Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →ECC用の証拠ベースのメールボックストリアージ、ドラフト作成、送信検証、および送信済みメールセーフフォローアップワークフロー。ユーザーがメールを整理したり、実際のメールサーフェスを通じてドラフトまたは送信したい、または送信済みメールに何が到着したかを証明したい場合に使用します。
.claude/skills/affaan-m-email-ops/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 122% | 0% |
当实际任务为邮箱工作时使用：分类、起草、回复、发送，或确认邮件已进入已发送文件夹。
这不是通用写作技能，而是围绕实际邮件界面的操作工作流。
在相关场景下调用这些ECC原生技能：
brand-voice 在起草任何面向用户的内容之前investor-outreach 用于面向投资者、合作伙伴或赞助商的邮件customer-billing-ops 当邮件线程属于账单/支持事件而非普通通信时knowledge-ops 当需要将消息或线程捕获到持久上下文中时research-ops 当回复依赖最新外部事实时messages-ops操作前明确：
若回复：
若创建新外发邮件：
brand-voice仅起草任务：
实时发送任务：
使用精确状态词：
若发送界面被阻止，保留草稿并报告确切阻止原因，而非未经说明即改用第二传输方式。
text邮件界面 - 账户 - 邮件线程/收件人 - 请求的操作 草稿 - 主题 - 正文 状态 - 已草拟/已发送/已拦截 - 适用时附上发送证明 下一步 - 发送 - 跟进 - 归档/移动
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 11,407 | 7,619 | -33% | 1 | 1 | 0% | 1,831 | 2,069 | +13% | 0 | 0 | — |
case-02 | fail→pass | 16,279 | 9,058 | -44% | 1 | 1 | 0% | 2,570 | 2,115 | -18% | 0 | 0 | — |
case-03 | fail→pass | 10,913 | 8,867 | -19% | 1 | 1 | 0% | 1,780 | 2,101 | +18% | 0 | 0 | — |
case-04 | pass→pass | 3,532 | 7,837 | +122% | 1 | 1 | 0% | 424 | 1,674 | +295% | 0 | 0 | — |
case-05 | fail→pass | 11,310 | 7,586 | -33% | 1 | 1 | 0% | 1,761 | 1,900 | +8% | 0 | 0 | — |
case-06 | pass→fail | 7,211 | 8,293 | +15% | 1 | 1 | 0% | 962 | 2,069 | +115% | 0 | 0 | — |
case-07 | fail→pass | 6,622 | 8,312 | +26% | 1 | 1 | 0% | 917 | 2,035 | +122% | 0 | 0 | — |
case-08 | fail→pass | 10,399 | 6,791 | -35% | 1 | 1 | 0% | 1,504 | 1,712 | +14% | 0 | 0 | — |
case-09 | pass→pass | 11,942 | 7,585 | -36% | 1 | 1 | 0% | 1,979 | 1,928 | -3% | 0 | 0 | — |
case-10 | fail→pass | 13,860 | 7,096 | -49% | 1 | 1 | 0% | 2,242 | 1,865 | -17% | 0 | 0 | — |
case-11 | pass→pass | 11,052 | 8,979 | -19% | 1 | 1 | 0% | 1,613 | 2,010 | +25% | 0 | 0 | — |
case-12 | pass→pass | 14,705 | 7,402 | -50% | 1 | 1 | 0% | 2,205 | 1,841 | -17% | 0 | 0 | — |
case-13 | fail→fail | 12,576 | 7,770 | -38% | 1 | 1 | 0% | 1,897 | 1,995 | +5% | 0 | 0 | — |
case-14 | fail→pass | 15,911 | 12,964 | -19% | 1 | 1 | 0% | 2,187 | 2,575 | +18% | 0 | 0 | — |
case-15 | pass→pass | 12,176 | 7,534 | -38% | 1 | 1 | 0% | 1,783 | 1,984 | +11% | 0 | 0 | — |
case-16 | fail→pass | 10,059 | 6,620 | -34% | 1 | 1 | 0% | 1,582 | 1,732 | +9% | 0 | 0 | — |
case-17 | fail→pass | 14,120 | 10,101 | -28% | 1 | 1 | 0% | 2,071 | 2,275 | +10% | 0 | 0 | — |
case-18 | pass→pass | 12,650 | 7,944 | -37% | 1 | 1 | 0% | 2,053 | 1,913 | -7% | 0 | 0 | — |
case-19 | pass→pass | 6,543 | 7,426 | +13% | 1 | 1 | 0% | 967 | 1,889 | +95% | 0 | 0 | — |
case-20 | pass→pass | 7,087 | 7,137 | +1% | 1 | 1 | 0% | 1,036 | 1,731 | +67% | 0 | 0 | — |
case-21 | pass→pass | 12,036 | 10,372 | -14% | 1 | 1 | 0% | 1,831 | 2,051 | +12% | 0 | 0 | — |
case-22 | fail→pass | 10,404 | 6,317 | -39% | 1 | 1 | 0% | 1,392 | 1,725 | +24% | 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. 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.