Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create and manage scheduled tasks with cron, interval, or one-time execution
.claude/skills/outworked-scheduler/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-20 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 8% | 0% |
| case-15 | ✓→✓ | ＝ Same ✓ | -55% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -14% | 0% |
You can create and manage scheduled tasks that fire on a cron schedule, at regular intervals, or at a specific time.
Standard 5-field cron: minute hour day-of-month month day-of-week
0 9 * * * — every day at 9:00 AM*/15 * * * * — every 15 minutes0 9 * * 1-5 — weekdays at 9:00 AM0 0 1 * * — first of every month at midnightMilliseconds between runs:
60000 — every minute300000 — every 5 minutes3600000 — every hourISO 8601 datetime: 2026-03-25T14:00:00Z
agentId is optional — omit it and the task will be routed to whichever agent handles itscheduler:task_fired trigger event that agents can respond toscheduler:run_now to test a task without waiting for its scheduleagentId to route the task to a specific agent when it fires| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,452 | 6,042 | -6% | 1 | 1 | 0% | 1,148 | 728 | -37% | 0 | 0 | — |
case-02 | fail→fail | 3,829 | 3,561 | -7% | 1 | 1 | 0% | 593 | 560 | -6% | 0 | 0 | — |
case-03 | fail→fail | 6,075 | 7,400 | +22% | 1 | 1 | 0% | 1,034 | 763 | -26% | 0 | 0 | — |
case-04 | pass→pass | 8,061 | 6,181 | -23% | 1 | 1 | 0% | 1,497 | 1,616 | +8% | 0 | 0 | — |
case-15 | pass→pass | 8,106 | 1,523 | -81% | 1 | 1 | 0% | 1,338 | 604 | -55% | 0 | 0 | — |
case-05 | pass→pass | 8,226 | 5,408 | -34% | 1 | 1 | 0% | 1,563 | 1,346 | -14% | 0 | 0 | — |
case-06 | pass→pass | 9,653 | 4,188 | -57% | 1 | 1 | 0% | 1,694 | 1,057 | -38% | 0 | 0 | — |
case-07 | fail→fail | 8,590 | 4,518 | -47% | 1 | 1 | 0% | 1,437 | 996 | -31% | 0 | 0 | — |
case-08 | fail→fail | 9,298 | 8,152 | -12% | 1 | 1 | 0% | 1,567 | 1,011 | -35% | 0 | 0 | — |
case-09 | fail→fail | 6,906 | 7,027 | +2% | 1 | 1 | 0% | 1,282 | 788 | -39% | 0 | 0 | — |
case-10 | fail→fail | 9,904 | 2,020 | -80% | 1 | 1 | 0% | 1,669 | 622 | -63% | 0 | 0 | — |
case-11 | fail→fail | 4,763 | 4,626 | -3% | 1 | 1 | 0% | 740 | 537 | -27% | 0 | 0 | — |
case-12 | fail→fail | 5,101 | 5,518 | +8% | 1 | 1 | 0% | 964 | 786 | -18% | 0 | 0 | — |
case-13 | fail→fail | 5,451 | 3,852 | -29% | 1 | 1 | 0% | 896 | 549 | -39% | 0 | 0 | — |
case-14 | fail→fail | 7,729 | 6,336 | -18% | 1 | 1 | 0% | 1,522 | 682 | -55% | 0 | 0 | — |
case-16 | fail→fail | 12,459 | 4,163 | -67% | 1 | 1 | 0% | 2,020 | 966 | -52% | 0 | 0 | — |
case-17 | fail→fail | 6,905 | 4,108 | -41% | 1 | 1 | 0% | 1,368 | 951 | -30% | 0 | 0 | — |
case-18 | fail→fail | 5,552 | 4,199 | -24% | 1 | 1 | 0% | 1,038 | 915 | -12% | 0 | 0 | — |
case-19 | pass→pass | 11,197 | 1,446 | -87% | 1 | 1 | 0% | 1,783 | 581 | -67% | 0 | 0 | — |
case-20 | fail→pass | 10,795 | 1,300 | -88% | 1 | 1 | 0% | 1,818 | 555 | -69% | 0 | 0 | — |
case-21 | fail→pass | 11,145 | 4,329 | -61% | 1 | 1 | 0% | 1,827 | 973 | -47% | 0 | 0 | — |
case-22 | fail→fail | 9,011 | 5,634 | -37% | 1 | 1 | 0% | 1,522 | 846 | -44% | 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, and 14 counted toward the lift figure. The other 8 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +9 percentage points is the difference between those two pass rates over the 14 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.