首页龙虾技能列表 › Data Completeness Check — 技能工具

Data Completeness Check — 技能工具

v1.0.0

数据完整性判断机制。通过前晚基准值 vs 早上值对比,判断 Power BI 数据是否回灌完毕。适用于任何有延迟回灌的数据源。避免发送不完整数据。

0· 53·0 当前·0 累计
by @mirowangl-ops·MIT-0
下载技能包
License
MIT-0
最后更新
2026/4/11
安全扫描
VirusTotal
无害
查看报告
OpenClaw
安全
high confidence
Instruction-only skill that describes a simple two-timepoint metric comparison to judge whether Power BI data has been fully backfilled; it declares no installs, credentials, or unusual access and is internally consistent with its purpose.
评估建议
This skill is high-level and coherent, but check these practical points before enabling it: 1) Decide and document how metrics will be fetched (Power BI API, DB query, exported file) and ensure any credentials used are minimal-scope and provided separately from this skill. 2) If you allow the agent to call connectors autonomously, confirm which tokens/accounts it will use and limit their scope. 3) Define and record where the sampled values and logs will be stored and who can view them. 4) Test t...
详细分析 ▾
用途与能力
The name/description (data completeness check for Power BI / T+1 sources) matches the instructions (capture a sentinel metric at 23:30 and 09:00 and compare). Nothing requested (no env vars, no binaries, no installs) is disproportionate to that stated purpose.
指令范围
The SKILL.md is high-level and prescribes a manual/automated sampling workflow (capture metric at two times, compare, use thresholds). It does not specify HOW to fetch metrics (e.g., Power BI API, DB query, or exported file) or where results are recorded. This ambiguity is not inherently malicious but means implementers must decide which connector/credentials to use.
安装机制
No install spec and no code files are present (instruction-only), so nothing is written to disk and no third-party packages are pulled. This is the lowest-risk install surface.
凭证需求
The skill declares no required environment variables, secrets, or config paths. The described task could require Power BI or database credentials in practice, but those are not requested by the skill itself—so the declared access is proportionate.
持久化与权限
always is false and the skill does not request persistent presence or modify other skills or agent-wide settings. Autonomous invocation is allowed by platform default but is not excessive given the simple nature of the instructions.
安全有层次,运行前请审查代码。

License

MIT-0

可自由使用、修改和再分发,无需署名。

运行时依赖

无特殊依赖

版本

latestv1.0.02026/4/11

- Introduced a data completeness checking mechanism based on dual time-point comparison (late night vs. morning values). - Outlined decision rules to verify if Power BI data has been fully updated before sending reports, preventing incomplete data delivery. - Provided practical reference thresholds for key metrics (e.g., Paid Subs, Revenue) on weekdays and weekends/holidays. - Made the approach generalizable for any delayed (T+1) data sources by selecting a key "sentinel" metric and establishing experience-based reference ranges.

● 无害

安装命令 点击复制

官方npx clawhub@latest install xiaozhua-data-completeness-check
镜像加速npx clawhub@latest install xiaozhua-data-completeness-check --registry https://cn.clawhub-mirror.com

技能文档

Overview

This skill provides a mechanism to check data completeness before sending reports. It compares a "sentinel" metric captured at two time points (late night vs. morning) to determine whether the data has been fully backfilled.

When to Use

  • Before sending Power BI reports or dashboards
  • When data sources have known T+1 delay (data available next day)
  • To prevent sending incomplete or stale data to stakeholders

How It Works

1. Select a Sentinel Metric

Choose a key metric that:

  • Is reliably updated when data refresh completes
  • Has predictable values (e.g., Paid Subscriptions, Total Revenue)
  • Serves as an indicator that the full data pipeline has finished

2. Capture Values at Two Time Points

Time PointTypical ValuePurpose
23:30 (previous night)Baseline/previous day final valueReference point
09:00 (morning)Current day partial or complete valueCheck if refresh completed

3. Compare and Decide

IF morning_value >= threshold_percentage of baseline_value:
    THEN data is complete -> proceed to send report
    ELSE data may be incomplete -> wait and retry or alert

Reference Thresholds

Weekdays (Monday - Friday)

  • Paid Subs: Expect 95-105% of previous day
  • Revenue: Expect 90-110% of previous day (depends on time of month)
  • Active Users: Expect 85-100% of previous day

Weekends/Holidays

  • Paid Subs: Expect 100-110% (may include weekend signups)
  • Revenue: Expect 80-120% (higher variance)
  • Active Users: Expect 60-90% (lower weekend activity)

Retry Logic

If data appears incomplete:

  • Wait 30 minutes
  • Re-capture the sentinel metric
  • Compare again against threshold
  • If still incomplete after 3 attempts, alert or skip sending

Generalization to Other Data Sources

This approach can be adapted for any T+1 data source:

  • Identify a suitable sentinel metric
  • Establish baseline reference ranges based on historical data
  • Define threshold percentages for "complete" vs "incomplete"
  • Implement the two-timepoint comparison workflow

Example Prompt

Check if the Power BI data is ready for today's sales report. Capture the total subscription count at 23:30 yesterday and 09:00 today. If the morning value is at least 95% of the previous night value, consider the data complete and send the report. Otherwise, wait 30 minutes and retry up to 3 times.

数据来源:ClawHub ↗ · 中文优化:龙虾技能库
OpenClaw 技能定制 / 插件定制 / 私有工作流定制

免费技能或插件可能存在安全风险,如需更匹配、更安全的方案,建议联系付费定制

了解定制服务