Support online 08:00-24:00 (UTC+8) - manual top-ups are usually done within 1 hour once details are complete, and night orders are handled after 8:00 the next morning

See pricing →
Home Top-up Skills Guides My orders

Requires

What you need before running this pack. Miss one and it cannot do the work.

  • Python python3 runs on your machine — one command, and it does not touch your files
  • Report / exported CSV You export a CSV or JSON from the platform admin as the input

Picked most often by

Who buys it most, by role. Click a role to see only that role's skills.

Similar skills

View all →
◫

亚马逊账号申诉包(POA / 视频验证 / 资金解冻)

跨境卖家收到亚马逊业绩通知后,判清 12 类违规中的哪一类、算出 9 类必备材料里自己缺哪几件、生成中英双语三段式 POA,并在提交前用 11 项闸门把驳回理由提前跑一遍。
Amazon Operations Owner / operator Account appeals & compliance
¥349 Al
Details Get it
◫

欧盟合规包:上架前的四关自查(GPSR / PPWR / EPR / 欧代)

把 listing 文本与你自己填的合规底账,逐条对着 GPSR / PPWR / EPR / WEEE / 电池法 / 宣称类的 19 条条款过一遍,判「必改 / 建议改 / 需人工
Amazon DTC store (Shopify) Multi-platform Operations +4
¥349 Al
Details Get it

What is included

  • ✓ metric-attribution — metric-attribution:指标变化归因
  • ✓ anomaly-alert — anomaly-alert:阈值告警与排查顺序
  • ✓ review-report — review-report:一页纸复盘 + 三条实验
  • ✓ lesson-ledger — lesson-ledger:经验沉淀

Who it is for

  • ✓ People who need steady output in a fixed scene (Owner / operator / Operations)
  • ✓ You already have assets — you need structure
  • ✓ Want to set it up once and reuse it, instead of re-explaining every time

Not for

  • – Want full automation with no input at all
  • – One-off use, no reuse planned
  • – Platform-specific ad console actions (do those manually)
Try saying this

Use “运营复盘包(每周 20 分钟,把一轮运营变成下一轮实验)” to make me Web pages — I will paste the input below.

Copy this line to your agent to start
  • Versionv0.1.0
  • Updated2026-10-24
  • OutputWeb pages

About this skill

两期经营数据 + 本期动作清单进,出五样东西:指标变化归因(带待验证假设)、阈值告警与排查顺序、一页纸三段式复盘、三条带有效/无效/负向判定标准的实验、可复用的经验条目。缺数据就写缺失,不编数、不承诺增长。

Sample output

Real input → real output, unpolished
真实输出节选|examples/01-weekly-full-cycle/after/attribution-notes.md Output
# 指标变化归因:2026-W42 对比 2026-W41

这份文件只做三件事:把两期数字摆出来、指出哪几项值得看、给每条变化写一个**待验证的假设**。
它不下因果结论——动作和指标同周发生只说明「时间上挨着」,能不能算数要看下期数据。

## 一、变化一览

| 指标 | 上期 | 本期 | 变化 | 幅度 | 判定 |
|---|---|---|---|---|---|
| 销量(件) | 218 | 246 | +28 | +12.84% | 改善 |
| 会话数(次) | 9480 | 9860 | +380 | +4.01% | 正常 |
| 转化率(%) | 2.3 | 2.49 | +0.19 | +8.26% | 改善 |
| 广告花费(USD) | 640 | 705 | +65 | +10.16% | 需关注 |
| 广告销售额(USD) | 1720 | 2105 | +385 | +22.38% | 改善 |
| 退货率(%) | 6.8 | 7.6 | +0.8 | +11.76% | 需关注 |
| 商品评分(星) | 4.42 | 4.31 | -0.11 | -2.49% | 需关注 |
| 评价数(条) | 87 | 95 | +8 | +9.20% | 改善 |
| ACOS(%) | 37.21 | 33.49 | -3.72 | -9.99% | 改善 |

共 9 项指标:需关注 3 / 改善 5 / 正常 1 / 本期缺失 0 / 无上期基线 0。

## 二、需要验证的假设

### 假设 1 · 退货率(%)(6.8 → 7.6,+11.76%)

- **可能对应的动作**:A-04 服装 3 个 SKU 补实测尺码表与尺寸图
- **为什么不能当成因果**:退货率在涨。先按原因拆开(尺码 / 图文不符 / 破损 / 物流延迟),不要笼统归成质量问题。 另有动作存在滞后:A-04(2026-10-16)距本期结束仅 2 天,退货率(%)通常要 10 天才反映,本期变化不能归给它。
- **怎么验证**:导出退货原因明
真实输出节选|examples/01-weekly-full-cycle/after/alerts.csv Output
metric,label,value,prev,delta,level,threshold,why,triage,owner
return_rate,退货率(%),7.6,6.8,+0.8,红,告警线 6.0、红线 7.5,退货率越过 7.5%:退货成本按单笔算通常远高于毛利里预留的那几个点,先止血再找原因。,① 先拉本期退货原因明细,按尺码 / 图文不符 / 破损 / 物流延迟 / 买家反悔五类算占比(半小时,最便宜) ② 按 SKU 排序,看退货是不是集中在 1–3 个 SKU;集中在少数 SKU 就是商品或图文问题,分散就是流程问题 ③ 核对本期是否改过尺码表、主图、包装或发货渠道;改过就先看那一项 ④ 最后才讨论下架或改文案,因为改文案不能解决破损和物流延迟,客服 + 运营
acos,ACOS(%),33.49,37.21,-3.72,黄,告警线 30.0、红线 45.0,ACOS 在 30%–45%:还能接受,但要看是哪个 campaign 拖的,不要全店一刀切。,① 先按 campaign 拆 ACOS,找出最贵的那一个(不要看全店平均) ② 再看该 campaign 的搜索词报告,把花费高、零转化的词挑出来加否定 ③ 核对出价与预算有没有在本期被改过;改过就先回滚再看 ④ 最后才考虑降整体预算,因为一刀切会把还在赚钱的词一起关掉,广告投放 + 运营
rating,商品评分(星),4.31,4.42,-0.11,黄,较上期下滑 ≥0.1 星告警、≥0.25 星红线;低于 4.0 星直接红线,评分较上期下滑 0.10 星以上:滚动平均值的小幅下滑,通常是某一批订单的时效或包装问题。,① 先把本期新增的 1–3 星评价原文逐条读完,按原因打标签(时效 / 破损 / 质量 / 描述不符) ② 对得上订单的,查那批订单的发货仓与承运商 ③ 看差评是否集中在某一个变体或生产批次 ④ 最后才讨论补偿或删评,因为评分是结果,原因在前面三步,客服 + 运营
units_sold,销量(件),246,218,+12.8%,绿,较上期下滑 ≥10.0% 告警、≥20.0% 红线

Install guide

One column per agent, each with a check that proves it installed. If the check fails, it is not installed.

Claude Code

Install to .claude/skills/ops-review/
mkdir -p .claude/skills
unzip ops-review.zip -d .claude/skills/

How to verify Open a new session and ask it to list the installed skills — seeing this pack's name means it worked.

For a global install, swap .claude/skills/ for ~/.claude/skills/ so every project of this user sees it.

Codex

Install to ~/.codex/skills/ops-review/
mkdir -p ~/.codex/skills
unzip ops-review.zip -d ~/.codex/skills/

How to verify Open a new codex session and ask it to list available skills — this pack showing up means it is live.

Codex only reads ~/.codex/skills/, not the project folder; on another machine you place it again.

Cursor

Install to .cursor/rules/ops-review.mdc
mkdir -p .cursor/rules
cp ops-review/SKILL.md .cursor/rules/ops-review.mdc

How to verify In chat, @ that rule file and ask what the pack's first hard rule is — a correct answer means it loaded.

Cursor has no separate skills folder: rules belong to the project, so you place them again per project — the biggest difference from the other two.

If it will not installNot opening a terminal at all? Paste the whole SKILL.md into the chat. Same result, you just re-paste it each new session.

Full steps and common errors live in the pack's playbook.md; the guides section also has ordering and usage guides。

How to use

4 steps; the first run takes about 10 minutes
  1. 1
    开工前准备 | 要做 | 怎么做 | 为什么 | |---|---|---| | 建一个复盘目录 | 桌面建 `复盘/`,里面建 `2026-W42/`,再建 `before/`、`after/` | 每周一个文件夹,下期对照上期时不会翻聊天记录 | | 备份原始数据 | 后台导出的 CSV 原文件单独存一份,不要直接改 | 复盘要改口径时能退回去 | | 抄指标 | 把 9 项指标抄进 `before/metrics.csv`:销量、会话数、转化率、广告花费、广告销售额、退货率、评分、评价数(ACOS 不用抄,脚本算) | 脚本按列名认字段,抄漏一列就少一项结论 | | 记动作 | 本期做过的事记进 `before/actions.csv`,每条写 `metric_hint`(想影响哪个指标) | 不记动作,归因就没有「
  2. 2
    装配 **Claude Code**:复制到项目的 `.claude/skills/`(没有就先建)。 **Codex CLI**:复制到 `.codex/skills/`。 **没有 shell**:不用装。把 `skills/metric-attribution/SKILL.md`、`skills/anomaly-alert/SKILL.md`、`skills/review-report/SKILL.md`、`skills/lesson-ledger/SKILL.md` 依次贴进对话,让它按规则手工执行(脚本那几步退化成口头执行,产出要你自己写进 CSV)。 装完确认(一条动作):新开一个对话,问 **「列出你现在可用的 skills」**,应看到 `metric-attribution`、`anomaly-al
  3. 3
    触发顺序 四关必须按顺序,原因不是习惯: 1. **先归因(1),再告警(2)**:告警要判「比上期跌了多少」,它读的是同一份 `metrics.csv`,但结论要挂在归因表里「这个指标本来就值得看」的前提下。反着做会怎样:告警把销量 +12.84% 也拉进排查清单,你花时间看一个没问题的指标。 2. **先 1 和 2,再报告(3)**:报告的三段式是把「变化」和「红线」合成人话,缺任何一份它只能写「建议优化」。反着做会怎样:报告里「做错了什么」整段空着,老板以为你这一期没犯错。 3. **沉淀(4)放最后**:它只引用**历史**实验的判定,不引用本期结果——本期刚做的动作还没被判过,不能进经验库。反着做会怎样:把本周刚做的事写成「有效经验」,下期拿它指导动作,等于自己给自己编因果。 你原话可以直接抄: 第一关:**
  4. 4
    每一步的输入与输出 ### 第 1 关 metric-attribution | 项 | 内容 | |---|---| | 输入 | `before/metrics.csv`(`period,metric,value[,unit,note]`,两期)+ `before/actions.csv`(`action_id,date,action,metric_hint`) | | 命令 | `python3 skills/metric-attribution/scripts/attribute.py --metrics before/metrics.csv --actions before/actions.csv --current 2026-W42 --previous 2026-W41 --current-end 2026-10-1

Reviews

Only verified buyers count; we do not edit, delete or pad them.
No reviews yet — buyers can review after purchase

This pack just launched and nobody has reviewed it yet. Once real buyers do, the stars, text and 1–5 distribution appear here; until then the rating shows an em dash instead of a made-up score.

Changelog

Updates are free while subscribed; a buyout includes later versions of this skill
v0.1.0 2026-10-24 按包内 manifest.json 上架

How to buy

Subscribe or buy outright — a buyout means no more subscription fees
Cancel anytime

Monthly

¥39 / mo

Try it for a month first

  • ✓ “运营复盘包(每周 20 分钟,把一轮运营变成下一轮实验)” — all files, samples and install instructions
  • ✓ Every skill unlocked while subscribed (including new releases)
  • ✓ Commercial licence: client work allowed, unlimited deliveries
  • ✓ Version updates pushed automatically, no repurchase
  • ✓ 1:1 support on Telegram, replies within 12 hours on business days
  • ✓ Cancel anytime, no next-period charge
Save 36% Most popular

Yearly

¥299 / yr

Pay for a year; prices are never converted to a monthly figure.

  • ✓ “运营复盘包(每周 20 分钟,把一轮运营变成下一轮实验)” — all files, samples and install instructions
  • ✓ Every skill unlocked while subscribed (including new releases)
  • ✓ Commercial licence: client work allowed, unlimited deliveries
  • ✓ Version updates pushed automatically, no repurchase
  • ✓ 1:1 support on Telegram, replies within 12 hours on business days
  • ✓ Cancel anytime, no next-period charge
One payment · founder price Founder price

Lifetime

¥899 / one-time

The founder price rises as the skill library grows.

  • ✓ “运营复盘包(每周 20 分钟,把一轮运营变成下一轮实验)” — all files, samples and install instructions
  • ✓ Every skill unlocked while subscribed (including new releases)
  • ✓ Commercial licence: client work allowed, unlimited deliveries
  • ✓ Version updates pushed automatically, no repurchase
  • ✓ 1:1 support on Telegram, replies within 12 hours on business days
  • ✓ No subscription or renewals — one payment, forever

No price conversion: yearly plans only show the saving, never a per-month figure. A buyout is one payment with no renewals.

FAQ

How do I install it into my agent?
Unzip the ZIP inside your project and put the folder into your agent skills directory (Claude Code: .claude/skills/, Codex: ~/.codex/skills/), then restart the session.
What does it depend on?
Just an agent that can read files. For images or video, connect your own generation tool; the skill ships with parameter templates.
Can I get a refund?
No no-reason refunds after digital delivery. If files are missing or broken, we verify within 48 hours and replace or refund.