Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Turn an overflowing inbox into a four-verb system — archive, reply-now, task, or park — with the two-minute rule enforced and a daily cadence that survives busy weeks. Use when asked help me get to inbox zero, my email is out of control, build me an email triage system, or process this backlog. Produces the triage pass on the actual inbox, the four-verb rules, the folder/label minimal set, and the daily cadence.
.claude/skills/mohitagw15856-email-triage-system/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 89% | 9 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 5% | 0% |
Inboxes overflow because every message gets the same treatment: opened, half-read, left "for later" — and later compounds. Triage replaces reading with deciding: every message gets exactly one of four verbs on first touch, the two-minute rule handles the small stuff instantly, and the system runs on a cadence measured in minutes per day, not heroic weekend purges.
Ask for these if not provided:
@waiting label and a weekly review. Parking without the review label is just procrastination with a name.Batches with verb counts: "archive ~3,100 (newsletters, notifications) · reply-now 12 · task 23 · park 4" · the bankruptcy call if warranted]
The four verbs with their tests · the two-minute rule · labels: @waiting, @task-filed, +1 optional]
Windows, lengths, the closed-between rule · the weekly @waiting 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 | 16,379 | 12,439 | -24% | 1 | 1 | 0% | 2,553 | 3,058 | +20% | 0 | 0 | — |
case-02 | fail→fail | 16,234 | 17,445 | +7% | 1 | 1 | 0% | 2,642 | 3,555 | +35% | 0 | 0 | — |
case-03 | fail→pass | 19,531 | 13,866 | -29% | 1 | 1 | 0% | 2,903 | 3,075 | +6% | 0 | 0 | — |
case-04 | pass→pass | 13,269 | 13,526 | +2% | 1 | 1 | 0% | 2,260 | 3,117 | +38% | 0 | 0 | — |
case-05 | pass→pass | 16,810 | 17,274 | +3% | 1 | 1 | 0% | 2,739 | 3,431 | +25% | 0 | 0 | — |
case-06 | pass→pass | 9,388 | 10,294 | +10% | 1 | 1 | 0% | 1,666 | 2,403 | +44% | 0 | 0 | — |
case-07 | pass→pass | 8,687 | 8,594 | -1% | 1 | 1 | 0% | 1,344 | 2,093 | +56% | 0 | 0 | — |
case-08 | pass→pass | 14,699 | 14,989 | +2% | 1 | 1 | 0% | 2,227 | 3,150 | +41% | 0 | 0 | — |
case-09 | pass→pass | 11,222 | 10,074 | -10% | 1 | 1 | 0% | 1,653 | 2,299 | +39% | 0 | 0 | — |
case-10 | fail→pass | 10,315 | 10,990 | +7% | 1 | 1 | 0% | 1,745 | 2,654 | +52% | 0 | 0 | — |
case-11 | pass→pass | 14,151 | 11,176 | -21% | 1 | 1 | 0% | 2,052 | 2,411 | +17% | 0 | 0 | — |
case-12 | fail→pass | 11,074 | 9,451 | -15% | 1 | 1 | 0% | 1,774 | 2,215 | +25% | 0 | 0 | — |
case-13 | fail→pass | 13,002 | 7,836 | -40% | 1 | 1 | 0% | 2,103 | 2,198 | +5% | 0 | 0 | — |
case-14 | fail→pass | 11,035 | 7,977 | -28% | 1 | 1 | 0% | 1,634 | 2,050 | +25% | 0 | 0 | — |
case-15 | pass→pass | 14,873 | 13,093 | -12% | 1 | 1 | 0% | 2,210 | 3,114 | +41% | 0 | 0 | — |
case-16 | pass→pass | 9,406 | 6,390 | -32% | 1 | 1 | 0% | 1,374 | 1,778 | +29% | 0 | 0 | — |
case-17 | fail→pass | 13,150 | 10,939 | -17% | 1 | 1 | 0% | 1,982 | 2,672 | +35% | 0 | 0 | — |
case-18 | pass→pass | 10,401 | 9,890 | -5% | 1 | 1 | 0% | 1,720 | 2,354 | +37% | 0 | 0 | — |
case-19 | pass→pass | 9,799 | 8,278 | -16% | 1 | 1 | 0% | 1,409 | 2,047 | +45% | 0 | 0 | — |
case-20 | pass→pass | 7,182 | 2,624 | -63% | 1 | 1 | 0% | 1,143 | 1,303 | +14% | 0 | 0 | — |
case-21 | pass→pass | 9,352 | 6,041 | -35% | 1 | 1 | 0% | 1,422 | 1,741 | +22% | 0 | 0 | — |
case-22 | fail→pass | 10,861 | 4,139 | -62% | 1 | 1 | 0% | 1,672 | 1,521 | -9% | 0 | 0 | — |
case-23 | fail→pass | 9,270 | 7,459 | -20% | 1 | 1 | 0% | 1,497 | 1,956 | +31% | 0 | 0 | — |
case-24 | pass→pass | 13,121 | 13,951 | +6% | 1 | 1 | 0% | 1,981 | 2,993 | +51% | 0 | 0 | — |
case-25 | pass→pass | 13,346 | 11,564 | -13% | 1 | 1 | 0% | 1,996 | 2,618 | +31% | 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. 25 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 25 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.