The Agent Skills Directory

[SAFE]: The skill implements a structured directory system (.harness/) intended to improve project maintainability and AI agent alignment without introducing dangerous behaviors or unauthorized access.
[SAFE]: Scripts provided in the templates utilize standard local execution patterns (shell and Python) to perform tasks such as directory scaffolding, linting, and architecture drift detection based on local configuration.
[SAFE]: The skill encourages security best practices, such as preventing the commitment of log files to version control and providing rules for detecting hardcoded secrets within the codebase.
[SAFE]: External dependencies and references point exclusively to established and trusted organizations in the AI and software engineering industries, posing no risk of supply chain or remote code execution attacks.

harness-engineering