# Source and maintenance

- Author/maintainer: Meyermedia
- Created on: 2026-07-21
- Original design source: `https://gist.github.com/mleem97/c5700fe7e58f50a02430fdf92e5a9348`
- Corroborating implementation source: `mleem97/ml-aimanager`
- License: original skill maintained in this repository

This skill generalizes the Zero-Touch, foundation-first, bounded self-healing, quality-gate, and explicit-approval model from the supplied design into a repository-independent Agent Skill.

The source architecture is software-development-oriented. This adaptation applies the same control model to research, analysis, writing, coding, file operations, and production workflows. Specialized roles are treated as internal checkpoints unless the runtime actually executes independent agents.

Before changing version-sensitive behavior, verify current Agent Skills and target-runtime documentation. Do not introduce external writes, destructive defaults, credential requests, opaque executables, or unbounded loops.