Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Send an email with multiple to, cc and bcc addresses.
.claude/skills/aiskillstore-extension-email-raw/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -13% | 0% |
Raw multi-recipient email extension for Caffeine AI.
This skill adds support for sending emails with multiple to, cc, and bcc recipients. Not suitable for bulk service emails (recipients see each other).
to, cc and bcc addresses.mo:caffeineai-email/emailClient.momodule { public type SendResult = { #ok; #err : Text; }; public func sendRawEmail( fromUsername : Text, to : [Text], cc : [Text], bcc : [Text], subject : Text, htmlBody : Text, ) : async SendResult; };
motoko filepath=src/backend/main.moimport Runtime "mo:core/Runtime"; import EmailClient "mo:caffeineai-email/emailClient"; actor { public func sendMeetingReminder( meetingSubject : Text, meetingTime : Text, confirmedAttendeeEmails : [Text], tentativeAttendeeEmails : [Text], ) : async () { let result = await EmailClient.sendRawEmail( "no-reply", confirmedAttendeeEmails, tentativeAttendeeEmails, [], meetingSubject, "Reminder the meeting will start at " # meetingTime, ); switch (result) { case (#ok) {}; case (#err(error)) { Runtime.trap("Failed to send meeting reminder email: " # error); }; }; }; };
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 21,680 | 11,853 | -45% | 1 | 1 | 0% | 3,367 | 1,821 | -46% | 0 | 0 | — |
case-17 | fail→pass | 14,002 | 11,270 | -20% | 1 | 1 | 0% | 2,431 | 1,794 | -26% | 0 | 0 | — |
case-18 | pass→pass | 16,260 | 11,549 | -29% | 1 | 1 | 0% | 2,062 | 1,760 | -15% | 0 | 0 | — |
case-02 | fail→pass | 19,460 | 10,904 | -44% | 1 | 1 | 0% | 3,038 | 1,785 | -41% | 0 | 0 | — |
case-03 | fail→pass | 23,855 | 13,201 | -45% | 1 | 1 | 0% | 3,828 | 2,103 | -45% | 0 | 0 | — |
case-04 | pass→pass | 15,951 | 13,154 | -18% | 1 | 1 | 0% | 1,615 | 1,729 | +7% | 0 | 0 | — |
case-05 | pass→pass | 20,873 | 12,803 | -39% | 1 | 1 | 0% | 2,638 | 1,953 | -26% | 0 | 0 | — |
case-06 | pass→pass | 17,446 | 11,661 | -33% | 1 | 1 | 0% | 1,879 | 1,341 | -29% | 0 | 0 | — |
case-07 | pass→pass | 16,768 | 8,432 | -50% | 1 | 1 | 0% | 1,606 | 1,157 | -28% | 0 | 0 | — |
case-08 | fail→pass | 15,316 | 9,037 | -41% | 1 | 1 | 0% | 1,438 | 1,256 | -13% | 0 | 0 | — |
case-09 | fail→pass | 20,498 | 8,015 | -61% | 1 | 1 | 0% | 2,295 | 1,079 | -53% | 0 | 0 | — |
case-10 | fail→pass | 22,994 | 10,928 | -52% | 1 | 1 | 0% | 2,504 | 1,574 | -37% | 0 | 0 | — |
case-11 | fail→pass | 8,174 | 9,079 | +11% | 1 | 1 | 0% | 1,350 | 1,442 | +7% | 0 | 0 | — |
case-12 | fail→pass | 13,277 | 4,864 | -63% | 1 | 1 | 0% | 1,541 | 1,498 | -3% | 0 | 0 | — |
case-13 | fail→pass | 16,373 | 5,605 | -66% | 1 | 1 | 0% | 2,660 | 1,644 | -38% | 0 | 0 | — |
case-14 | fail→pass | 10,467 | 5,423 | -48% | 1 | 1 | 0% | 1,954 | 1,644 | -16% | 0 | 0 | — |
case-15 | fail→pass | 19,307 | 12,077 | -37% | 1 | 1 | 0% | 2,471 | 1,905 | -23% | 0 | 0 | — |
case-16 | fail→pass | 19,508 | 6,638 | -66% | 1 | 1 | 0% | 2,682 | 1,887 | -30% | 0 | 0 | — |
case-19 | fail→pass | 20,638 | 6,066 | -71% | 1 | 1 | 0% | 2,844 | 1,759 | -38% | 0 | 0 | — |
case-20 | fail→pass | 22,409 | 7,768 | -65% | 1 | 1 | 0% | 3,510 | 2,073 | -41% | 0 | 0 | — |
case-21 | fail→pass | 13,500 | 2,660 | -80% | 1 | 1 | 0% | 1,822 | 1,018 | -44% | 0 | 0 | — |
case-22 | fail→pass | 9,225 | 7,798 | -15% | 1 | 1 | 0% | 1,545 | 1,011 | -35% | 0 | 0 | — |
case-23 | fail→pass | 13,669 | 2,621 | -81% | 1 | 1 | 0% | 2,153 | 882 | -59% | 0 | 0 | — |
case-24 | fail→pass | 11,938 | 10,600 | -11% | 1 | 1 | 0% | 1,687 | 1,398 | -17% | 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. 24 cases were attempted. The headline lift of +79 percentage points is the difference between those two pass rates over the 24 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.