Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Chase unanswered emails without being annoying — the escalating-gently sequence with timing rules, the re-ask that makes replying easy, and the close-the-loop discipline that ends zombie threads. Use when asked they haven't replied what do I send, write a follow-up that isn't pushy, how long do I wait before chasing, or manage my waiting-on list. Produces the follow-up sequence with dates, drafts per rung, and the give-up-gracefully exit.
.claude/skills/mohitagw15856-follow-up-chaser/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 44% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 46% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 100% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 43% | 0% |
Unanswered email dies in two ways: never chased (the ask silently expires) or chased badly ("just bumping this!" seven times, each one easier to ignore than the last). Good chasing is engineered: each follow-up lowers the cost of replying — restating the ask smaller, offering a default, adding a deadline with a reason — and the sequence escalates channel or stakes, not temperature. Three rungs, then a graceful close that converts silence into an answer.
Ask for these if not provided:
| Rung | Send date | Channel | The move | |---|---|---|---|
Rung 1: the shrunk ask + default, verbatim · Rung 2: urgency + smallest ask · Rung 3: the channel ping · The close]
The @waiting entry: what, who, next rung date — reviewed weekly]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 14,057 | 12,191 | -13% | 1 | 1 | 0% | 2,812 | 3,049 | +8% | 0 | 0 | — |
case-02 | fail→pass | 11,770 | 10,772 | -8% | 1 | 1 | 0% | 2,335 | 2,891 | +24% | 0 | 0 | — |
case-03 | fail→fail | 19,406 | 13,903 | -28% | 1 | 1 | 0% | 3,439 | 3,206 | -7% | 0 | 0 | — |
case-04 | fail→fail | 12,482 | 16,016 | +28% | 1 | 1 | 0% | 2,174 | 3,606 | +66% | 0 | 0 | — |
case-05 | pass→pass | 9,346 | 9,585 | +3% | 1 | 1 | 0% | 1,522 | 2,437 | +60% | 0 | 0 | — |
case-06 | pass→fail | 5,676 | 12,052 | +112% | 1 | 1 | 0% | 988 | 3,059 | +210% | 0 | 0 | — |
case-07 | fail→pass | 9,702 | 8,259 | -15% | 1 | 1 | 0% | 1,722 | 2,474 | +44% | 0 | 0 | — |
case-08 | fail→pass | 11,026 | 11,392 | +3% | 1 | 1 | 0% | 1,883 | 2,753 | +46% | 0 | 0 | — |
case-09 | pass→pass | 12,018 | 12,531 | +4% | 1 | 1 | 0% | 2,190 | 3,076 | +40% | 0 | 0 | — |
case-10 | pass→pass | 12,878 | 9,527 | -26% | 1 | 1 | 0% | 2,171 | 2,555 | +18% | 0 | 0 | — |
case-11 | fail→pass | 4,993 | 5,803 | +16% | 1 | 1 | 0% | 937 | 1,875 | +100% | 0 | 0 | — |
case-12 | pass→pass | 10,920 | 9,926 | -9% | 1 | 1 | 0% | 1,714 | 2,612 | +52% | 0 | 0 | — |
case-13 | pass→pass | 9,639 | 10,480 | +9% | 1 | 1 | 0% | 1,726 | 2,572 | +49% | 0 | 0 | — |
case-14 | fail→fail | 8,316 | 7,048 | -15% | 1 | 1 | 0% | 1,378 | 2,143 | +56% | 0 | 0 | — |
case-15 | pass→pass | 9,744 | 7,838 | -20% | 1 | 1 | 0% | 1,592 | 2,205 | +39% | 0 | 0 | — |
case-16 | pass→pass | 9,270 | 5,003 | -46% | 1 | 1 | 0% | 1,614 | 1,950 | +21% | 0 | 0 | — |
case-17 | fail→pass | 11,902 | 11,582 | -3% | 1 | 1 | 0% | 2,173 | 3,103 | +43% | 0 | 0 | — |
case-18 | pass→pass | 10,921 | 9,625 | -12% | 1 | 1 | 0% | 1,853 | 2,671 | +44% | 0 | 0 | — |
case-19 | pass→pass | 10,047 | 10,101 | +1% | 1 | 1 | 0% | 1,636 | 2,524 | +54% | 0 | 0 | — |
case-20 | fail→fail | 5,954 | 8,785 | +48% | 1 | 1 | 0% | 1,152 | 2,530 | +120% | 0 | 0 | — |
case-21 | fail→pass | 12,293 | 11,255 | -8% | 1 | 1 | 0% | 2,133 | 2,963 | +39% | 0 | 0 | — |
case-22 | fail→pass | 11,438 | 10,980 | -4% | 1 | 1 | 0% | 2,191 | 2,878 | +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. 22 cases were attempted. The headline lift of +27 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.