cfgit

Detect and reconcile drift in live datastore records with version control.

2|Updated Jun 21, 2026
One-click install
npx skills add https://github.com/AusafMo/cfgit --skill cfgit
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cfgit
Source: https://github.com/AusafMo/cfgit/tree/main/skills/cfgit
Command: npx skills add https://github.com/AusafMo/cfgit --skill cfgit

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a safety layer for live datastore records, enabling version control, history, and drift detection for data that is typically edited out-of-band by scripts or admin tools.

Core Features & Use Cases

  • Drift Reconciliation: Detects and adopts changes made to database records outside of the cfgit workflow.
  • Branching & PRs: Allows for safe, draft-based experimentation on runtime configuration before merging changes.
  • Use Case: Manage agent prompts, model routing, or feature flags in a production database with the ability to diff, rollback, and audit every mutation as if it were source code.

Quick Start

Use the cfgit skill to inspect the current drift status of your database records and adopt any out-of-band changes into the version history.

Frequently Asked Questions about cfgit

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

FAQPage Schema
How do I version control live database records for configuration changes?

Version control for live database records is achieved by applying diffing, branching, and pull request validation workflows to opaque JSON configurations, preventing unauthorized or accidental data clobbering in production environments.

What is drift reconciliation for production database configurations?

Drift reconciliation is the process of detecting out-of-band changes made to database records outside the version control workflow, allowing you to inspect and adopt those modifications into the official version history.

Can I use branching and pull requests to manage runtime configuration like feature flags?

Yes, you can use draft-based branching and pull request validation to safely experiment with runtime configurations such as feature flags, model routing, and agent prompts before merging changes to production.

How do I detect out-of-band changes to JSON records in a live datastore?

You can detect out-of-band changes to opaque JSON records by inspecting the drift status of your live datastore, which identifies modifications made by external scripts or admin tools outside the versioned workflow.

What is the best way to rollback accidental mutations to model routing configurations?

The best way to rollback accidental mutations is by treating database configurations like source code, using version history and diffing to audit every change and safely revert unauthorized modifications to model routing or policy configurations.

Do I need external dependencies to manage versioning for opaque JSON records?

No external dependencies are required to manage versioning for opaque JSON records, as the workflow operates independently to provide diffing, branching, and drift detection for control-plane data within production environments.