agentleFS
zerlinpizerlinpi/amazon-ads-skills/skills/campaign-health-monitor/SKILL.md

campaign-health-monitor

Skill8 starsChanged 57 days agoMITView on GitHub

Monitor routine campaign health status across Amazon Ads traffic, spend, conversion, budget and efficiency using historical baselines; use for scheduled checks, watchlists and broad campaign triage, not deep root-cause work. 中文:用于日常 Campaign 健康巡检、状态分级和周趋势监控。

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When a conclusion depends on live MCP/API/connector data, load ../../references/connector-capability.md before interpreting missing or …

    line 18
  2. Distinctive:

    search term、placement 或 delivery-surface mix 是否变化。

    line 124

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Campaign Health Monitor

Connector capability gate

When a conclusion depends on live MCP/API/connector data, load ../../references/connector-capability.md before interpreting missing or empty fields. If a machine-readable capability snapshot is available, identify the exact required capability IDs and run ../../scripts/evaluate_connector_capability_gate.py before metric interpretation.

  • Pass — continue with the Skill's normal evidence, sufficiency, lineage and safety checks.
  • Degraded — do not make a high-confidence dependent recommendation; keep the result to Directional, Hold, Alternate Source, Missing Data, or Manual Review.
  • Blocked — do not treat the dependent observation as action-safe until the capability is resolved or an allowed alternate source is verified.
  • missing_evidence_policy = never_zero — Partial, Unsupported, Unknown, or absent connector capability is never a numeric zero, unchanged state, or proof that the platform lacks the feature.

This gate is read-only and does not authorize live Amazon Ads mutation.

目标

发现“真正需要运营人员注意”的 campaign,而不是对正常波动产生告警疲劳。

默认模式

Suggest。只输出状态、证据和动作候选。

Routing boundary

  • 这是例行状态扫描 / watchlist / campaign triage Skill:回答“哪些 Campaign 健康、需要观察或值得升级调查”。
  • 如果用户主要是在判断一个突然信号究竟是不是异常、业务事件还是正常波动,交给 ../anomaly-detection/SKILL.md。
  • 如果已经确认存在持续且有业务影响的下降,并且问题是“为什么下降、损失从哪里来、哪一个原因有证据”,交给 ../performance-drop-diagnosis/SKILL.md。
  • Healthy 只表示当前没有明确健康阻塞,不自动等于存在可盈利扩量空间;扩量判断交给 ../growth-opportunity-finder/SKILL.md。

Progressive loading

只在需要时加载共享规则:

  • ../../references/benchmark-policy.md:当 Amazon Benchmark reporting、同行/类目中位数或外部 benchmark 影响健康判断时;
  • ../../references/data-lineage.md:当当前窗口与基线来自不同 report/source/semantic definition 或成熟度时;
  • ../../references/marketing-stream-dataset-coverage.md:当 Amazon Marketing Stream 被用于小时健康判断时,先分清 v1 独立 traffic/conversion dataset 与 v2 统一 performance dataset;对 v2 必须确认逐记录最高 streamBatch.version 已选出,再进行任何累计指标或异常判断;
  • ../../references/decision-boundaries.md:当输出具体 bid/budget/placement 动作候选时。

输入

至少需要:

  • campaign 当前窗口数据;
  • 一个可比较历史窗口或滚动基线;
  • 日期、Marketplace、币种;
  • campaign state、budget;
  • impressions、clicks、spend、orders、sales。

最好额外提供:placement、budget status、bid strategy、库存/促销/价格变化。

如果使用 peer/category benchmark,还应记录 benchmark source、metric、percentile/statistic、peer/cohort fit、window/retrieval date 和 eligibility/availability status。缺失 benchmark 不自动补 0。

健康维度

Traffic

  • impressions 是否异常下降/上升;
  • clicks/CTR 是否偏离基线;
  • campaign 是否突然失去流量。

Cost

  • CPC 是否突增;
  • spend 增长是否由流量增加还是 CPC 增加驱动。

Conversion

  • CVR 是否显著低于自己的历史基线;
  • orders/sales 下滑是否只是流量下滑导致。

Efficiency

  • ACOS/ROAS 相对目标与历史变化;
  • 不在 sales=0 时计算伪 ACOS。

Budget

  • 是否频繁预算耗尽;
  • 预算受限 campaign 是否本身高效;
  • 是否存在低效 campaign 占用大量预算。

Delivery

  • state/eligibility 是否异常;
  • 是否存在 Listing/Buy Box/库存影响。

Peer context(可选)

  • Amazon Benchmark reporting 或其他可比较 peer cohort 可用于补充“市场上下文”;
  • peer median / 25th / 75th percentile 是比较证据,不是仓库默认健康阈值;
  • 低于 peer median 不代表 campaign 自动为 Critical,也不证明 bid、budget、placement、targeting 或 negative 中的任何一个是根因;
  • 高于 peer benchmark 也不等于存在可盈利扩量空间。

状态规则

不要依赖固定百分比阈值或 peer benchmark 作为唯一依据。结合:

  • 自身历史波动区间;
  • 流量体量;
  • 目标与单位经济;
  • 业务事件;
  • 连续性;
  • benchmark 的 cohort fit 与方法适配度(若使用)。

状态含义:

  • Healthy:核心指标在合理区间,无明确阻塞;
  • Watch:出现值得观察的偏离,但证据不足以做强动作;
  • Critical:持续、显著且有业务影响的问题,或 delivery/数据层硬故障。

单纯“低于同行中位数/25th percentile”不足以构成 Critical。

根因拆解

当 sales/ROAS 下降时按顺序拆:

  1. impressions 是否下降;
  2. CTR 是否下降;
  3. CPC 是否上升;
  4. clicks 是否不足;
  5. CVR 是否下降;
  6. price/promotion/inventory/Buy Box 是否改变;
  7. search term、placement 或 delivery-surface mix 是否变化。

这里的拆解用于决定是否升级调查以及升级到哪个 Skill,不要把 campaign-health-monitor 本身当成深度根因分析器。持续业务影响需要交给 performance-drop-diagnosis。

不要把结果指标或 benchmark gap 当根因。

输出格式

对每个需要关注的 campaign 输出:

Campaign: <name/id>
Status: Healthy | Watch | Critical
Primary signal: <主要异常>
Evidence: <当前值 vs 基线>
Peer context: <optional benchmark + cohort fit + what it does not prove>
Likely causes: <按可能性排序>
Business impact: <影响>
Confidence: 0-1
Next action: <调查或优化建议>
Validation window: <再次观察条件>

然后给出:

  • Critical campaigns;
  • Watch list;
  • Healthy campaigns that may warrant growth review;
  • Account-wide pattern。

对第三类只标记“值得进入增长审查”的候选;真正的 scale/headroom 判断必须路由到 growth-opportunity-finder,不能从 Healthy 状态直接推出加预算/加 bid。

特殊事件保护

如果窗口覆盖 Prime Day、Best Deal、LD、Coupon、价格促销或断货,先标记事件,再解释相对基线偏差,避免将正常活动波动误报。

安全

任何 bid/budget/placement 变化只作为 Suggest action proposal。实际执行交给外部 executor,并遵守 ../../references/decision-boundaries.md。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.