Load context
On first use: Read ../../CLAUDE.md to obtain scenario-level configuration (alert threshold definitions/escalation paths/responsible persons).
/metric-alert-handling — Metric Alert Handling
Examples
→ Example: The user says "several metrics have triggered alerts and need to be highlighted in the management report," and the system should invoke this skill to generate alert handling explanations and response measures.
→ Example: The user says "the Revenue alert was later verified to be a false alarm and the data needs to be corrected and explained to management," and the system should invoke this skill to generate a false alarm correction explanation.
→ Example: The user says "the cash flow alert has been triggered for 2 consecutive months; does it need to be escalated," and the system should invoke this skill to determine whether the escalation trigger conditions have been met.
Step 1: Alert Information Confirmation
Basic alert information:
□ 触发时间：[YYYY-MM-DD HH:MM]
□ 预警指标：[KPI/指标名称]
□ 预警类型：[跌破阈值/异常波动/趋势恶化/数据缺失]
□ 预警详情：
→ 触发值：[X]
→ 预警阈值：[X]
→ 偏差幅度：[X]%（触发值 vs 阈值）
→ 本期 vs 上期：[X]（[+/-X]%）
→ 预警级别：[🔴 严重 / 🟡 中等 / ⚪ 轻微]
Alert history:
□ 同指标历史预警：[X] 次 / 过去 [X] 期
→ 最近一次预警：[YYYY-MM] — [描述]
→ 是否重复预警：[是/否]
Step 2: Data Verification
Data accuracy check:
□ 数据来源：[系统名称]
□ 数据提取时间：[YYYY-MM-DD HH:MM]
□ 数据口径：[与标准定义一致/存在口径差异]
□ 异常数据排查：
→ 重复计算：[有/无] — [描述]
→ 数据遗漏：[有/无] — [描述]
→ 录入错误：[有/无] — [描述]
→ 口径变化：[有/无] — [描述]
□ 数据核实结果：[✅ 数据准确 / ⚠️ 数据有误但可修正 / 🔴 数据不可用]
After excluding system/data issues, confirm:
□ 核实后指标值：[X]（vs 阈值 [X]）
□ 是否仍触发预警：[✅ 是 / 🔴 否（数据问题导致误报）]
Step 3: Cause Analysis
Rapid attribution:
□ 归因方向：
→ [业务因素]：实际业务表现低于预期
→ [数据因素]：数据质量问题
→ [预算/目标因素]：预算设定过于激进
→ [季节性因素]：季节性波动
→ [一次性因素]：偶发事件
□ 核心原因定位：
→ 最可能原因：[描述]
→ 置信度：[高/中/低]
→ 依据：[描述]
Impact scope assessment:
□ 影响业务范围：[全公司/特定业务线/特定区域]
□ 影响金额估算：[X] 万（对毛利/利润的影响）
□ 可逆性：[完全可逆/部分可逆/不可逆]
Step 4: Preliminary Response Actions
Immediate actions:
□ 已采取行动：[描述] — 时间 [YYYY-MM-DD]
□ 初步应对是否有效：[✅ 是 / ⚠️ 待观察 / 🔴 无效]
□ 持续监控计划：
→ 监控频率：[每日/每周]
→ 关注指标：[指标列表]
→ 预警阈值调整：[是/否 — 调整至 [X]]
Escalation judgment:
□ 是否需要升级：[是/否]
□ 升级条件检查：
→ KPI 缺口 > [X]%：[✅/🔴]
→ 预计影响金额 > [X] 万：[✅/🔴]
→ 趋势持续恶化（> [X] 期）：[✅/🔴]
→ 涉及多个业务线：[✅/🔴]
□ 升级路径：
→ 升级至：[直属上级/CFO/CEO]
→ 升级理由：[描述]
→ 建议行动：[描述]
Step 5: Generate Alert Handling Report
═══════════════════════════════════════
指标预警处理报告
预警时间：[YYYY-MM-DD HH:MM]
处理时间：[YYYY-MM-DD]
═══════════════════════════════════════
【预警信息】
□ 指标：[名称]
□ 预警类型：[类型]
□ 预警级别：[🔴 严重 / 🟡 中等 / ⚪ 轻微]
□ 触发值：[X] | 阈值：[X]
【数据核实】
□ 数据准确性：[✅ 确认 / 🔴 误报]
□ 核实后是否仍触发：[✅ 是 / 🔴 否]
【原因分析】
□ 归因方向：[描述]
□ 核心原因：[描述]
□ 影响范围：[范围]
□ 影响金额：[X] 万
【应对行动】
□ 已采取行动：[描述]
□ 初步效果：[✅/⚠️/🔴]
□ 持续监控：[是/否]
→ 频率：[每 X]
→ 关注指标：[列表]
【升级决策】
□ 是否升级：[是/否]
□ 升级至：[对象]
□ 升级理由：[描述]
□ 建议行动：[描述]
□ 下次评估时间：[YYYY-MM-DD]
═══════════════════════════════════════
置信度：[✅ 高 / ⚠️ 中 / 🔴 低]
预警状态：[✅ 已处理 / ⏳ 监控中 / 🔴 已升级]
═══════════════════════════════════════
Escalation Trigger Conditions
- Alert metric gap > [X]% (e.g., 20%) and no clear sign of improvement
- The same metric triggers an alert for [X] consecutive periods (e.g., 3 periods)
- The alert involves multiple metrics deteriorating simultaneously
- Estimated impact amount > [X] ten thousand
- Preliminary response actions are ineffective and the trend shows no improvement
- The cause of the alert is an external uncontrollable factor (market/policy)
Finance Skills — metric-alert-handling atomic skill