audit-agent

This skill configures Claude Code sandbox network isolation by setting up trusted domain whitelists and custom access policies. It helps developers secure code execution by managing environment variables and preventing unauthorized network access. Use it to implement zero-trust architecture and prevent prompt injection attacks via network controls.

This skill provides automated quality verification for code and agent outputs using truth scoring and quality checks. It automatically rolls back changes that fall below a 0.95 accuracy threshold, ensuring codebase reliability. Use it for CI/CD integration and maintaining high-quality standards in development workflows.

The sandbox-configurator skill automatically configures Claude Code's execution environment with security boundaries for file system and network isolation. It enables developers to define trusted domains, manage file permissions, and control network access for secure code execution. Use this skill when you need to set up a secure, isolated sandbox for running untrusted code or building applications.

This skill automates GitHub Actions workflows with AI swarm coordination for intelligent CI/CD pipelines and repository management. It generates, analyzes, and orchestrates workflows using adaptive automation capabilities. Use it when you need to streamline GitHub automation with self-organizing, multi-agent coordination.