 | +17% | 0 | 0 | — |
case-20 | pass→pass | 13,023 | 21,082 | +62% | 1 | 1 | 0% | 2,224 | 4,699 | +111% | 0 | 0 | — |
case-21 | pass→pass | 10,900 | 7,293 | -33% | 1 | 1 | 0% | 1,733 | 2,627 | +52% | 0 | 0 | — |
case-22 | pass→pass | 12,259 | 12,634 | +3% | 1 | 1 | 0% | 2,034 | 3,463 | +70% | 0 | 0 | — |
case-23 | pass→pass | 15,001 | 8,572 | -43% | 1 | 1 | 0% | 2,500 | 2,807 | +12% | 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. 23 cases were attempted. The headline lift of +26 percentage points is the difference between those two pass rates over the 23 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.