Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →GitHub、Linear、デスクトップアラート、フック、接続された通信インターフェースを網羅する、統合されたECCネイティブワークフローとして通知を運用する。真の問題がアラートルーティング、重複排除、エスカレーション、またはインボックス崩壊である場合に使用する。
.claude/skills/affaan-m-unified-notifications-ops/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 67% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 69% | 0% |
当真正的问题不是缺少通知，而是通知系统碎片化时，使用此技能。
任务是将分散的事件整合到一个操作员界面上，包含：
从已有资源出发：
优先使用 ECC 原生编排，而非建议用户采用独立的通知产品。
将通道视为：
目标是更少但更好的通知。
| 等级 | 示例 | 默认处理方式 | | --- | --- | --- | | 严重 | 默认分支 CI 损坏、安全问题、发布受阻、部署失败 | 立即中断 | | 高 | 请求审查、PR 失败、阻塞责任人的交接 | 当日提醒 | | 中 | 问题状态变更、重要评论、积压变动 | 摘要或队列 | | 低 | 重复成功、常规噪音、冗余生命周期标记 | 抑制或折叠 |
如果工作区没有严重等级模型，请先构建一个，再提出自动化方案。
列出：
指出 ECC 已拥有的部分。
针对每个事件族，回答：
使用以下默认值：
检查：
优先选择：
针对每个真实通知需求，定义：
如果 ECC 已有原语，优先使用：
最终输出：
text当前表面 - 来源 - 渠道 - 重复项 - 缺口 事件模型 - 严重 - 高 - 中 - 低 路由计划 - 来源 -> 渠道 - 原因 - 操作员/负责人 整合 - 抑制 - 合并 - 规范摘要 下一步ECC行动 - 技能/钩子/代理/MCP - 下一步要构建的具体工作流
project-flow-opsworkspace-surface-auditworkspace-surface-auditproject-flow-opsgithub-opsknowledge-opscustomer-billing-ops 当通知痛点涉及计费/客户运营而非工程时| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 30,589 | 18,113 | -41% | 1 | 1 | 0% | 4,322 | 4,040 | -7% | 0 | 0 | — |
case-02 | fail→pass | 37,194 | 10,980 | -70% | 1 | 1 | 0% | 5,946 | 2,978 | -50% | 0 | 0 | — |
case-08 | pass→pass | 13,350 | 14,130 | +6% | 1 | 1 | 0% | 1,880 | 3,414 | +82% | 0 | 0 | — |
case-03 | pass→pass | 14,331 | 12,405 | -13% | 1 | 1 | 0% | 2,083 | 3,306 | +59% | 0 | 0 | — |
case-04 | fail→pass | 14,494 | 16,077 | +11% | 1 | 1 | 0% | 2,229 | 3,733 | +67% | 0 | 0 | — |
case-05 | fail→pass | 14,067 | 13,452 | -4% | 1 | 1 | 0% | 1,992 | 3,292 | +65% | 0 | 0 | — |
case-06 | fail→pass | 13,280 | 11,705 | -12% | 1 | 1 | 0% | 1,829 | 3,087 | +69% | 0 | 0 | — |
case-07 | pass→pass | 11,588 | 9,434 | -19% | 1 | 1 | 0% | 1,845 | 2,780 | +51% | 0 | 0 | — |
case-09 | fail→fail | 14,347 | 16,088 | +12% | 1 | 1 | 0% | 2,189 | 3,629 | +66% | 0 | 0 | — |
case-10 | fail→fail | 14,406 | 9,994 | -31% | 1 | 1 | 0% | 2,210 | 2,696 | +22% | 0 | 0 | — |
case-11 | fail→pass | 14,514 | 11,684 | -19% | 1 | 1 | 0% | 2,138 | 2,997 | +40% | 0 | 0 | — |
case-12 | pass→pass | 14,672 | 7,336 | -50% | 1 | 1 | 0% | 2,036 | 2,496 | +23% | 0 | 0 | — |
case-13 | pass→pass | 12,372 | 18,155 | +47% | 1 | 1 | 0% | 1,941 | 4,092 | +111% | 0 | 0 | — |
case-14 | pass→pass | 11,950 | 13,448 | +13% | 1 | 1 | 0% | 1,806 | 3,290 | +82% | 0 | 0 | — |
case-15 | pass→pass | 21,367 | 14,010 | -34% | 1 | 1 | 0% | 3,190 | 3,196 | +0% | 0 | 0 | — |
case-16 | pass→pass | 10,848 | 9,478 | -13% | 1 | 1 | 0% | 1,484 | 2,694 | +82% | 0 | 0 | — |
case-17 | fail→pass | 29,065 | 11,241 | -61% | 1 | 1 | 0% | 2,961 | 2,934 | -1% | 0 | 0 | — |
case-18 | pass→pass | 7,310 | 9,769 | +34% | 1 | 1 | 0% | 1,047 | 2,761 | +164% | 0 | 0 | — |
case-19 | fail→pass | 15,253 | 11,959 | -22% | 1 | 1 | 0% | 2,159 | 2,855 | +32% | 0 | 0 | — |
case-20 | pass→fail | 8,449 | 17,653 | +109% | 1 | 1 | 0% | 1,531 | 3,533 | +131% | 0 | 0 | — |
case-21 | pass→pass | 18,725 | 28,172 | +50% | 1 | 1 | 0% | 3,424 | 4,427 | +29% | 0 | 0 | — |
case-22 | pass→fail | 12,605 | 17,853 | +42% | 1 | 1 | 0% | 2,011 | 4,200 | +109% | 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. 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.