bledden

Anomaly Hunter — Data skill for Claude Code

Data community

Multi-agent anomaly detection with real-time investigation.

How to install Anomaly Hunter

This entry records only its repository, not the path inside it, so there is no exact command to give. Open bledden/anomaly-hunter and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Anomaly Hunter does

Multi-agent anomaly detection with real-time investigation. Built with 8 AI/ML sponsors for autonomous data quality monitoring.

Alternatives in Data

Claude Code Usage Monitor — by Maciek-roboblog - A real-time terminal-based tool for monitoring Claude Code token usage 7k ★
Check Data Quality — Renamed → /portaljs-check-data-quality 2.3k ★
Azure Event Hub Java — Real-time high-throughput streaming for Java 1.8k ★

README

Anomaly Hunter - Autonomous Data Quality Monitor

**Multi-agent anomaly detection with real-time investigation**

[![Python 3.9+](https://img.shields.io/badge/python-3.9+-blue.svg)](https://www.python.org/downloads/) [![9 Sponsors](https://img.shields.io/badge/sponsors-9-green.svg)](#sponsors) [![60+ Detections](https://img.shields.io/badge/detections-60+-orange.svg)](#autonomous-learning) [![5 Domains Validated](https://img.shields.io/badge/domains-5%20validated-purple.svg)](#beyond-network-monitoring-universal-anomaly-detection) [![Built with Corch](https://img.shields.io/badge/built%20with-Corch-orange.svg)](https://github.com/bledden/weavehacks-collaborative)

Built for **Production Readiness** - October 2025

[>>] What It Does

Anomaly Hunter autonomously detects and investigates data anomalies using 3 specialized AI agents working in parallel:

Pattern Analyst (GPT-5 Pro) - Statistical anomaly detection
Change Detective (Claude 4.5 Sonnet) - Time-series drift analysis
Root Cause Agent (Claude 4.5 Sonnet) - Dependency graph investigation

**Result:** Real-time anomaly detection with confidence-scored root cause analysis.

Multi-Model Collaboration Architecture

**Inspired by Facilitair's proven sequential collaboration approach** (73% pass rate, +36.8% quality improvement), Anomaly Hunter implements **multi-agent consensus** instead of relying on a single model for critical decisions.

**Why Multiple Models Beat Single-Model Approaches:**

**Diverse Perspectives**: GPT-5 Pro excels at statistical pattern recognition, while Claude 4.5 Sonnet provides nuanced time-series analysis and causal reasoning. Each model brings different strengths.
**Error Correction Through Consensus**: When agents disagree, the orchestrator synthesizes findings and flags uncertainty. A single model can hallucinate or miss context - three independent analyses reduce false positives.
**Specialization Over Generalization**: Rather than asking o