cvm-operator

Automate the CVM pipeline with reproducible scripts for scoring and KPI summaries.

3|1|Updated Feb 26, 2026

Search Tags:#pipeline #scoring #orchestrator #exports #kpi #cvm #top-n

Ready to run

npx skills add https://github.com/KiselevAleksey/cvm --skill cvm-operator

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: cvm-operator
Source: https://github.com/KiselevAleksey/cvm/tree/main/skills/cvm-operator
Command: npx skills add https://github.com/KiselevAleksey/cvm --skill cvm-operator

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires duckdb, pandas, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Operates the CVM pipeline with deterministic scripts to avoid ad-hoc command sequences and ensure reproducible results across runs.

Core Features & Use Cases

End-to-end orchestration of data preparation, scoring, and artifact export
Support for real and demo modes, including top-N customer exports
Reproducible, auditable runs with standardized outputs and KPI summaries

Quick Start

Run the end-to-end CVM workflow using the provided run_cvm_flow.sh script in scripts, selecting real or demo mode as needed.

Frequently Asked Questions about cvm-operator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate a CVM pipeline for churn and cross-sell scoring end-to-end?▼

You can automate a CVM pipeline end-to-end by running deterministic scripts that handle data preparation, churn and cross-sell scoring, and artifact export. Use the provided shell script to execute the entire workflow reproducibly in real or demo mode.

What is the best way to generate reproducible top-N customer lists from CSV drops?▼

Generating reproducible top-N customer lists requires standardized scripts that onboard CSV drops and run scoring models. This approach ensures auditable runs with consistent outputs and KPI summaries across single-use-case or all-use-cases executions.

Do I need pandas and duckdb to run CVM scoring workflows?▼

Yes, pandas and duckdb are required dependencies for running CVM scoring workflows. These frameworks support the underlying data processing and orchestration logic executed by the deterministic scripts located in the scripts directory.

Can I run customer valuation models in demo mode before processing real data?▼

Yes, you can run customer valuation models in demo mode before processing real data. The orchestration scripts support both real and demo modes, allowing you to test top-N customer exports and KPI summaries safely.

How does the CVM orchestrator ensure reproducible data-to-decision workflows?▼

The CVM orchestrator ensures reproducible data-to-decision workflows by using deterministic scripts and adhering to a strict output contract. This standardizes run artifacts and KPI summaries, avoiding ad-hoc command sequences across multiple runs.

What outputs should I expect from an automated reactivation scoring run?▼

From an automated reactivation scoring run, you should expect top-N actionable customer lists, run artifacts, and KPI summaries. All outputs adhere to the standardized contract defined in the output-contract reference document.