When should I use the Anomaly Detection skill?

Activate this skill at the very beginning of any data ingestion or processing flow to prevent poor data from affecting downstream analysis or wasting resources.

Does this skill fix broken data?

No, the Anomaly Detection skill is non-destructive and acts specifically as a pre-filter; it identifies, flags, or blocks records rather than modifying the original content.

How does it handle suspicious but valid data?

It uses a 'Bias for Flow' approach: records that are suspicious but not logically impossible are flagged and passed with a reduced confidence score rather than being blocked.

What kind of anomalies can it detect?

It detects structural errors (missing IDs), temporal anomalies (future timestamps), value outliers (negative prices), pattern spikes, and systemic source failures.

Anomaly Detection

Name: Anomaly Detection
Author: dvcrn

bydvcrn

•

분석 및 모니터링

Inspects and filters raw data records to identify structural, temporal, and value-based anomalies before downstream processing.

The Anomaly Detection skill acts as a high-performance pre-filter for data-intensive applications, ensuring that malformed or logically impossible records do not degrade the performance of downstream models or waste computational resources. By categorizing anomalies into structural, temporal, value, and pattern-based issues, it provides a structured assessment with confidence scores, allowing developers to maintain high data integrity in their pipelines. It is designed for speed and safety, blocking data that could cause system crashes while flagging suspicious records for further manual or automated review.

주요 기능

01Automated data blocking for malformed or unsafe records

023 GitHub stars

03Systemic failure signaling for batch ingestion monitoring

04Standardized output schema with confidence modifiers and plain-English details

05Multi-category anomaly classification including structural and temporal checks

06Non-destructive assessment that preserves original data integrity

사용 사례

01Validating incoming transaction records to prevent database poisoning

02Monitoring data ingestion pipelines for systemic source failures or schema changes

03Preprocessing raw time-series data for financial or blockchain applications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dvcrn/openclaw-skills-marketplace anomaly-detection

For use in Claude.ai and ChatGPT

주요 기능

01Automated data blocking for malformed or unsafe records

023 GitHub stars

03Systemic failure signaling for batch ingestion monitoring

04Standardized output schema with confidence modifiers and plain-English details

05Multi-category anomaly classification including structural and temporal checks

06Non-destructive assessment that preserves original data integrity

사용 사례

01Validating incoming transaction records to prevent database poisoning

02Monitoring data ingestion pipelines for systemic source failures or schema changes

03Preprocessing raw time-series data for financial or blockchain applications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dvcrn/openclaw-skills-marketplace anomaly-detection

For use in Claude.ai and ChatGPT