Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Design safe WeChat follower-message automation and send customer-service text only when the user provides an explicit target and confirmation.
.claude/skills/devin-axis-wechat-official-messages/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -29% | 0% |
用于设计私信、欢迎语和客服回复策略。
send-customer-text 向明确的 OpenID 发送经用户确认的文本。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 14,625 | 12,109 | -17% | 1 | 1 | 0% | 1,646 | 1,187 | -28% | 0 | 0 | — |
case-02 | fail→pass | 37,764 | 27,467 | -27% | 1 | 1 | 0% | 3,038 | 3,862 | +27% | 0 | 0 | — |
case-03 | fail→pass | 19,845 | 19,708 | -1% | 1 | 1 | 0% | 2,161 | 2,436 | +13% | 0 | 0 | — |
case-04 | fail→pass | 18,176 | 16,273 | -10% | 1 | 1 | 0% | 2,034 | 1,806 | -11% | 0 | 0 | — |
case-05 | pass→pass | 18,418 | 14,763 | -20% | 1 | 1 | 0% | 2,174 | 1,645 | -24% | 0 | 0 | — |
case-06 | fail→pass | 22,922 | 17,803 | -22% | 1 | 1 | 0% | 2,913 | 2,119 | -27% | 0 | 0 | — |
case-07 | fail→pass | 25,643 | 17,765 | -31% | 1 | 1 | 0% | 3,203 | 2,282 | -29% | 0 | 0 | — |
case-08 | fail→pass | 15,204 | 13,561 | -11% | 1 | 1 | 0% | 1,545 | 1,528 | -1% | 0 | 0 | — |
case-09 | pass→pass | 20,135 | 14,978 | -26% | 1 | 1 | 0% | 2,215 | 1,729 | -22% | 0 | 0 | — |
case-10 | fail→pass | 17,518 | 14,151 | -19% | 1 | 1 | 0% | 1,952 | 1,577 | -19% | 0 | 0 | — |
case-11 | pass→pass | 15,939 | 11,457 | -28% | 1 | 1 | 0% | 1,644 | 1,169 | -29% | 0 | 0 | — |
case-12 | fail→pass | 15,495 | 7,915 | -49% | 1 | 1 | 0% | 1,675 | 612 | -63% | 0 | 0 | — |
case-13 | pass→pass | 35,385 | 23,548 | -33% | 1 | 1 | 0% | 3,876 | 2,889 | -25% | 0 | 0 | — |
case-14 | fail→pass | 26,987 | 21,807 | -19% | 1 | 1 | 0% | 3,083 | 2,910 | -6% | 0 | 0 | — |
case-15 | pass→pass | 17,771 | 11,765 | -34% | 1 | 1 | 0% | 1,829 | 1,318 | -28% | 0 | 0 | — |
case-16 | pass→pass | 21,678 | 15,986 | -26% | 1 | 1 | 0% | 2,314 | 1,750 | -24% | 0 | 0 | — |
case-17 | pass→pass | 24,024 | 16,956 | -29% | 1 | 1 | 0% | 2,869 | 1,985 | -31% | 0 | 0 | — |
case-18 | pass→pass | 28,790 | 27,292 | -5% | 1 | 1 | 0% | 3,375 | 3,510 | +4% | 0 | 0 | — |
case-19 | fail→pass | 13,550 | 14,438 | +7% | 1 | 1 | 0% | 1,359 | 1,680 | +24% | 0 | 0 | — |
case-20 | pass→pass | 29,016 | 29,807 | +3% | 1 | 1 | 0% | 4,546 | 5,009 | +10% | 0 | 0 | — |
case-21 | pass→pass | 16,991 | 14,307 | -16% | 1 | 1 | 0% | 1,827 | 1,617 | -11% | 0 | 0 | — |
case-22 | pass→pass | 19,242 | 22,334 | +16% | 1 | 1 | 0% | 3,082 | 3,105 | +1% | 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 +45 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.